PERFORMANCE of VARIOUS COMPUTERS in COMPUTATIONAL CHEMISTRY

Martyn F. Guest,
Computational Science and Engineering Department,
CCLRC Daresbury Laboratory, Daresbury,
Warrington WA4 4AD, Cheshire, UK

February 2005

• ABSTRACT
• 1. INTRODUCTION
• 2. The SPEC BENCHMARKS
• 2.1 SPEC CPU2000 Benchmarks
• 2.2 SPEC CPU95 Benchmarks
• 3. The SERIAL Chemistry Benchmarks
• 3.1 The MATRIX-89 and MATRIX-97 Benchmarks
• 3.1.1 Whetstone Benchmark
• 3.1.2 Sparse Matrix Multiply Benchmark
• 3.1.3 Matrix Multiply Benchmark II - Similarity Transformations
• 3.1.4 Diagonalization Benchmark
• 3.1.5 Relative Performance on Matrix Operations
• 3.2 COMPUTATIONAL CHEMISTRY KERNELS
• 3.2.1 The STREAM Benchmark
• 3.3 THE CHEMISTRY APPLICATIONS BENCHMARK
• 3.3.1 The Quantum Chemistry Benchmark - GAMESS-UK-89
• 3.3.1.1 GAMESS-UK-97
• 3.3.2 The Molecular Simulation Benchmark - DL_POLY
• 3.4 SUMMARY of the SERIAL Benchmarks
• 4. The Chemistry RATE Benchmarks
• 4.1 The Individual ChemRATE Benchmarks
• 4.2 The Overall ChemRATE Benchmark
• 4.3 SUMMARY of the ChemRATE Benchmark
• 5. CONCLUSIONS
• References
• Table 1. SPEC CPU2000 - SPECfp and SPECint. Absolute and Relative Values.
• Table 2. SPEC CPU2000 - SPECfp_base and SPECint. Absolute and Relative Values.
• Table 3. SPEC CPU2000 - SPECfp_rate and SPECint_rate. Absolute and Relative Values .
• Table 4. Vector Whetstone-97 Benchmark.
• Table 5. The Matrix-97 Benchmark - Sparse MMO.
• Table 6. The Matrix-97 Benchmark - Similarity Transformation.
• Table 7. The Matrix-97 Benchmark - Diagonalization.
• Table 8. The Matrix-97 Benchmark: Performance Relative to the HP RX5670 Itanium2/1500-H.
• Table 9. The Matrix-89 Benchmark: Performance Relative to the HP RX5670 Itanium2/1500-H.
• Table 10. The Computational Chemistry Kernels Benchmark.
• Table 11. The STREAM Benchmark. TRIAD rates (MBytes/sec).
• Table 12. The GAMESS-UK-89 Benchmark.
• Table 13. The GAMESS-UK-97 Benchmark.
• Table 14. The DL_POLY Benchmark.
• Table 15. Summary of the Serial Chemistry Benchmarks.
• Table 16. The Matrix Multiplication RATE Benchmark 1. FORTRAN MMO.
• Table 17. The Matrix Multiplication RATE Benchmark 2. MXMB0.
• Table 18. The Matrix Multiplication RATE Benchmark 3. DGEMM.
• Table 19. The Matrix Multiplication RATE Benchmark 4. QHQ (Vector Algorithm).
• Table 20. The Matrix Multiplication RATE Benchmark 5. QHQ (Scalar Algorithm).
• Table 21. The Matrix Diagonalisation RATE Benchmark.
• Table 22. The DLPOLY RATE Benchmark.
• Table 23. The GAMESS-UK RATE Benchmark.
• Table 24. The Chemistry RATE Benchmark. Overall RATE Estimates.
• Table 25. APPENDIX: Machine Configurations under Evaluation.



ABSTRACT

This report compares the performance of a number of different computer systems using a variety of software from the discipline of computational chemistry. The software includes matrix operations, a variety of chemistry kernels from quantum chemistry and molecular dynamics and a set of twelve quantum chemistry (QC) calculations and six molecular dynamics (MD) simulations. The QC calculations have been carried out using the GAMESS-UK electronic structure code, the MD calculations using the DL_POLY molecular dynamics program.

The comparison involves approximately one hundred and fifty computers, ranging from vector supercomputers such as the NEC SX-6i to scientific workstations from HP/Compaq, IBM, Silicon Graphics and SUN, and both 32- and 64-bit CPUs from Intel and AMD.

1. INTRODUCTION

This report presents a performance evaluation through benchmarking of a number of different computer systems specifically in the area of computational chemistry. This work has been ongoing since 1988, when the intention was to include representative hardware typifying supercomputers, superminis, workstations and the emerging class of parallel, novel architecture machines. Throughout the 1980's the cost effective debate in scientific computing centred on the relative merits of conventional vector supercomputers [1] and so-called superminis, machines costing some 10% of supercomputers, and exhibiting some 10% of supercomputer performance [1]. Over the past 8 years the supermini category of computational resource has to a large extent disappeared from both the vocabulary and offerings of most hardware vendors, to be replaced by high-end workstation servers. In contrast to earlier reports in this series, we have now discarded those machines that originally belonged in the supermini bracket (e.g., the Convex C-220 and Convex C-3860, the FPS M64/60 [1] and Alliant FX2808).

Vector supercomputers used in this report include the NEC SX-5 and SX-4. A large number of workstations and workstation servers have been evaluated, including systems featuring proprietary CPUs from HP/Compaq, IBM, Silicon Graphics and SUN (see below). During the 1990s, workstations based on proprietary RISC CPUs repeatedly demonstrated their cost-effectiveness against vector supercomputers in many scientific applications. Throughout this time, the performance of the family of Intel x86 processors became increasingly competitive, so much so that recent reports in this series have witnessed the x86 CPUs challenging the leading proprietary CPUs from IBM, SGI, Sun, Compaq and HP.

A wide variety of both Intel and AMD commodity CPUs feature in our benchmarking. Historically, these have evolved from Intel's Pentium-based CPUs, with the Pentium Pro (200 MHz), Pentium MMX (233 MHz), the Pentium II (266 - 450 MHz), and Pentium III CPUs (500 - 1 GHz). The Pentium III has now of course been surpassed by its successor, the Pentium 4 - systems benchmarked include those based on the dual processor Xeon CPUs at clock speeds increasing from 2.0 to 3.2 GHz. In addition to Intel's 32-bit IA32 products, we have also benchmarked a number of 32-bit CPUs from AMD, comprising the MP range of dual processor products (with the MP2400+ (2.0 GHz), MP2000+ (1667 MHz) and MP1800+ (1533 MHz) CPUs). Older AMD systems feature the K7 Athlon CPUs (500 - 1400 MHz).

Much has been made of the potential of the new generation of commodity 64-bit processors in scientific and technical computing. Intel led the way with their 64-bit IA64 processor, but have recently been challenged with the emergence of AMD's 64-bit Opteron processor. In response to this challenge Intel have now released their own x86-64 based products, the so-called "Nocona" or EM64T CPUs, with clock speeds increasing from 3.2GHz - 3.6 GHz (the HP DL380 dual-CPU Xeon EM64T running RedHat AS 3.0). All three 64-bit products now feature in these benchmarks.

• The first of Intel's IPF products, the Itanium 1 CPU, clearly under delivered; we include here benchmarks based on both the 733 and 800 MHz CPUs. The former appeared in an RX4610 server from Hewlett Packard, a 4-way system with 733 MHz CPUs, featuring 2 MB L3 cache and running the HPUX OS. The 800 MHz Itanium CPU with 4 MB L3 cache also appeared in a 4-way system, running the Linux OS with Intel's 64-bit compilers. The release in 2002 of the second generation Itanium processor from Intel, the Itanium 2 (or McKinley) processor addressed many of the shortcomings evident in its predecessor. Included here are benchmarks from a number of systems featuring the processor, including those from HP (running both the HPUX and Linux operating system), plus a system with the so-called Tiger board from Intel. Systems from HP include the 4-way RX5670 and 2-way RX2600, both featuring 1 GHz CPUs and 4 MByte of L3 cache; the RX5670 is Linux-based, the RX2600 HPUX-based. The corresponding 900 MHz single-CPU Linux system is the ZX6000, with 1.5 MByte L3 cache, running the Linux OS. The system featuring the Intel Tiger board is a 4-way, 1 GHz machine (4 MB L3 cache) running the Linux OS.
  Further improvements to the IPF product line arrived in 2003 with the Madison processor. We again include benchmarks from a number of systems featuring this processor, including those from HP (running both HPUX and Linux), from SGI (the Itanium2, Linux64-based Altix systems) and the Intel-based Tiger systems. Systems from HP include both 2- and 4-way systems with a variety of Itanium2 CPUs - 1.2, 1.3, 1.5 and 1.6 GHz. Systems based on the 1.6 GHz Itanium2 include the 2-way HP RX1620 (running both HPUX and Linux64) and the 4-way HP RX4640, running Linux64. The latter system features a 9 MByte L3 cache, while the RX1620 systems feature a 3 MByte L3 cache. The HP RX1620 is enhanced by a FSB clocked at 533 MHz.

  1.5 GHz Itanium2 CPU-based systems include the 4-way RX5670 and 2-way RX2600, (with 6 MByte L3 cache), the 4-way RX4640 with 1.3 GHz CPU, plus the 1 GHz CPU-based RX5670 and RX2600 systems (with 4 MByte of L3). The RX5670 and RX4640 are both HPUX-based, while the RX2600 is Linux-based. The SGI system evaluated is the SGI Altix 3700 with both 1.3 and 1.5 GHz CPUs. Intel-based 4-way Tiger solutions include those with both 1.5 and 1.2 GHz CPUs, the latter a prototype.

• The 64-bit AMD Opteron systems benchmarked include those on 1.4, 1.6. 1,8, 2.0, 2.2 and 2.4 GHz CPUs in both 2-way and 4-way systems. Two way systems include the Opteron244 (1.8 GHz) and Opteron248 (2.2 GHz), four-way systems the Opteron840 (1.4 GHz), 842 (1.6 GHz), 846 (2.0 GHz), 848 (2.2 GHz) and the Opteron850 (2.4 GHz). These systems are from a variety of vendors, including the HP DL145 dual-CPU 2.2GHz (RedHat AS 3.0), and the V20 and V40z Sun Fire servers (2.2 GHz).

A large number of workstations and workstation servers based on proprietary CPUs have been benchmarked over the past decade. The systems from HP/Compaq, IBM, Silicon Graphics and SUN are listed below.

• HP/Compaq/DEC. Several generations of machines featuring the Alpha processor have been evaluated, the most recent including the HP AlphaServer ES45, with both 1250 MHz and 1000 MHz A21264C EV68 CPUs (with 16 MByte L2 cache in the former and 8 MByte L2 cache in the latter). Perhaps a more interesting system was the AlphaServer SC based on EV7 technology. Code-named "Marvel", the 1 GHz processor promised significantly enhanced memory bandwidth compared to the EV68. EV67 processors clocked at 833 MHz feature in the AlphaServer ES40 (again with 8 MByte L2 cache) and in the Linux-based UP2000 dual-processor CPU from API (in the present case with 4 MByte L2 DDR cache). The slower 667 MHz EV67 featured in the API UP2000 and in the Compaq PW XP1000, the DS20E AlphaServer, and the ES40 AlphaServer (with 8 MByte L2 cache). Systems housing the A21264 EV6 CPU include the DEC Alpha 8400/6-575, the entry-level 466 MHz Compaq AlphaServer DS10, the 500 MHz A21264-based machines (the single CPU XP1000 personal workstation, the dual processor Compaq DS20/6-500 and quad processor ES40/6-500 AlphaServer), plus the Compaq GS140 with 525 MHz A21264. Note that the DS10 features a 2 MByte L2 cache, in contrast to the 4 MByte cache of the XP1000, DS20/6-500 and ES40/6-500. Also included is the Samsung AlphaPC 264DP-500, which with 2 MByte L2 cache runs Linux, in contrast to the Tru64 OS on the other EV6-based machines. Retired Alpha systems include the AXP Alpha EV4-based 3000 series and EV5-based 500, 600, 2100, PW AU and 8400 series.
• Hewlett Packard, with the most recent PA-RISC based system evaluated being the RP3440 dual-CPU + dual-core PA8800 clocked at 1.0GHz (HP-UX 11.11). Other systems include the Model RP7410 featuring the 850 MHz PA8700+, PA8600 and PA8700 CPUs feature in the HP9000/J6000 (552MHz) and HP9000/J6700 (750MHz) respectively. Older PA8500-systems include the HP9000/785 Model 360 (367MHz), the HP9000/C3000 (400MHz), HP9000/J5000 and 9000/N4000 (440MHz) models. Retired systems include the HP model 9000 series with the PA8000 and PA8200 CPUs, the former in the C160 (160MHz), the latter in the 200MHz HP9000/V2200 (the Exemplar-V class), the HP9000/C200 and in the 236 MHz HP9000/C240 and HP9000/V2250.
• IBM, with a wide range of RS6000-based models centred on Power CPU technology. The latest addition to the power series is the power-5 CPU, featured here in the IBM eServer p5 570 (1900 MHz). Power-4+ and power-4 CPUs include the 1.7 GHz power4+ CPU in the 32-way pSeries 690Turbo, and the 1.3 GHz power-4 CPU based-systems, the 8-way IBM pSeries 690Turbo and the 32-way Regatta H node. Benchmark results from both machine are included in this report, together with those from the 16-way Regatta HPC node (in which one of the dual core CPUs is disabled), and the 1 GHz power4 CPU in the 4-way IBM eServer pSeries 630/6C4. Older Power3 RS6000-based models include the 4-way SP Winterhawk2 (WH2) node and IBM RS/6000 44P-270, both clocked at 375 MHz, the 8-way Nighthawk node, clocked at 222 MHz, and the dual processor IBM RS/6000 43P Model 260 (200 MHz). Retired systems evaluated include those based on Power, Power2 and Power2 Super chip (P2SC) CPUs.
• Silicon Graphics, with desktop and server machines featuring the MIPS range of processors. The most recent R14k-based machines evaluated include the 600 MHz SGI Onyx 300, and the 500 MHz Origin 3800 and dual processor Origin 300. R12k-based machines include the Origin 3800 (400 MHz), Origin 2000 (300 and 400 MHz), the 270 MHz Octane R12k and O2 R12k, and the 400 MHz R12k Octane 2. Systems featuring the R10k include the Power Challenge, Origin200, O2000, Octane and the ill-fated O2; the R10k CPUs include those rated at 175, 195, 225 and 250 MHz. R5k-based machines include the Indy and O2. Retired machines with the R3000-, R4000-, and R8k-based systems include the Challenge, Power Challenge (PChall.), Indy, Indigo and Indigo2, Crimson, and 4D series.
• SUN, with systems based on the UltraSPARC family of processors. The latest UltraSPARC-III Cu-based systems evaluated include the dual 1.056 GHz Sun Blade 2000 and the 8-way Sun Fire V880 with 900 MHz CPUs. Other UltraSPARC-III systems benchmarked include the Sun Blade 1000 Model 1750, with 750 MHz CPU, and SUN Fire 6800/900-Cu. All of these systems feature 8 MByte L2 cache. UltraSPARC-IIi based systems feature the HPC4500 with 400 MHz CPU and Ultra80 (with 450 MHz CPU), both with 4 MByte L2 caches. Retired systems benchmarked include the SPARC-5 and SPARC-10, the HyperSPARC, Ultra-SPARC-1 (Ultra-1/140 and 170) and Ultra-SPARC-2 processors (Ultra-2/200, Ultra-2/300, Ultra30/300 and Enterprise HPC4500, 336MHz).

While now somewhat dated, we also include the evaluation of a number of individual nodes of a variety of Massively Parallel Processors (MPP), including the CRAY T3D, T3E/900 and T3E/1200, the IBM SP2 with both power2-based TN2 and P2SC (120 and 160 MHz) and power3 Nighthawk (222 MHz) and Winterhawk2 (WH2) (375 MHz) CPUs, and the Hitachi SR2201. Figures for the Intel iPSC/860, the KSR-2 from Kendall Square Research, and transputer and i860-based Meiko Computing Surfaces may be found in reference 3.

Machines featuring in this exercise, together with associated configurations, are given in the Appendix. We should stress from the outset that our access to much of the hardware evaluated herein has been at best short lived, and has often involved the temporary loan or donation of machines as part of one of the hardware evaluation exercises run at the Daresbury Laboratory. In many cases these machines were not optimally configured in terms of either memory, or high speed disk, and consideration of the results presented here should be viewed in that light.

Following an introductory evaluation of hardware based on the SPEC Benchmarks (Section 2), we present in Section 3 a number of serial, single CPU benchmarks using a variety of chemistry-oriented software. This 64-bit floating point precision FORTRAN-based code may be classified into three distinct categories - Matrix, chemistry kernels and chemistry applications - each is designed to provide a pointer to the relative hardware capabilities in the discipline of computational chemistry. We then address the issue raised above - that of appropriate RATE benchmarks, and describe in section 4 a series of benchmarks designed to provide a so-called ChemRATE metric, based on extended versions of serial benchmarks used in Section 3.

2. The SPEC BENCHMARKS

One of the most useful indicator of CPU performance is provided by the SPEC (``Standard Performance Evaluation Corporation'') benchmarks. This benchmark suite contains non-tuned application-based code to measure processor speed for both integer (SPECint) and floating point (SPECfp) arithmetic. While earlier versions of the suite (e.g. SPECmark89) had certain well-advertized flaws, SPECfp95 and SPECint95, and their successors, SPECfp2000 and SPECint2000, have become industry standards in measuring primarily the performance of a system's processor, memory architecture, operating system and compiler.

A SPEC FAQ describing the SPEC benchmark suite and the SPEC consortium is periodically posted to comp.benchmarks, and can be found on the WWW at

http://www.specbench.org/spec/faq

An excellent summary of the SPEC benchmarks that was periodically updated is available via anonymous ftp from ftp.cs.toronto.edu in the file /pub/spectable. This posting terminated in December 2000 with the arrival of SPEC2000. More SPEC-related information is available at the SPEC WWW site,

http://www.specbench.org

and at the Performance Database Web site,

http://performance.netlib.org/performance/html/spec.html#specsite.

2.1 SPEC CPU2000 Benchmarks

The current generation of SPEC benchmarks, SPEC CPU2000,

http://www.specbench.org/osg/cpu2000/results/cpu2000.html

comprises a number of separate listings including CFP2000 and CINT2000. The former is derived from the results of fourteen floating-point benchmarks compiled with aggressive optimization. It is the geometric mean of fourteen normalized ratios (one for each floating-point benchmark). CINT2000 is derived from the results of twelve integer benchmarks compiled with aggressive optimization. It is the geometric mean of twelve normalized ratios (one for each integer benchmark). Note that the level of optimization is not mandated. While highly aggressive optimization is permitted, results derived from benchmarks compiled with conservative optimization (as in SPECfp_base2000) can be submitted.

It is perhaps worthwhile at the outset to consider the potential shortcomings of using single processor benchmarks, such as SPECfp2000 or SPECfp2000_base, on the current generation of SMP-based clusters. The nature of the associated cache hierarchy and memory architecture need to be considered, both arguing for a consideration of not only the SPECfp values, but also the related RATE benchmarks (SPECfp_rate). The complex cache hierarchy of systems such as the 8-way IBM 690Turbo means in practice that with 7 of the 8 CPUs disabled (according to the SPECfp2000 rules), a given single processor benchmark is actually running in an environment comprising the total L3 cache associated with the 8-way MCM i.e. 128 MByte. Since many of the SPECfp2000 benchmarks have a total memory requirement of this order of magnitude, the recorded level of performance arguably bears little resemblance to what might be seen if, for example, all 8 CPUS were running the same job. A second effect is that of the memory architecture of the SMP node in question; in many instances the effective memory bandwidth available to a given job will decline significantly as more of the CPUs are involved in processing. The impact of this effect will be critically dependent on the nature and bandwidth demands exhibited by the application. The impact of L3 cache utilisation and memory architecture may be partially quantified by considering the associated SPECfp2000_rate figures for such systems. Both effects should be considered when considering all the single CPU results presented in this paper.
The current set of CPU2000 results are given in Tables 1 (SPECfp2000) and 2 (SPECfp_base2000) (with the baseline system the Ultra 10 333MHz). In each case we have normalised the values relative to those of the 1.5GHz Itanium2 CPU in the HP Integrity RX5670 (both SPECfp2000 and SPECfp_base2000 values of 2108).

An examination of the SPECfp2000 values of Table 1 demonstrates the dominant position of the Intel's 2nd generation Itanium2 processor (Madison) with 20 of the leading 30 systems featuring this processor. The only other CPU to feature in the top 30 systems is IBM's power5 processor (the 1.9 GHz CPU in the eServer p5 570, the 1.65 GHz CPU in the i5/p5 550 and 520 eServer systems, and the 1.5 GHz CPU in the p5 520 Express). Considering the leading systems of Table 1, we note:

1. Within the leading 30 systems, SPECfp2000 ratings for the Itanium2 processors may be broadly classied according to both clock speed and size of L3 cache. For the 1.6 GHz CPU, we find ratings of between 2712 (HP RX4640 with 9 MB L3 cache) to 2545 (Dell PowerEdge 7250 with 3MB cache), although we note the rather poor rating (2001) reported for the I2X2 system from ION Computers. It is noticeable that the size of cache (and associated cost) does not have a major impact on the SPECfp ratings e.g. the SGI Altix 3700 Bx2 ratings of 2647 and 2600 for systems with 9MB and 6MB of L3 cache respectively. Systems based on the 1.5 GHz processor exhibit SPECfp2000 ratings from 2502 (HP Integrity RX4640 with 4MB L3) to as low as 1830 (Supermicro SS6113M-i with 6MB L3). The latter is little better than the value of 1817 (in the 1.4 GHz-based I2X2 system from ION Computers) and 1814 (the 1.3 GHz Integrity RX5670 system from HP). These may be contrasted with figures for the older Itanium2 systems ranging from 1410 (the 1GHz CPU in the SGI Altix 3000) to 1086 (the 900 MHz CPU in the HP ZX2000 system).
2. Historically two sets of values were typically given for the Itanium2-based systems from HP - corresponding to execution under HPUX using HP's f90 compiler (tagged "H" in the tables), and to execution under Linux running Intel's efc/ifort compiler ("L"). While the HPUX-based figures were initially higher than those based on ifort, the improvements from version 7.0 onwards of the Intel compiler have reversed this trend. We would expect this situation to remain fluid as new releases of both compilers are made available. It is noticeable that the vast majority of the efc-based figures are all SPECfp_base figures; few optimised SPECfp2000 ratings have been published to date.
3. With a SPECfp rating of 2702, the 1.90 GHz power5-based IBM eServer p5 570 outperforms the 1.65 GHz power5 CPU by factors of between 1.13 (the eServer p5 590) and 1.28 (the p5 520 Express). These factors increase to 1.59 and 2.13 respectively when compared to the older 1.7 GHz power4+ and 1.3 GHz power4 CPUs, both in the pSeries 690Turbo series.
4. Considering the Pentium 4 and associated Xeon systems, we find ratings as a function of clock speed of 1842 (3.80 GHz Intel D925XECV2), 1724 (Intel D925XECV2 3.46 GHz P4), 1631 (Supermicro Super P8SC8 3.80 GHz P4) and 1630 (Intel D925XCV, AA-400 3.6GHz P4) and 1597 (HP ProLiant BL20p G3 3.6GHz Xeon). The 3.80 GHz D925XECV2 is seen to exhibit a rating close to that of the 1.3GHz Itanium2 in the SGI Altix 3000 series (1854).
5. The leading x86-64 system from AMD is the 2.6GHz Athlon64 (MSI K8N Neo2 FX-55) which, with a rating of 1878, is slighly ahead of the 3.80 GHz Intel D925XECV2 (1842) and of comparable speed to the Itanium2 1.5GHz/6ML3 in the Dell PowerEdge3250 (1875). Other x86-64 ratings as a function of clock speed include the 2.6GHz FX-55 (1782, AMD MSI K8N Neo2) and ratings for the 2.4GHz Opteron systems ranging from 1644-1849 (1849, Fujitsu CELSIUS V810 / Opteron250; 1787, Sun Java Wrkst. W1100z; 1771, IBM eServer 326; 1701, HP ProLiant DL145 Opteron250; 1693, Sun Fire V20z Opteron250, and 1644, AMD ASUS SK8V/Opteron150)
.
6. While the leading UltraSPARC-based systems from Sun are now well off the leading pack (see below), the SPARC64-based systems from Fujitsu remain competitive, with the 1890 MHz CPU in the PrimePower900 series exhibiting a rating of 1803.
7. Considering the final variant of the Alpha processor, the EV7, we find ratings of 1684 in the 1.3 GHz-based AlphaServer GS1280, 1494 in the 1.1GHz-based ES80 and ES47, and 1288 in the 1 GHz ES47. Ratings for the older EV68 processor include values of 1365 for the AlphaServer ES45 Model 68/1250 (A21264C) - this CPU outperforms the DS25 68/1000 (985) and ES45/1000 (960) by factors of 1.35 and 1.39. These factors are greater than the clock speed ratio of 1.25, caused in large part by the increased cache on the 68/1250 (16 vs. 8 MByte).

While the top 50 systems exhibit greater than 63% of the performance of the 1.6 GHz-based HP Integrity RX4640 (i.e. SPECfp2000 figures higher than 1700), only 6 distinct CPUs are involved in these systems, namely IBM's power5, Intel's IA64 Itanium2 and IA32/EM64T Pentium 4 Xeon processors, AMD's Athlon64 and Opteron CPUs, and finally the SPARC64 GP from Fujitsu.

Forty of the next 50 entries (SPECfp ratings from 1441-1659) feature either the Pentium 4 / Xeon IA32 CPUs or x86-64 systems from AMD, with 30 systems from Intel and 10 from AMD. The remaining 10 systems include four pwr4+ systems from IBM, four EV7 systems, and two of the slower Itanium2 systems (1.4 GHz with 1.5 MB L3 cache).

All other CPUs exhibit SPECfp2000 ratings of less than 1440 i.e. are a at least a factor 1.88 times slower than the 1.6GHz/9MB Itanium2 in the HP Integrity RX4640. Other figures worth noting on the leading machines from other vendors include:
• 1353 and 1349 for the 1600 MHz UltraSPARC IIIi CPU in the Sun Blade Model 2500 and Sun Fire V440 respectively. Based on the former figure, the 1600 MHz CPU is outperformed by the HP RX4640 Itanium2 by a factor of two.
• An estimated value of 1000 for the 1 GHz dual-core PA8800 CPU in the HP 9000 Model RP3440 and 674 for the 875 MHz PA8700+ in Hewlett Packards HP 9000 Model C3750. The HP 9000 Model RP3440 would thus be ranked at around position 250 in the SPECfp2000 list. Note that the PA8800 is essentially identical to the other PA2.0 CPUs - with higher frequency, more L2 cache, dual-core and uses the same bus the Itanium2 uses (200MHz@128bit double pumped). Thus in most cases the PA8800 is performing roughly like a Itanium2 1.0GHz system, not unexpected given use of the identical memory bus, virtually the same number of functional units and the same frequency (considering the old 1.0GHz Itanium2).
• 982 for the x86-32 2.20 GHz CPU from AMD (ASUS A7N8X Motherboard in the XP 3200+) and 659 for the dual processor 2.00 GHz CPU in the MP2400+).
• 529 for the MIPS R14k/600 MHz CPU and 463 for the R14k/500 MHz CPU, both in SGI's Origin 3200. The SGI Origin 3200 / 600MHz R14k is ranked at position 403 in the SPECfp2000 list.
• 456 and 440 for the 1.4 GHz and 1.26 GHz Pentium III respectively in Dell's Precision PowerEdge 1500SC WorkStation. The PentiumIII is now a factor of four slower than the 3.8 GHz Pentium 4.
• 434 for the 450 MHz RS/6000 power3 CPU in IBM's RS/6000 44P-170, and 382 for the 375 MHz CPU in the IBM RS/6000 SP-375MHz T/W.
Summarising the above discussion, normalised ratings for the leading proprietary and commodity-based CPUs from each vendor normalised with respect to the HP Integrity RX4640 1.6 GHz Itanium2 (9MB L3 cache) and the values of Table 1 are as follows:
1. 100% - HP Integrity RX4640/Itanium2 1.6 GHz/9MB (2712);
2. 100% - IBM eServer p5 570 (1900 MHz) (2702);
3. 98% - SGI Altix 3700 Bx2/Itanium2 1.6 GHz/9MB (2647);
4. 95% - Bull NovaScale 4040, 1.6 GHz Itanium2 with 9MB L3 cache (2566);
5. 94% - Dell PowerEdge 7250, 1.6 GHz Itanium2 with 9MB L3 cache (2545);
6. 75% - ION Computers I2X4, 1.5 GHz Itanium2 with 9MB L3 cache (2043);
7. 69% - AMD MSI K8N Neo2 motherboard with 2.6 GHz Athlon64 FX-55 (1878);
8. 69% - Intel D925XECV2 3.80 GHz Pentium4 (1873);
9. 66% - Fujitsu PrimePower900, 1.89 GHz SPARC64 V with 3MB L2 cache (1803);
10. 66% - SUN Java Workstation W1100z (2.4 GHz Opteron) (1787);
11. 62% - HP AlphaServer GS1280, 1.3 GHz EV7 (1684);
12. 50% - Sun Blade Model 2500, 1.6GHz UltraSPARC IIIi (1353);
13. 37% - HP 9000 Model RP3440, 1.0GHz PA8800 (1000); and,
14. 19% - SGI Origin 3200/R14k-600 (529) and 17% for the Origin 3200/R14k-500 (463).

In terms of leading proprietary CPUS from each vendor, it is clear that both MIPS-based systems from SGI and PA-RISC systems from HP, and, to a lesser extent, both Alpha EV7/EV68 systems from HP/Compaq and UltraSPARC systems from Sun, are now eclipsed by the commodity IA64 and x86-systems from Intel and AMD. Indeed, the main products now offered by all four vendors feature some combination of the latter CPUs.

Finally, we would point out a number of issues around the present benchmarks and their potential relationship to the SPEC benchmarks:

• Some issues need to be considered when comparing the SPECfp2000 ratings with the results of the present benchmarking exercise. Many of the published figures rely on utilities that were not available in the current benchmarking exercise e.g., "spike", and it may be more meaningful in such cases to draw comparisons with the corresponding SPECfp_base values.
• When considering the present benchmarking results, there are several factors we wish to consider in assessing the usefulness of the SPEC ratings;
  • Do the SPECfp values provide a reliable metric for evaluating the capabilities of hardware in computational chemistry? If so, we would expect to find a close mapping of the ratios for the various chemistry benchmarks onto the SPECfp ratios;
  • Does any particular CPU consistently ``underperform'' based on the SPECfp criteria? - this would manifest itself as the ratios from the chemistry benchmarks falling below the SPECfp ratios e.g., the impact of the memory problems of the SGI O2-R12k and R10k [6] impacting on the benchmarks.
  • As discussed above, we should consider the potential shortcomings of using single processor benchmarks such as SPECfp2000 on the current generation of SMP hardware. The only way to quantify this effect is to generate the associated "rate" figures for the machine in question. We note at this stage that the current set of SPECfp2000_rate figures are given in Table 3; these will be discussed later in this document.

We will attempt to address these issues below.

2.2 SPEC CPU95 Benchmarks

With an increasing number of vendors publishing CPU2000 results for new systems, we clearly had a problem in terms of relating older machines, for which only SPECfp95 results are available, with these newer systems. In practice we have discontinued our analysis based on the older CPU95 metric, but have attempted to generate extrapolated CPU2000 values for these older machines based on a comparison of those processor families where both CPU95 and CPU2000 values were available.

We note for completeness that SPECfp95 was derived from the results of ten floating-point benchmarks compiled with aggressive optimization. It is the geometric mean of ten normalized ratios (one for each floating-point benchmark). SPECint95 was derived from the results of eight integer benchmarks compiled with aggressive optimization. It is the geometric mean of eight normalized ratios (one for each integer benchmark) Note that the level of optimization is not mandated. While highly aggressive optimization was permitted, results derived from benchmarks compiled with conservative optimization (as in SPECfp_base) were often submitted.

SPECfp95 and SPECint95 results for many of the CPUs discussed in this paper are available in previous versions of thie report.

3. The SERIAL Chemistry Benchmarks

We first focus on a number of serial benchmarks using a variety of chemistry-oriented software. This 64-bit floating point precision FORTRAN-based code may be classified into three distinct categories, each designed to provide a pointer to the relative hardware capabilities in the discipline of computational chemistry;

1. The first category (the MATRIX Benchmark, section 3.1) reflects the dependence of many of the algorithms in the area of electronic structure calculations on matrix operations, and includes both matrix multiplication and matrix diagonalization. Note that the MATRIX Benchmark has now been extended to include the treatment of larger matrices, designed to exercise both memory bandwidth and cache more rigorously and, in common with many other standard benchmarks, to reflect the limited run times of the original benchmark (hereafter referred to as MATRIX-89) whereby modern CPU's completed the tasks in just a few seconds. We refer to this extended MATRIX Benchmark as MATRIX-97.
2. The second category (the Computational Chemistry Kernels, section 4) includes four chemistry `kernels', each comprising less than a 1000 lines of FORTRAN code, and intended to be representative of the typical calculations undertaken in the area of computational chemistry. Described in more detail in section 3.2, these kernels include direct-SCF, molecular dynamics (MD), quantum monte carlo (QMC), and a Jacobi eigen solver (JACOBI);
3. Finally, we include complete quantum chemistry and molecular simulation applications (section 3.3). The QC calculations use the GAMESS-UK electronic structure program [4]. Twelve typical applications are included, featuring both conventional Hartree Fock self-consistent field (SCF) and direct SCF, complete active Space SCF (CASSCF) and multiconfiguration SCF (MCSCF), configuration interaction calculations, both direct-CI and conventional table-driven MRD-CI, Moller Plesset perturbation theory (MP2), and both SCF and MP2 analytic 2nd derivatives. In the same way that the MATRIX benchmarks have been extended to include the treatment of larger matrices, we have recently extended the GAMESS-UK benchmark to handle larger more representative calculations. We refer to this extended GAMESS-UK benchmark as GAMESS-UK-97 (see section 3.3.1.1 below), and to the original benchmark as GAMESS-UK-89.

   The MD calculations described in section 3.3.2 use the DL_POLY simulation package [5] with six typical MD calculations.

3.1 The MATRIX-89 and MATRIX-97 Benchmarks

Previous versions of this report presented timings from both the older MATRIX-89 and extended MATRIX-97 Benchmarks. We now present only the detailed timings from the latter (see Tables 4-8) given the limited run times of the original benchmark.

3.1.1 Whetstone Benchmark

A comparison of the single processor Whetstone performance on a variety of machines, including workstations, both IA32 and IA64 commodity CPUs, vector CPUs and a variety of MPP processors is given in Table 4 (MATRIX-97). Data provided includes the Mflop performance on a variety of floating point vector loops (V_L =1024), together with the total CPU time to execute the benchmark, and the MWIPS performance. The primary aim of this benchmark is to provide a performance measure of both floating point (FP) and integer arithmetic; thus while trends in the V_L Mflop ratings are of interest, only a small part of the total CPU time is actually involved in these operations. The wide variety of standard functions exercised (abs, sqrt, exp, alog, sin, cos, atan etc.) consume a far larger fraction of the reported times; note the latter provide a close mapping onto the measured rate of instruction processing (MWIPS, million whetstones instructions per second). Note also that the MATRIX-97 benchmark was generated from the original MATRIX-89 code by increasing the outer loop variable, i2, by an order of magnitude.

3.1.2 Sparse Matrix Multiply Benchmark

The matrix multiply operation (MMO) is central to the efficient operation of modern QC codes on vector processors [1], it being possible both to extract near peak performance for this kernel and to formulate many QC steps around this operation. A comparison of the single processor sparse MMO performance on a variety of machines, is given in Table 5. In this benchmark a series of MMOs (R = A X B) involving matrices of increasing order were performed, with each MMO conducted a number of times, this number being inversely proportional to the order of the matrices,

In the MATRIX-89 benchmark, the matrices were of order 10, 20, 30, ...,100, with the summed CPU time metric referring to 100 MMOs of order 10 matrices, 90 of order 20 matrices, and so on up to 10 MMOs for matrices of order 100. The order of the matrices involved in the MATRIX-97 benchmark were increased by a factor of five compared to MATRIX-89, with the CPU times of Table 5 referring to 10 MMOs of order 50, 9 MMOs of order 100 etc., up to a single MMO for matrices of order 500. Figures are presented for both `full' (0% sparse) and 50%-sparse B matrices, with the performance figures referring to code written entirely in Fortran.

3.1.3 Matrix Multiply Benchmark II - Similarity Transformations

The second matrix multiplication benchmark is designed in part to show the importance of well-optimized library software in the goal of cost-effective performance. This benchmark involves performing a series of similarity transforms (Q*HQ) using both a scalar and vector algorithm. The scalar code collapses the matrix transposition and multiplications to yield an algorithm with fewer FLOPS than the vector code, which adopts a brute force approach by explicitly performing the transposition and two matrix multiplications. In the latter case we utilize the BLAS library routine DGEMM (where available) for performing the requisite MMOs, in the former case the dot product BLAS routine, DDOT. In the MATRIX-89 benchmark, the matrices were of order 10, 20, 30, ...,150, the summed CPU time metric referring to 150 transforms of order 10 matrices, 140 of order 20 matrices, and so on up to 10 transforms for matrices of order 150. The order of the matrices involved in the MATRIX-97 benchmark were doubled compared to MATRIX-89, with the CPU times of Table 6 referring to 300 transforms of order 20, 280 transforms of order 40 etc., up to 20 transforms for matrices or order 300. Again, given the small execution times of the MATRIX-89 benchmark, we focus the discussion below on the MATRIX-97 results.

3.1.4 Diagonalization Benchmark

Tables 10 and 11 present the results of a matrix diagonalization benchmark intended to supplement the previous analysis conducted by Dunning and co-workers [7]. We consider a similar benchmark, based on diagonalizing a series of real symmetric matrices, using 64-bit floating point arithmetic. Again the CPU time was measured for the diagonalization of each size matrix, with the summed times used as the benchmark execution time. Results include the compiler options used on the depicted hardware.

In the MATRIX-89 benchmark, the matrices were of order 10, 20, 30,..., 100, the summed CPU times of Table 10 referring to 100 diagonalizations of order 10 matrices, 90 of order 20 matrices, and so on up to 10 diagonalizations for matrices of order 100. The order of the matrices involved in the MATRIX-97 benchmark were increased compared to MATRIX-89, with the CPU times of Table 11 referring to three diagonalisations for matrices of order 50, 100, 150, 200, 250 and 300.

While the previous analysis was restricted to the EISPACK RS routine, we consider below the performance of eight diagonalization routines available in various mathematical libraries and quantum chemistry codes:

1. EIGRS, an unoptimized FORTRAN version of the library routine RS (available in SCILIB on the CRAY) from the IMSL library of routines [8].
2. F02ABF from the NAG library [9].
3. HQRII [10], as implemented in the semiempirical MOPAC program [11].
4. GIVENS, adapted from the QCPE program exchange (number 62.1).
5. SDIAG2, as implemented in the MUNICH system of programs.
6. JACOBI, from the ATMOL system of programs [12].
7. JACO, the diagonalization routine from the direct-SCF program DISCO [13].
8. ERDUW, as taken from the Berkeley System of Quantum Chemistry codes.

The first four routines are all based on the Householder QR method, whilst the last four use the Jacobi method. Note that the only optimization performed involved inserting calls to the BLAS for two-dimensional rotations (DROT) and vector interchange (DSWAP).

3.1.5 Relative Performance on Matrix Operations

To summarize the performance of the various workstations on the matrix multiply and diagonalization benchmarks detailed above, we show in Table 12 (MATRIX-89) and Table 13 (MATRIX-97) the performance of each relative to the Compaq Alpha ES40/6-833 (100%). The "Total" is derived by simply averaging the individual figures for the MMO, (Q*HQ) and Diagonalisation benchmarks.

3.2. COMPUTATIONAL CHEMISTRY KERNELS

One of the crucial requirements in evaluating the increasingly broad range of hardware platforms, whether these be parallel machines (true MIMD message-passing machines, workstation clusters, shared-memory multiprocessors etc), or simply the lastest workstation, is the availability of portable benchmarking codes that are representative of the application area under consideration. In an attempt to provide such capabilities in computational chemistry, we have described previously four representative codes, each of which is less than 1000 lines of FORTRAN, and is sufficiently portable that migration to any hardware platform can typically be achieved in a matter of hours. In this report we limit our discussion to the performance of these codes on a variety of single CPUs.

The benchmarking kernels comprise the following programs that are realistic models of actual chemical applications or algorithms;

1. SELF CONSISTENT FIELD (SCF);  This Self Consistent Field (SCF) electronic structure kernel uses distributed primitive 1s gaussian functions as a basis (thus emulating use of s,p,... functions) and computes integrals to essentially full accuracy. It is a direct SCF code, with an atomic density used for a starting guess. There are two available problem sizes, corresponding to 60 basis functions Be₄ and 240 basis functions (Be₁₆). The timings of Table 14 refer to the former.
2. MOLECULAR DYNAMICS (MD);  This program bounces a few thousand argon atoms around in a box with periodic boundary conditions. Pairwise interactions (Leonard-Jones) are used with a simple integration of the Newtonian equations of motion.
3. MONTE CARLO (QMC);  This code evaluates the energy of the simplest explicitly correlated electronic wavefunction for the He atom ground state using a variational monte-carlo method without importance sampling.
4. JACOBI ITERATIVE LINEAR EQUATION SOLVER (JACOBI);  Uses a naive jacobi iterative algorithm to solve a linear equation. All the time is spent in a large matrix vector product.

Total CPU timings for the SCF, MD and QMC benchmarks are presented in Table 14, together with the Mflop ratings from the JACOBI benchmark. Historically, these results have presented a somewhat confusing picture, with the detailed processor ordering very much a function of the particular chemistry kernel under examination.

To summarize the performance of the various workstations on the computational chemistry kernels as a whole, we show in Table 14 an overall single figure for each machine under the heading "Relative Performance". This is derived by normalising each of the individual benchmarks (SCF, MD, QMC and Jacobi) to the performance observed on the HP RX5670 Itanium2/1500-H, and taking the average of the four figures, expressed as a percentage.

3.2.1 The STREAM Benchmark

To provide a more quantitative insight into the issue of memory bandwidth that characterized the behaviour Jacobi benchmark above, we have adopted the widely used STREAM benchmark due to John McCalpin, University of Virginia, Charlottesville, Virginia. STREAM is a simple synthetic benchmark program that measures sustainable memory bandwidth (in MB/s) and the corresponding computation rate for simple vector kernels. The benchmark is specifically designed to work with datasets much larger than the available cache on any given system, so that the results are (presumably) more indicative of the performance of very large, vector style applications.

For more details on STREAM, see: Sustainable Memory Bandwidth in Current High Performance Computers

A somewhat broader look on the issue is given in : Memory Bandwidth and Machine Balance in Current High Performance Computers, with a version of this paper published in the newsletter of the IEEE Technical Committee on Computer Architecture (TCCA) in December 1995.

A subset of the STREAM results, those for the TRIAD vector kernel (a(i) = b(i) + q*c(i)) for many of the CPUs discussed in this paper are given in Table 15. Note that for the older CPUs, we have taken values from the STREAM web site; for the more recent CPUs we have, where possible, repeated the runs to confirm the reported values.

The TRIAD values of Table 15 have been normalized to the measured value on the HP RX5670 Itanium2/1500-H. We note that the figures of Table 15 say nothing about the multiprocessor capabilities of the CPUs. A significant performance degradation is observed when running multiple copies of the benchmark on those multiprocessor systems with "inferior" memory sub-systems. This effect must be considered when assessing the intuitive cost-effectiveness of multi-processor nodes. The delivered TRIAD rate per CPU for the dual-processor Xeon nodes (i.e. when running two concurrent STREAM benchmarks) is typically half the figure of Table 15.

3.3 THE CHEMISTRY APPLICATIONS BENCHMARK

3.3.1 The Quantum Chemistry Benchmark - GAMESS-UK-89

The GAMESS-UK-89 benchmark summarized below was designed to reflect the typical range of calculations performed by the ab initio quantum chemist. It includes 12 calculations carried out using the GAMESS-UK electronic structure code, and includes the following functionality;

• Conventional SCF Calculations, on morphine and 2,4,6 tri-nitro-toluene (Calculations 1. and 2. respectively);
• Valence-only ECP calculations, with a geometry optimization of Na ₇Mg⁺ in an ECP-DZ+D(Mg) basis of 70 GTOs (calculation 3).
• A Direct-SCF calculation on Cytosine (82 GTOs, 6-31G basis) Calculation 4);
• Multi-configuration SCF calculations, with a CASSCF geometry optimization on H₂CO (Calculation 5), and a larger CASSCF calculation, also on H₂CO (Calculation 6);
• Configuration interaction calculations, both Direct-CI on the H₂CO/H₂ +CO TS transition state (Calculation 7), and conventional table driven-CI on TiCl₄ (Calculation 8);
• Moller Plesset calculations, with a MP2 geometry optimization of H₃SiNCO (Calculation 9);
• Analytic second derivatives, at both the SCF (C₅ H₅N molecule in a 6-31G basis, Calculation 10) and MP2 level (C₄ in a 6-31G* basis, Calculation 11);
• A Direct-MP2 calculation on C₅H₅N in a DZ + D(N) basis set (Calculation 12).

The data presented in Table 12 is collected under control of the UNIX command time where available, and includes CPU time (both user and system), total elapsed time and Efficiency, measured as CPU versus elapsed. While our original aim was to base comparisons strictly on elapsed times, such timings could not be consistently gathered over the range of machines considered. For example, the range of disk configurations varies enormously, from primitive SCSI disks to striped high-speed raid disks, and the loading on the machines varies, from effectively single-user loading on many of the workstations, to multi-user environments. Thus while reporting the elapsed times, we use such figures to identify, where appropriate, the requirement for enhanced disk configurations, rather than as any implied criticism of the machine in question.

The total user CPU timings of Table 12 refer to the summed user CPU timings over all 12 calculations of the GAMESS-UK-89 benchmark.

3.3.1.1 GAMESS-UK-97

The GAMESS-UK-97 benchmark summarized below is designed to reflect a more demanding range of calculations performed by the ab initio quantum chemist. It includes 10 calculations carried out using the GAMESS-UK electronic structure code (Version 6), and includes the following functionality;

• A Direct-SCF calculation on Morphine (227 GTOs, 6-31G basis); the calculation converges in 9 SCF cycles (Calculation 1);
• Conventional SCF Calculation on 2,4,6 tri-nitro-toluene (6-31G** basis, 265 GTOs); requires 11 iterative SCF cycles and uses 2e-integral format (Calculation 2);
• DFT B3LYP calculation on Cytosine (167 GTOs, 6-311G* basis) conventional SCF converging in 10 SCF cycles, high quadrature (Calculation 3);
• CASSCF Multi-configuration SCF calculation on H₂CO (Calculation 4) conducted in a CC-PVTZ basis of 100 GTOs. The 16 electrons in 13 MOS (109857 csfs) converged in 3 macro-iterations;
• Direct Configuration interaction calculation on the H₂CO/H₂ +CO TS transition state (Calculation 5). Conducted in a CC-PVTZ basis of 100 GTOS (88 spherical harmonics), the in-store calculation included 1,629,857 CSFs generated from 16 reference functions;
• A coupled cluster CCSD(T) calculation on the C₄ molecule conducted in a tzv+2D+1F basis of 144 GTOs; a total of 172,811 configurations (CCSD) (Calculation 6);
• Moller Plesset calculations, with a MP2 geometry optimization of H₃SiNCO (Calculation 7); TZVP basis (105 GTOs), 5 energy and 5 gradient calculations for convergence;
• Analytic second derivatives, at both the SCF (C₅ H₅N molecule in a 6-311G** basis, (144 GTOs, Calculation 8) and MP2 level (C₄ in a TZVP basis (2D on carbon, 104 GTOs, Calculation 9);
• A Direct-MP2 energy calculation on Cytosine in a 6-31G* basis set (130 GTOS, Calculation 10) requiring 2 integral passes.

The data presented in Table 17 is collected under control of the UNIX command time where available, and includes CPU time (both user and system), total elapsed time and Efficiency, measured as CPU versus elapsed. The total User CPU timings of Table 17 refer to the summed user CPU timings over all 10 calculations of the GAMESS-UK-97 benchmark. An examination of the results leads to similar conclusions to those based on the GAMESS-UK-89 benchmarks that have been described in detail above.

3.3.2 The Molecular Simulation Benchmark - DL_POLY

The benchmark summarized below is designed to reflect the typical range of simulations undertaken by the molecular dynamicist. It includes 6 calculations carried out using the DL_POLY molecular dynamics code, and includes the following functionality;

• Benchmark 1: Simulation of a sodium-potassium disilicate glass (1080 atoms, 300 time steps);
• Benchmark 2: Simulation of metallic aluminium with Sutton-Chen potential at 300K (256 atoms, 8000 time steps);
• Benchmark 3: Simulation of valinomycin in 1223 water molecules (3837 atoms, 100 time steps);
• Benchmark 4: Dynamic Shell model water structure (768 atoms, 1024 sites, 1000 time steps);
• Benchmark 5: Dynamic Shell model MgCl₂ structure (768 atoms, 1280 sites, 1000 time steps);
• Benchmark 8: Simulation of a model membrane with 2 membrane chains, 202 solute molecules and 2746 solvent molecules (3148 atoms, 1000 time steps).

The data presented in Table 14 is collected under control of the UNIX command time where available, and includes CPU time (both user and system), total elapsed time and Efficiency, measured as CPU versus elapsed. The total user CPU timings of Table 14 refer to the summed user CPU timings over all 6 calculations of the benchmark. Note that in contrast to the QC benchmark, little I/O is performed by the DL_POLY calculations, so that efficiency should always be high assuming the benchmarks were conducted on a dedicated resource.

3.4 SUMMARY of the SERIAL Benchmarks

As a summary of this work, we present in Table 15 the relative performance of 120 CPUs against the HP RX5670 Itanium2/1500-H in terms of the CPU2000 benchmarks, and those from the present Matrix-97, Chemistry Kernels, GAMESS-UK and DL_POLY benchmarks. Note that while previous versions of this report have driven this comparison across the CPU95 ratings, we have now attempted to transition to CPU2000. In doing this, we clearly have a problem in terms of relating older machines, for which only SPECfp95 results are available, with the newer systems. Thus many of the SPECfp entries of Table 19 are estimated CPU2000 figures based on extrapolations from the older CPU95 values. This extrapolation process has involved identifying systems for which both SPECfp2000 and SPECfp95 figures are available, deriving appropriate scaling factors based on these values, and applying this scaling to the older CPUs. While not an ideal metric, we would suggest this approach is preferable to the alternative of trying to derive extrapolated SPEC95 figures from measured SPECfp2000 figures for the more recent systems.

4. The Chemistry RATE Benchmarks

4.1 The Individual ChemRATE Benchmarks

The Matrix Multiplication RATE Benchmark - 1. FORTRAN MMO

The order of the matrices involved in the MMO RATE Benchmark were increased still further from the MATRIX-97 benchmark, with the elapsed times of Table 16 referring to 10 MMOs of order 100, 9 MMOs of order 200 etc., up to a single MMO for matrices of order 1200.

The Matrix Multiplication RATE Benchmark - 2. FORTRAN MXMB0

The order of the matrices involved in the MXMB0 RATE Benchmark were increased still further from the MATRIX-97 benchmark, with the elapsed times of Table 17 referring to the time taken to perform the same set of matrix multiplications as specified in the MMO benchmark.

The Matrix Multiplication RATE Benchmark - 2. FORTRAN DGEMM

The order of the matrices involved in the DGEMM RATE Benchmark were as defined in the two RATE benchmarks above. Now the multiplications made no attempt to utilize sparsity in the B matrix, but resorted in each case to a single call to the BLAS3 library routine, DGEMM.

4.2 The Overall ChemRATE Benchmark

4.3 SUMMARY of the ChemRATE Benchmark

5. CONCLUSIONS

REFERENCES

1

M.F. Guest and S. Wilson, Daresbury Laboratory Preprint, DL/SCI/P290T; Supercomputers in Chemistry, ed. P. Lykos and I. Shavitt, A.C.S. Symposium series 173 (1981) 1; V.R. Saunders and M.F. Guest, Comp.Phys.Comm., 26 (1982) 389: M.F. Guest, in ``Supercomputer Simulations in Chemistry'', Ed. M. Dupuis, Lecture Notes in Chemistry, 44, Springer Verlag (1986) 98.

2

M.F. Guest, R.J. Harrison, J.H. van Lenthe and L.C.H. van Corler, ``Computational Chemistry on the FPS-X64 Scientific Computers: Experience on single- and multi-processor systems'', Theoret. Chim. Acta 71 (1987) 117.

3

M.F. Guest, ``Performance of Various Computers in Computational Chemistry,'' Proceedings of the Daresbury Machine Evaluation Workshop, November 1996.

4

GAMESS-UK is a package of ab initio programs written by M.F. Guest, J.H. van Lenthe, J. Kendrick, K. Schoeffel and P. Sherwood, with contributions from R.D. Amos, R.J. Buenker, M. Dupuis, N.C. Handy, I.H. Hillier, P.J. Knowles, V. Bonacic-Koutecky, W. von Niessen, R.J. Harrison, A.P. Rendell, V.R. Saunders, and A.J. Stone. The package is derived from the original GAMESS code due to M. Dupuis, D. Spangler and J. Wendoloski, NRCC Software Catalog, Vol. 1, Program No. QG01 (GAMESS), 1980.

5

DL_POLY is a parallel molecular dynamics simulation package developed at Daresbury Laboratory by W. Smith and T.R. Forester under the auspices of the Engineering and Physical Sciences Research Council (EPSRC) for the EPSRC's Collaborative Computational Project for the Computer Simulation of Condensed Phases (CCP5) and the Molecular Simulation Group (MSG) at Daresbury Laboratory. The package is the property of the Central Laboratory of the Research Councils.

6

In theory the O2-R10k should have outperformed the corresponding Indigo2; with better memory bandwidth, superior I/O and more tightly coupled integration it should have done well. However SGI made a design decision which has seriously impaired the performance of the O2 in some application areas. It took about 3 months for this "flaw" to be fully identified. Until December 1996, SGI were claiming that the O2 R10k would perform in the region of 10, 12 or even 15 SPECfp95. Indeed on some benchmarks it does indeed achieve performance that matches these figures. However, the O2 has a Unified Memory Architecture which uses main system memory as memory for the graphics display and operations. Despite the impressive bandwidth figures for the O2, it does seem that the O2 memory architecture severely impedes the performance of the R10k processor, particularly when compared with the Octane, Indigo2 and Origin systems. This is shown by the SPEC comparisons of Table 1; we suspect that the two main factors limiting performance in the memory subsystem are the main memory speed and the CRIME chip.

The CRIME chip, which acts as the memory interface between the memory and the three drains on it - the CPU (800 MByte/second), I/O engine (500 MByte/sec) and the monitor display (700 MByte/second) - is probably the main bottleneck. This chip was designed to work as a built in memory controller, but the design was biased toward the R5k; it can't work directly with the R10k because the R5k expects 32 byte cache refills while the R10k wants to have 64 or 128 byte refills. Therefore SGI supply a custom ASIC with the R10k daughter board. This interfaces the R10k's level 2 cache with the CRIME chip. Performance problems are caused by the ASIC having to break each 128 byte cache refill operation into 4, 32 byte refills. The net impact of this effect is that the O2 R10k will only work well with problems that fit into the L2 cache (1 MByte). Not surprisingly, the memory intensive SPECfp95 figures are badly affected, although the impact on less memory intensive applications is not so severe.

7

R. Shepard, R.A. Bair, R.A. Eades, A.F. Wagner, M.J. Davis, C.B. Harding and T.H. Dunning Jr., Int. J. Quant. Chem. , (1983) 17; R.A. Bair and T.H. Dunning Jr., J. Comp. Chem., 5 (1984) 44; R. Bair, ``FPS-164 Matrix Multiplication Subroutine Guide'', Argonne National Laboratory, (1984); T.H. Dunning Jr. and R.A. Bair, ``Advanced Theories and Computational Approaches to the Electronic Structure of Molecules'', NATO ASI Series, D. Reidel, 1984, p1.

8

IMSL Program Library, 1978.

9

NAG Fortran Library, Numerical Algorithms Group Ltd, 1984.

10

Y. Beppu, Computers and Chemistry, 6 (1982).

11

J.P. Stewart, ``MOPAC - A General Molecular Orbital Package'', QCPE 455, 1984.

12

V.R. Saunders and M.F. Guest, ``ATMOL3 Part 9'', RL-76-106, 1976; M.F. Guest and V.R. Saunders, Mol. Phys., 28 (1974) 819; D. Moncrieff and V.R. Saunders, ``ATMOL-Introduction Notes'', UMRCC, May, 1986; Cyber-205 Note Number 32, UMRCC, September, 1985.

13

J. Almlof et al, J. Comp. Chem., 3 (1982) 385. Performance of Various Computers in Computational Chemistry

14

see the following web page to obtain Pentium Pro Optimized BLAS and FFTs for Intel Linux: http://www.cs.utk.edu/ ghenry/distrib

15

the ATLAS blas http://www.netlib.org/atlas/index.html gives a 60% speed improvement over the latest intel blas library, given the changes suggested at http://www.cs.utk.edu/~rwhaley/ATLAS/errata.html

16

More evidence of the poor performance of Sun Fortran f90 Compiler is shown in the following figures Linpack 100X100 all Fortran benchmark kindly provided by Hans-Hermann Frese of ZIB, Berlin. The results using different Sun Fortran compiler releases on a Sun Ultra 60 Model 2360 used one UltraSPARC CPU at 360 MHz. The performance of the latest NAGWare f95 is also compared:

Linpack 100X100 Performance as a function of FORTRAN compiler.

Compiler	Options	Linpack 100X100
Sun f77 4.2	-fast -O5	164 Mflop/s
Sun f77 5.0	-fast -O5	166 Mflop/s
Sun f90 1.2	-fast	61 Mflop/s
Sun f90 2.0	-fast -ftrap=no%division	68 Mflop/s
NAGWare f95 4.0	-Wc,-fast	166 Mflop/s

These figures underline that the performance of the Sun Fortran 90 compiler is still poor compared with Sun's Fortran 77 compiler. Surprisingly, the performance of the NAGWare Fortran 95 compiler which produces intermediate C code is as good as Sun's native Fortran 77 compiler.



Table 1. SPEC CPU2000 - SPECfp and SPECint. Absolute and Relative Values.

SPEC CPU2000 - SPECfp and SPECint. Absolute Values and Values Relative to the HP RX5670 Itanium2/1500-H.

Rank	Machine	SPECfp	SPECint	Test	Relative Values (%)
				Date	SPECfp	SPECint
1	HP Integrity RX4640/Itanium2 1.6G/9MB	2712.0	1590.0	Nov-04	129%	121%
2	IBM eServer p5 570 (1900 MHz) 2CC	2702.0	1452.0	Jul-04	128%	111%
3	HP Integrity RX1620/Itanium2 1.6G/3MB	2692.0	1452.0	Nov-04	128%	111%
4	HP Integrity RX2620/Itanium2 1.6G/6MB	2675.0	1535.0	Dec-04	127%	117%
5	SGI Altix 3700 Bx2/Itanium2 1.6G/9MB	2647.0	1441.0	Nov-04	126%	110%
6	SGI Altix 3700 Bx2/Itanium2 1.6G/6MB	2600.0	1410.0	Nov-04	123%	107%
7	Bull NovaScale 4040 Itanium2/1.6Hz 9MBL3	2566.0	1457.0	Nov-04	122%	111%
8	HP Integrity RX2620/Itanium2 1.6G/3MB	2553.0	1408.0	Nov-04	121%	107%
9	Dell PowerEdge 7250 Itanium2/1600 9MB L3	2545.0	1403.0	Nov-04	121%	107%
10	HP Integrity RX4640/Itanium2 1.5G/4MB	2502.0	1372.0	Nov-04	119%	105%
11	IBM eServer p5 590 (1650 MHz, 1 CPU)	2398.0	1259.0	Oct-04	114%	96%
12	SGI Altix 3700 Bx2/Itanium2 1.5G/4MB	2361.0	1242.0	Dec-04	112%	95%
13	HP Integrity RX2620/Itanium2 1.3G/3MB	2229.0	1170.0	Dec-04	106%	89%
14	IBM eServer p5 550 (1650 MHz) 2CC	2221.0	1248.0	Jul-04	105%	95%
15	IBM eServer i5 550 (1650 MHz) 2CC	2221.0	1248.0	Aug-04	105%	95%
16	HP Integrity RX1620/Itanium2 1.3G/3MB	2214.0	1178.0	Dec-04	105%	90%
17	HP Integrity RX4640/Itanium2,1.5GHz	2161.0	1404.0	Jan-04	103%	107%
18	SGI Altix 3000 (1.5GHz, Itanium 2)	2148.0	1243.0	Mar-04	102%	95%
19	IBM eServer p5 520 (1650 MHz) 2CC	2138.0	1248.0	Jul-04	101%	95%
20	IBM eServer i5 520 (1650 MHz) 2CC	2138.0	1248.0	Aug-04	101%	95%
21	IBM eServer p5 550 Express 1500MHz	2127.0	-	Oct-04	101%	-
22	HP Integrity RX2600/Itanium2,1.5GHz	2119.0	1408.0	Jun-03	101%	107%
23	HP Integrity RX5670/Itanium2,1.5GHz	2108.0	1312.0	Jun-03	100%	100%
24	HP ZX6000/Itanium2, 1.5GHz,6MB-L3, H	2106.0	1315.0	Jun-03	100%	100%
25	IBM eServer p5 520 Express 1500MHz	2103.0	-	Oct-04	100%	-
26	ION Computers I2X4 (1.5GHz Itanium2)	2043.0	1232.0	May-04	97%	94%
27	Bull NovaScale 4040 Itanium2/1.5GHz	2015.0	1232.0	Jul-03	96%	94%
28	ION Computers I2X2 (1.6GHz Itanium2)	2001.0	1178.0	May-04	95%	90%
29	IBM eServer OpenPower 720 (1650 MHz)	1966.0	1138.0	Oct-04	93%	87%
30	SGI Altix 350/Itanium2 1.4GHz/3MB	1942.0	1078.0	May-04	92%	82%
31	HP Integrity RX4640/Itanium2,1.3GHz	1891.0	1132.0	Jan-04	90%	86%
32	AMD MSI K8N Neo2 Athlon64/2.6GHz FX-55	1878.0	1735.0	Nov-04	89%	132%
33	Dell PowerEdge3250/Itanium2 1.5GHz/6ML3	1875.0	1099.0	Jul-03	89%	84%
34	Dell PowerEdge3250/Itanium2 1.4GHz/3ML3	1868.0	1022.0	Mar-04	89%	78%
35	SGI Altix 3000 (1.3GHz, Itanium 2)	1854.0	1019.0	Mar-04	88%	78%
36	Fujitsu CELSIUS V810 / 2.4GHz Opteron 250 L	1849.0	1619.0	Nov-04	88%	123%
37	Intel D925XECV2 3.80 GHz P4	1842.0	1671.0	Nov-04	87%	127%
38	Supermicro SS6113M-i(1.5GHz/6MB, Itanium 2)	1830.0	1090.0	Dec-03	87%	83%
39	ION Computers I2X2 (1.4GHz Itanium2)	1817.0	926.0	Jul-03	86%	71%
40	HP Integrity RX5670/Itanium2,1.3GHz	1814.0	1066.0	Nov-03	86%	81%
41	HP Integrity RX2600/Itanium2,1.3GHz	1808.0	1073.0	Nov-03	86%	82%
42	PrimePower900 (SPARC64 GP/1890MHz)	1803.0	1345.0	May-04	86%	103%
43	PrimePower650 (SPARC64 GP/1890MHz)	1803.0	1345.0	May-04	86%	103%
44	Sun Java Wrkst. W1100z Opteron2.4GHz	1787.0	1582.0	Jul-04	85%	121%
45	AMD MSI K8N Neo2 Athlon64/2.6GHz FX-55	1782.0	1854.0	Sep-04	85%	141%
46	IBM eServer pSeries655 (1.7GHz/1CPU)	1776.0	1158.0	Jan-04	84%	88%
47	IBM eServer 326 2.4 GHz Opteron EKO	1771.0	1584.0	Oct-04	84%	121%
48	Intel D925XECV2 3.46 GHz P4	1724.0	1772.0	Nov-04	82%	135%
49	IBM xSeries 336 (3.6GHz Xeon)	1721.0	-	Sep-04	82%	-
50	HP ProLiant DL145 Opteron250/2.4GHz	1701.0	1561.0	Oct-04	81%	119%
51	IBM eServer pSeries 690Turbo 1.7 GHz	1699.0	1113.0	May-03	81%	85%
52	Sun Fire V20z (Opteron250/2.4GHz)	1693.0	1521.0	Jul-04	80%	116%
53	Sun Fire V40z (Opteron850/2.4GHz)	1692.0	1515.0	Jul-04	80%	115%
54	PathScale ASUS SK8N/Opteron248 2.2G	1691.0	1452.0	Apr-04	80%	111%
55	HP AlphaServer GS1280 7/1300	1684.0	994.0	Jun-04	80%	76%
56	SGI Altix 350/Itanium2 1.4GHz/1.5MB	1684.0	986.0	Dec-03	80%	75%
57	IBM eServer pSeries 655 1700 MHz	1678.0	1103.0	May-03	80%	84%
58	AMD ASUS SK8V/Opteron150 2.4GHz SES	1644.0	1487.0	May-04	78%	113%
59	IBM xSeries 336 (3.6GHz,800MHz FSB)	1641.0	-	Aug-04	78%	-
60	AMD MSI K8T Neo2-FIR Athlon64 2.4GHz	1634.0	1717.0	Sep-04	78%	131%
61	Supermicro Super P8SC8 (3.80 GHz Pentium 4)	1631.0	1639.0	Nov-04	77%	125%
62	Intel D925XCV (AA-400 3.6GHz,P4)	1630.0	1575.0	Jun-04	77%	120%
63	HP ProLiant BL20p G3 3.6GHz Xeon	1597.0	1543.0	Nov-04	76%	118%
64	Intel D925XCV (AA-400 3.4GHz,P4)	1581.0	1705.0	Jun-04	75%	130%
65	Intel D925XECV2 3.0E GHz P4	1571.0	1369.0	Nov-04	75%	104%
66	Intel D875PBZ (AA-301 3.4GHz,P4)	1561.0	-	Feb-04	74%	-
67	Intel D875PBZ (AA-206 3.4GHz P4)	1556.0	1705.0	Feb-04	74%	130%
68	IBM eServer 325 2.4GHz Opteron 64-PGI	1553.0	1472.0	May-04	74%	112%
69	Dell Prec. Workst. 360/3.4GHz P4 Extreme	1548.0	1689.0	Jun-04	73%	129%
70	Supermicro P8SAA ( 3.6 GHz Pentium 4, 1MBL2)	1541.0	1560.0	Nov-04	73%	119%
71	AMD Tyan K8QS Opteron850/2.4GHz	1531.0	1663.0	Sep-04	73%	127%
72	HP ProLiant ML350 G4 3.4GHz Xeon	1531.0	1452.0	Nov-04	73%	111%
73	AMD ASUS SK8V/Opteron150 2.4GHz WS	1528.0	1655.0	May-04	72%	126%
74	ION Computers SR1325TP (3.4GHz Pentium 4 H	1524.0	1671.0	May-04	72%	127%
75	HP ProLiant ML370 G4 3.6GHz Xeon	1523.0	1556.0	Oct-04	72%	119%
76	Fujitsu CELSIUS M420/3.4EGHz P4(800MHz Bus)	1522.0	1501.0	Jun-04	72%	114%
77	IBM eServer pSeries 670 1.5 GHz	1520.0	981.0	May-03	72%	75%
78	AMD ASUS SK8N/Opteron148 2.2GHz SES	1514.0	1409.0	Nov-03	72%	107%
79	HP ProLiant DL380 G4 3.6GHz Xeon,L3	1514.0	1553.0	Jul-04	72%	118%
80	Fujitsu CELSIUS M430/3.4FGHz P4(800MHz Bus)	1512.0	1484.0	Aug-04	72%	113%
81	Dell PowerEdge 1850/3.6GHz P4Xeon	1503.0	1538.0	Aug-04	71%	117%
82	Intel D925XECV2 2.8E GHz P4	1497.0	1290.0	Nov-04	71%	98%
83	HP AlphaServer ES80 7/1150	1494.0	882.0	Jun-04	71%	67%
84	HP AlphaServer ES47 7/1150	1494.0	882.0	Jun-04	71%	67%
85	Intel D875PBZ (AA-206 3.2GHz P4)	1493.0	1620.0	Nov-03	71%	123%
86	Dell Prec. Workst. 360/3.40E GHz P4	1492.0	1484.0	Jun-04	71%	113%
87	HP ProLiant ML370 G4 3.4GHz Xeon	1490.0	1489.0	Aug-04	71%	113%
88	AMD ASUS SK8N/Opteron148 2.2GHz WS	1490.0	1477.0	Nov-03	71%	113%
89	Dell PowerEdgeSC 1425 (3.6 GHz Xeon)	1490.0	1541.0	Nov-04	71%	117%
90	Dell PowerEdge 1855/3.6GHz Xeon	1489.0	1540.0	Nov-04	71%	117%
91	IBM eServer pSeries 655 1500 MHz	1488.0	970.0	May-03	71%	74%
92	Dell Prec. Workst. 360/3.2GHz P4 Extreme	1486.0	1601.0	Feb-04	70%	122%
93	Intel D875PBZ (AA-301:3.4EGHz,P4)	1485.0	1491.0	Apr-04	70%	114%
94	HP AlphaServer GS1280 7/1150	1482.0	877.0	Jan-03	70%	67%
95	Intel D925XCV (AA-400 3.0EGHz,P4)	1459.0	1353.0	Jun-04	69%	103%
96	Supermicro SUPER P4SCT+II 3.4GHz P4,800MHz bus	1456.0	1493.0	Jun-04	69%	114%
97	Dell PowerEdge 2850/3.6 GHz P4Xeon	1455.0	1522.0	Aug-04	69%	116%
98	HP ProLiant DL380 G4 3.4GHz Xeon,L3	1453.0	1477.0	Jul-04	69%	113%
99	Dell PowerEdge3250/Itanium2 1.4GHz/1.5ML3	1444.0	824.0	Sep-03	69%	63%
100	Dell Prec. Workst. 360/3.20E GHz P4	1441.0	1404.0	Jan-04	68%	107%
101	Intel D875PBZ (AA-301 3.2EGHz,P4)	1437.0	1421.0	Feb-04	68%	108%
102	HP RX5670/Itanium2, 1GHz,3MB-L3, L	1431.0	-	Nov-02	68%	-
103	HP RX2600/Itanium2, 1GHz,3MB-L3, L	1427.0	-	Nov-02	68%	-
104	AMD ASUS SK8N Athlon FX-51/2.2GHz(W)	1423.0	1447.0	Sep-03	68%	110%
105	HP ZX6000/Itanium2, 1GHz,3MB-L3, L	1422.0	-	Nov-02	67%	-
106	Fujitsu CELSIUS V810 / 2.2GHz Opteron 248 L	1420.0	1354.0	Dec-03	67%	103%
107	SGI Altix 3000 (1GHz, Itanium 2)	1410.0	683.0	May-03	67%	52%
108	IBM eServer pSeries655 1700MHz/LPR/L	1405.0	1023.0	Jul-03	67%	78%
109	Bull NovaScale 4040 Itanium2/1.0GHz	1404.0	719.0	Mar-03	67%	55%
110	IBM eServer pSeries655 1700MHz/L	1401.0	1031.0	Sep-03	66%	79%
111	Fujitsu CELSIUS V810 / 2.2GHz Opteron 248 W	1397.0	1429.0	Dec-03	66%	109%
112	AMD ASUS SK8N/Opteron146 2.0GHz HE	1394.0	1354.0	Feb-04	66%	103%
113	HP ProLiant DL360 G4 3.6GHz Xeon	1389.0	1467.0	Oct-04	66%	112%
114	Dell Prec. Workst. 360/3.00E GHz P4	1385.0	1329.0	Jan-04	66%	101%
115	Intel D875PBZ (AA-301 3.0EGHz,P4)	1384.0	1345.0	Feb-04	66%	103%
116	HP Integrity RX1600/Itanium2,1.0GHz	1382.0	837.0	Feb-04	66%	64%
117	AMD ASUS SK8N Athlon FX-51/2.2GHz(L)	1378.0	1395.0	Sep-03	65%	106%
118	SGI Altix 350 (1.0GHz, Itanium 2)	1374.0	743.0	Feb-04	65%	57%
119	IBM x306 (3.0EGHz,Pentium4,1CC)	1370.0	1336.0	Mar-04	65%	102%
120	HP AlphaServer ES45 68/1250	1365.0	928.0	Nov-02	65%	71%
121	HP ProLiant BL30p (3.2GHz Xeon,2MB L3)	1357.0	1550.0	Nov-04	64%	118%
122	HP ProLiant DL360 G3 3.2GHz/Xeon,2ML3 1	1350.0	1541.0	Apr-04	64%	117%
123	HP ProLiant DL360 G4 3.4GHz Xeon,L3	1350.0	1402.0	Jul-04	64%	107%
124	PrimePower450 (SPARC64 GP/1320MHz)	1350.0	882.0	Dec-03	64%	67%
125	Sun Fire V440/UltraSPARC-IIi 1.6GHz	1349.0	845.0	Nov-04	64%	64%
126	HP ProLiant DL380 G3 3.2GHz/Xeon,2ML3 1	1349.0	1534.0	Apr-04	64%	117%
127	Dell PowerEdge 2650/3.2 GHz P4Xeon,2ML3	1347.0	1420.0	Feb-04	64%	108%
128	Dell Prec. WorkSt.650/3.2GHz P4Xeon 2ML3	1346.0	1563.0	Jan-04	64%	119%
129	PrimePower650 (SPARC64 GP/1350MHz)	1340.0	905.0	Jun-03	64%	69%
130	AMD ASUS SK8N Opteron146 2.0GHz-I,WS	1339.0	1335.0	Sep-03	64%	102%
131	Intel D875PBZ (AA-301 2.8EGHz,P4)	1328.0	1268.0	Feb-04	63%	97%
132	Intel D875PBZ (AA-301:3.4GHz,Pentium	1328.0	-	Feb-04	63%	-
133	Dell Prec. Workst. 360/2.80E GHz P4	1327.0	1254.0	Jan-04	63%	96%
134	PrimePower250 (SPARC64 GP/1320MHz)	1325.0	881.0	Dec-03	63%	67%
135	HP ProLiant ML370 G3 (3.2GHz/2MB L3	1323.0	1530.0	Jun-04	63%	117%
136	PrimePower900 (SPARC64 GP/1350MHz)	1322.0	892.0	May-03	63%	68%
137	Intel D875PBZ (3.4 GHz, Pentium4)	1308.0	1393.0	Feb-04	62%	106%
138	Dell Prec. Workst. 360/3.40 GHz P4	1306.0	1369.0	Jan-04	62%	104%
139	HP RX5670/Itanium2, 1GHz,3MB-L3, H	1305.0	807.0	Jul-02	62%	62%
140	HP RX2600/Itanium2, 1GHz,3MB-L3, H	1301.0	810.0	Jul-02	62%	62%
141	AMD ASUS SK8N Opteron146 2.0GHz-I,SES	1300.0	1262.0	Sep-03	62%	96%
142	IBM eServer pSeries 650/6M2 1450MHz	1295.0	935.0	Nov-02	61%	71%
143	AMD Rioworks HDAMA Opteron 246 / 2.0 G	1293.0	1317.0	Jul-03	61%	100%
144	IBM x335 (3.2GHz, 533MHz FSB)	1292.0	1517.0	Feb-04	61%	116%
145	HP AlphaServer ES80 7/1000	1288.0	761.0	Jan-03	61%	58%
146	HP AlphaServer ES47 7/1000	1288.0	761.0	Jan-03	61%	58%
147	Fujitsu Celsius M420 / 3.2GHz P4(800MHz bus)	1287.0	1261.0	Jul-03	61%	96%
148	Dell Prec. Workst. 360/3.2 GHz P4 DDR400	1285.0	1249.0	Jun-03	61%	95%
149	Dell PowerEdge 6600/3.0 GHz Xeon MP	1283.0	1408.0	Feb-04	61%	107%
150	IBM eServer pSeries 655/651 1300 MHz	1281.0	848.0	Nov-02	61%	65%
151	IBM eServer pSeries 690Turbo 1.3 GHz	1266.0	839.0	Apr-02	60%	64%
152	Intel D875PBZ (3.2 GHz, Pentium4)	1260.0	1329.0	Nov-03	60%	101%
153	AMD MSI K8 Neo Athlon64 3200+/2 GHz	1250.0	1335.0	Sep-03	59%	102%
154	Dell Prec. Workst. 360/3.0 GHz P4 DDR400	1248.0	1193.0	Jun-03	59%	91%
155	IBM BladeCenter JS20 (PPC-970,2.2GHz)	1241.0	1040.0	Oct-04	59%	79%
156	IBM eServer 325 2.0GHz Opteron 64-gcc	1231.0	1239.0	Sep-03	58%	94%
157	ION Computers BX21 (3.2GHz Xeon/2ML3 1C)	1229.0	1495.0	May-04	58%	114%
158	Intel D875PBZ (3.0 GHz, Pentium4)	1229.0	1265.0	May-03	58%	96%
159	ION Computers SR2300WV2 (3.2GHz Xeon/2ML3)	1223.0	1455.0	Feb-04	58%	111%
160	Dell PowerEdge 2650/3.2 GHz P4Xeon,1ML3	1220.0	1306.0	Oct-03	58%	100%
161	AMD Einox A4800/Opteron144 1.8GHz-I,WS	1219.0	1170.0	Apr-03	58%	89%
162	Dell Prec. WorkSt.650/3.2GHz P4 Xeon L3	1213.0	1333.0	Sep-03	58%	102%
163	Dell Prec. Workst. 360/2.80E GHz DDR533	1212.0	1211.0	Jan-04	57%	92%
164	HP ProLiant DL580 G2 3.0GHz XeonMP 1CC	1208.0	1491.0	Mar-04	57%	114%
165	Dell Prec. Workst. 360/2.8 GHz P4 DDR400	1204.0	1129.0	Jun-03	57%	86%
166	Dell Prec. WorkSt.650/3.2GHz P4Xeon 1ML3	1202.0	1415.0	Jan-04	57%	108%
167	HP ProLiant BL20p G2 3.2GHz Xeon,1ML3	1198.0	1321.0	Oct-03	57%	101%
168	HP ProLiant DL380 G3 3.2GHz Xeon,L3	1197.0	1313.0	Oct-03	57%	100%
169	HP ProLiant DL360 G3 3.2GHz Xeon,L3	1197.0	1319.0	Oct-03	57%	101%
170	Dell PowerEdge 2650/3.06 GHz P4Xeon,1ML3	1195.0	1263.0	Oct-03	57%	96%
171	HP ProLiant BL20p G2 3.06GHzXeon,1ML3	1193.0	1256.0	Aug-03	57%	96%
172	Dell PowerEdge 1750/3.2 GHz P4Xeon,1ML3	1192.0	1289.0	Oct-03	57%	98%
173	Dell Prec. WorkSt.650/3.06GHz P4 Xeon L3	1186.0	1294.0	Jun-03	56%	99%
174	HP ProLiant DL380 G3 3.06GHz Xeon,L3	1184.0	1258.0	Jun-03	56%	96%
175	HP ProLiant DL360 G3 3.06GHz Xeon,L3	1183.0	1258.0	Jun-03	56%	96%
176	IBM eServer 325 2.0GHz Opteron 64-PGI	1180.0	1226.0	Jul-03	56%	93%
177	IBM IntelliStation POWER 275 1.45GHz	1180.0	978.0	Jun-03	56%	75%
178	ION Computers BX41 (3GHz Xeon MP/4ML3 1C)	1178.0	1486.0	May-04	56%	113%
179	Dell Prec. WorkSt.650/3.06GHz P4Xeon 1ML3	1176.0	1368.0	Feb-04	56%	104%
180	AMD Einox A4800/Opteron144 1.8GHz-I,SE	1168.0	1100.0	Apr-03	55%	84%
181	Dell Prec. Workst. 360/3.2 GHz P4 DDR333	1168.0	1201.0	Jun-03	55%	92%
182	Dell Prec. Workst. 360/2.6 GHz P4 DDR400	1160.0	1067.0	Jun-03	55%	81%
183	IBM eServer pSeries 630/6C4 1450MHz	1158.0	910.0	Feb-03	55%	69%
184	IBM eServer pSeries 630/6E4 1450MHz	1158.0	910.0	Feb-03	55%	69%
185	AMD Einox A4800/Opteron144 1.8GHz-I,SE	1154.0	1100.0	Apr-03	55%	84%
186	Dell PowerEdge 1750/3.06 GHz Xeon L3	1152.0	1243.0	Jul-03	55%	95%
187	HP RX5670/Itanium2,900MHz,1.5MB-L3,L	1151.0	-	Nov-02	55%	-
188	HP RX2600/Itanium2,900MHz,1.5MB-L3,L	1151.0	-	Nov-02	55%	-
189	HP ZX6000/Itanium2,900MHz,1.5MB-L3,L	1139.0	-	Nov-02	54%	-
190	Dell Prec. Workst. 360/3.0 GHz P4 DDR333	1137.0	1149.0	Jun-03	54%	88%
191	PrimePower650 (SPARC64 GP/1080MHz)	1129.0	737.0	May-03	54%	56%
192	Dell PowerEdge 6650/2.8 GHz Xeon MP	1120.0	1234.0	Jun-03	53%	94%
193	AMD Einox A4800/Opteron142 1.6GHz-I,WS	1120.0	1053.0	Apr-03	53%	80%
194	Sun Blade Model 2000 / 1200 MHz	1118.0	722.0	Mar-03	53%	55%
195	HP ProLiant BL40p (2.8GHz Xeon MP)	1112.0	1249.0	Sep-03	53%	95%
196	Dell Prec. Workst. 360/2.4 GHz P4 DDR400	1111.0	1003.0	Jun-03	53%	76%
197	Dell Prec. WorkSt.650/2.8GHz P4Xeon 1ML3	1107.0	1270.0	Jan-04	53%	97%
198	Sun Netra 20 / 1200 MHz	1106.0	712.0	Mar-03	52%	54%
199	Sun Fire 280R / 1200 MHz Cu	1106.0	712.0	Mar-03	52%	54%
200	HP ProLiant DL560 (2.8GHz Xeon MP)	1105.0	1247.0	Sep-03	52%	95%
201	Dell Prec. Workst. 360/2.8 GHz P4 DDR333	1104.0	1093.0	Jun-03	52%	83%
202	IBM eServer pSeries 655/651 1100 MHz	1103.0	722.0	Nov-02	52%	55%
203	Dell Prec. WorkSt. 350/3.06 GHz P4 Xeon	1103.0	1130.0	Nov-02	52%	86%
204	HP ProLiant DL580 G2 2.8GHz Xeon MP	1094.0	1233.0	Sep-03	52%	94%
205	Intel D850EMVR(3.06 GHz,Pentium4)	1092.0	1098.0	Apr-03	52%	84%
206	IBM eServer pSeries 650/6M2-1 1450MH	1091.0	886.0	Apr-03	52%	68%
207	Intel D850EMVR(3.06GHz,Pentium4,MT)	1091.0	1107.0	Dec-02	52%	84%
208	SGI Altix 3000 (900MHz, Itanium 2)	1090.0	-	Feb-03	52%	-
209	Dell Prec. Mob. Workst. M60 Pentium M 755	1088.0	1541.0	Jun-04	52%	117%
210	HP ZX2000/Itanium2,900MHz,1.5MB-L3,L	1086.0	-	Nov-02	52%	-
211	Sun Fire V480 / 1200MHz Cu	1084.0	702.0	Oct-03	51%	54%
212	Sun Fire V880 / 1200 MHz Cu	1082.0	700.0	Sep-03	51%	53%
213	ION Computers SR2300WV2 (3.2GHz Xeon/1ML3,	1078.0	1330.0	Feb-04	51%	101%
214	AMD ASUS SK8N/Opteron140 1.4GHz EE	1075.0	990.0	Feb-04	51%	75%
215	IBM eServer pSeries 670 Pwr4/1.1 GHz	1075.0	708.0	Apr-02	51%	54%
216	HP ProLiant BL20p G2 3.06GHz Xeon	1074.0	1075.0	Aug-03	51%	82%
217	Dell Prec. Workst. 360/2.6 GHz P4 DDR333	1067.0	1034.0	Jun-03	51%	79%
218	HP ProLiant DL380 G3 3.06GHz Xeon	1065.0	1068.0	Jun-03	51%	81%
219	Sun Fire V240 / 1.28GHz Cu	1063.0	704.0	Jan-04	50%	54%
220	Dell Prec. WorkSt.650/3.06 GHz P4 Xeon	1063.0	1138.0	Apr-03	50%	87%
221	Intel D850EMVR (2.8 GHz, Pentium4)	1060.0	1041.0	Apr-03	50%	79%
222	Dell Prec. WorkSt. 350/2.8 GHz P4 Xeon	1060.0	1061.0	Nov-02	50%	81%
223	Sun Fire V250 / 1280MHz Cu	1054.0	702.0	Sep-03	50%	54%
224	HP ProLiant DL360 G3 3.06GHz Xeon	1048.0	1070.0	Jun-03	50%	82%
225	Supermicro X5DPA-GG (3.2GHz, 533MHz bus)	1046.0	1270.0	Jan-04	50%	97%
226	Dell PowerEdge 1750/3.06 GHz Xeon	1044.0	1067.0	Jul-03	50%	81%
227	PrimePower450 (SPARC64 GP/1100MHz)	1037.0	682.0	May-03	49%	52%
228	Dell Prec. WorkSt. 350/2.66 GHz P4 Xeon	1036.0	1026.0	Nov-02	49%	78%
229	Dell Prec. WorkSt.650/2.4GHz P4Xeon 1ML3	1031.0	1123.0	Jan-04	49%	86%
230	PrimePower250 (SPARC64 GP/1100MHz)	1030.0	680.0	May-03	49%	52%
231	Sun Fire V1280 / 1200MHz Cu	1027.0	676.0	Jan-04	49%	52%
232	Dell Prec. Workst. 360/2.4 GHz P4 DDR333	1027.0	972.0	Jun-03	49%	74%
233	Intel D850EMVR (2.67 GHz, Pentium4)	1024.0	1005.0	Oct-02	49%	77%
234	Sun Blade Model 2500 / 1.28 GHz	1021.0	696.0	Aug-03	48%	53%
235	Dell Prec. WorkSt.650/2.8 GHz P4 Xeon	1020.0	1068.0	Apr-03	48%	81%
236	IBM eServer pSeries 615/6C3 1.2 GHz	1018.0	822.0	May-03	48%	63%
237	IBM eServer pSeries 615/6E3 1.2 GHz	1018.0	822.0	May-03	48%	63%
238	HP AlphaServer GS160 68/1224	1014.0	833.0	Nov-02	48%	63%
239	IBM eServer pSeries 630/6E4 1200MHz	1014.0	767.0	Apr-03	48%	58%
240	HP AlphaServer GS320 68/1224	1014.0	833.0	Nov-02	48%	63%
241	HP AlphaServer GS80 68/1224	1014.0	833.0	Nov-02	48%	63%
242	IBM eServer pSeries 630/6C4 1200MHz	1014.0	767.0	Apr-03	48%	58%
243	AMD Einox A4800/Opteron140 1.4GHz-I,WS	1012.0	933.0	Apr-03	48%	71%
244	Sun Fire V65x / (3.06 GHz Xeon)	1011.0	1066.0	May-03	48%	81%
245	Dell Prec. WorkSt. 350/2.53 GHz P4 Xeon	1005.0	986.0	Nov-02	48%	75%
246	ION Computers SR2300WV2 (3.066GHz Xeon/512	1003.0	1139.0	Feb-04	48%	87%
247	Dell PowerEdge 2650/3.06 GHz P4 Xeon	1003.0	1056.0	Apr-03	48%	80%
248	Dell Prec. WorkSt.650/2.66 GHz P4 Xeon	1002.0	1033.0	Apr-03	48%	79%
249	Intel D850EMVR (2.53 GHz, Pentium4)	999.0	956.0	Apr-03	47%	73%
250	HP AlphaServer DS25 68/1000	985.0	678.0	Aug-02	47%	52%
251	HP ProLiant DL140 (3.2GHz Xeon,1ML3)	984.0	1258.0	Nov-03	47%	96%
252	IBM eServer pSeries 630/6C4-1 1.45GH	984.0	856.0	Apr-03	47%	65%
253	IBM eServer pSeries 630/6E4-1 1450MH	984.0	856.0	Apr-03	47%	65%
254	AMD ASUS A7N8X XP 3200+ / 2200 MHz	982.0	1080.0	May-03	47%	82%
255	Sun Fire V880 / 1050 MHz Cu	982.0	626.0	Jun-03	47%	48%
256	Dell Prec. WorkSt. 340/3.06 GHz P4 Xeon	982.0	1074.0	Nov-02	47%	82%
257	Dell Prec. WorkSt. 350/2.4 GHz P4 Xeon	979.0	951.0	Nov-02	46%	72%
258	Dell Prec. Workst. 360/2.26 GHz P4 Xeon	965.0	904.0	Jun-03	46%	69%
259	Intel D850EMVR (2.4 GHz, Pentium4)	964.0	927.0	Oct-02	46%	71%
260	Sun Fire V480 / 1050MHz Cu	962.0	619.0	Jul-03	46%	47%
261	ION Computers BX21 (2.8GHz Xeon/512kB L2,1	960.0	1050.0	May-04	46%	80%
262	Compaq AlphaServer ES45 Model 68/1000	960.0	679.0	Oct-01	46%	52%
263	Dell Prec. WorkSt. 350/2.26 GHz P4 Xeon	949.0	909.0	Nov-02	45%	69%
264	Supermicro X5DPA-GG (3.06GHz, 533MHz bus)	949.0	1056.0	Jan-04	45%	80%
265	Dell Prec. WorkSt. 340/2.8 GHz P4 Xeon	947.0	1010.0	Sep-02	45%	77%
266	Dell Prec. WorkSt.650/2.4 GHz P4 Xeon	944.0	955.0	Apr-03	45%	73%
267	Intel D850EMVR (2.26 GHz, Pentium4)	934.0	888.0	Oct-02	44%	68%
268	Fujitsu CELSIUS R610 / 2.8GHz Pentium4 Xeon	930.0	1016.0	Feb-03	44%	77%
269	Intel D850EMV2 (2.8 GHz, Pentium4)	929.0	984.0	Aug-02	44%	75%
270	Dell Prec. WorkSt. 340/2.66 GHz P4 Xeon	927.0	978.0	Sep-02	44%	75%
271	Intel D850EMV2 (2.67 GHz, Pentium4)	911.0	952.0	Aug-02	43%	73%
272	Dell Prec. WorkSt. 340/2.53 GHz P4 Xeon	905.0	941.0	Sep-02	43%	72%
273	IBM IntelliStation POWER 275 1.0GHz	901.0	683.0	Jun-03	43%	52%
274	Intel D850EMV2 (2.53GHz, Pentium4)	889.0	914.0	Aug-02	42%	70%
275	Dell Prec. WorkSt. 530/2.8 GHz P4 Xeon	887.0	957.0	Sep-02	42%	73%
276	IBM eServer pSeries 630/6C4 1.0GHz	886.0	639.0	Jul-02	42%	49%
277	IBM eServer pSeries 630/6E4 1.0GHz	886.0	639.0	Jul-02	42%	49%
278	Sun Blade 1500 / 1.062GHz	884.0	589.0	Jun-03	42%	45%
279	Dell Prec. WorkSt. 340/2.4 GHz P4 Xeon	882.0	905.0	Sep-02	42%	69%
280	HP AlphaServer DS15/1000	876.0	631.0	Jun-04	42%	48%
281	Intel D850EMV2 (2.4 GHz, Pentium4)	870.0	882.0	Aug-02	41%	67%
282	Dell Prec. WorkSt.650/2.0 GHz P4 Xeon	870.0	837.0	Apr-03	41%	64%
283	AMD ASUS A7N8X XP 3000+ / 2167 MHz	869.0	995.0	Feb-03	41%	76%
284	Dell Prec. WorkSt. 530/2.6 GHz P4 Xeon	860.0	911.0	Sep-02	41%	69%
285	IBM eServer pSeries 630/6C4-1 1.2GHz	852.0	720.0	Apr-03	40%	55%
286	IBM eServer pSeries 630/6E4-1 1200MH	852.0	720.0	Apr-03	40%	55%
287	Intel D850EMV2 (2.26GHz, Pentium4)	846.0	847.0	Aug-02	40%	65%
288	Sun Fire V1280 / 900MHz Cu	844.0	535.0	Feb-03	40%	41%
289	AMD ASUS A7N8X XP 2800+ / 2250 MHz	843.0	933.0	Sep-02	40%	71%
290	Sun Fire V210 / 1002MHz Cu	841.0	555.0	Mar-03	40%	42%
291	Sun Fire V240 / 1002MHz Cu	836.0	553.0	Mar-03	40%	42%
292	AMD ASUS A7N8X XP 2700+ / 2166 MHz	831.0	913.0	Sep-02	39%	70%
293	Dell Prec. WorkSt. 530/2.4 GHz P4 Xeon	827.0	859.0	Sep-02	39%	65%
294	Sun Blade Model 2050 / 1050 MHz	827.0	610.0	Nov-01	39%	46%
295	Dell Prec. WorkSt. 340/2.2 GHz P4	818.0	833.0	Sep-02	39%	63%
296	Sun Blade Model 2000 / 1050 MHz	817.0	-	Apr-02	39%	-
297	Intel D850MD (2.4 GHz, Pentium4)	812.0	833.0	Apr-02	39%	63%
298	Dell PowerEdge 2650/2.8 GHz P4 Xeon	810.0	907.0	Oct-02	38%	69%
299	Intel D850EMV2 (2.2 GHz, Pentium4)	808.0	811.0	Aug-02	38%	62%
300	Dell Prec. WorkSt. 530/2.2 GHz P4	795.0	810.0	Sep-02	38%	62%
301	Sun Fire 280R / 1015 MHz Cu	785.0	574.0	Nov-02	37%	44%
302	Compaq AlphaServer DS20E Model 68/833	784.0	571.0	Jun-01	37%	44%
303	Intel D850EMVR (2.0A GHz, Pentium4)	782.0	828.0	Apr-03	37%	63%
304	Dell Prec. WorkSt. 340/2.0 GHz P4	779.0	779.0	Sep-02	37%	59%
305	Intel D850MD (2.2 GHz, Pentium4)	777.0	784.0	Jan-02	37%	60%
306	Compaq AlphaServer ES40 Model 6/833	777.0	565.0	Jun-01	37%	43%
307	Sun Blade Model 2000 / 1015 MHz	775.0	576.0	Oct-02	37%	44%
308	Intel D850EMV2 (2.0A GHz, Pentium4)	773.0	820.0	Aug-02	37%	63%
309	Dell Prec. WorkSt. 530/2.0 GHz P4	758.0	757.0	Sep-02	36%	58%
310	Compaq AlphaServer GS160 Model 16 68/1001	756.0	621.0	Jun-01	36%	47%
311	Compaq AlphaServer GS320 Model 32 68/1001	756.0	621.0	Jun-01	36%	47%
312	Compaq AlphaServer GS80 Model 8 68/1001	756.0	621.0	Jun-01	36%	47%
313	Dell PowerEdge 2650/2.4 GHz P4 Xeon	750.0	815.0	Apr-02	36%	62%
314	Intel D850EMV2 (2.0 GHz, Pentium4)	745.0	684.0	Aug-02	35%	52%
315	Intel D850MD (2.0AGHz, Pentium4)	743.0	735.0	Jan-02	35%	56%
316	Sun Blade 1000 Model 900 / 900MHz Cu	731.0	533.0	Nov-01	35%	41%
317	Sun Fire V480 / 900 MHz Cu	727.0	531.0	Jun-02	34%	40%
318	Intel D850EMV2 (1.9 GHz, Pentium4)	726.0	660.0	Aug-02	34%	50%
319	Compaq AlphaServer DS20L Model 68/833	724.0	537.0	Apr-02	34%	41%
320	Dell Prec. WorkSt. 530/1.8 GHz P4	717.0	703.0	Sep-02	34%	54%
321	Intel D850GB (2.0 GHz Pentium4)	714.0	656.0	Aug-01	34%	50%
322	Intel D850MD (2.0 GHz, Pentium4)	713.0	653.0	Jan-02	34%	50%
323	Sun Netra 20 / 900 MHz Cu	713.0	533.0	May-02	34%	41%
324	Dell Prec. WorkSt. 340/1.8 GHz P4	711.0	650.0	Sep-02	34%	50%
325	AMD Epox 8KHA+ XP2600+ / 2133 MHz	710.0	839.0	Aug-02	34%	64%
326	Intel D850EMV2 (1.8 GHz, Pentium4)	708.0	638.0	Aug-02	34%	49%
327	Dell PowerEdge 7150 Itanium/800 4MB L3	703.0	-	Jun-01	33%	-
328	HP SERVER RX4610 Itanium/800 4MB L3	701.0	379.0	May-01	33%	29%
329	Sun Fire 280R / 900 MHz Cu	700.0	529.0	Nov-01	33%	40%
330	Dell PowerEdge 6650/2.0 GHz Xeon MP	699.0	842.0	Oct-02	33%	64%
331	Intel D850MD (1.9 GHz, Pentium4)	698.0	633.0	Jan-02	33%	48%
332	Sun Fire V880 / 900 MHz Cu	697.0	507.0	May-02	33%	39%
333	Intel D850GB (1.9 GHz Pentium4)	696.0	634.0	Aug-01	33%	48%
334	AMD Epox 8KHA+ XP2400+ / 2000 MHz	693.0	808.0	Aug-02	33%	62%
335	Dell PowerEdge 4600/2.0 GHz P4 Xeon	693.0	701.0	Jan-02	33%	53%
336	Intel D850EMV2 (1.7 GHz, Pentium4)	687.0	613.0	Aug-02	33%	47%
337	Intel D850MD (1.8 GHz, Pentium4)	679.0	612.0	Jan-02	32%	47%
338	Intel D850GB (1.8 GHz Pentium4)	678.0	613.0	Aug-01	32%	47%
339	Dell Prec. WorkSt. 340/1.7 GHz P4	675.0	596.0	Oct-01	32%	45%
340	HP 9000 Model C3750 / PA8700+/875	674.0	678.0	Oct-02	32%	52%
341	Dell Prec. WorkSt. 530/1.7 GHz P4	672.0	592.0	Oct-01	32%	45%
342	AMD Epox 8KHA+ XP2200+ / 1800 MHz	671.0	765.0	Jul-02	32%	58%
343	Dell Prec. WorkSt. 330/2.0 GHz P4	669.0	659.0	Aug-01	32%	50%
344	Intel D850EMV2 (1.6 GHz, Pentium4)	666.0	588.0	Aug-02	32%	45%
345	AMD Epox 8KHA+ XP2100+ / 1733 MHz	660.0	749.0	Apr-02	31%	57%
346	Intel D850GB (1.7 GHz Pentium4)	659.0	587.0	Aug-01	31%	45%
347	Intel D850MD (1.7 GHz, Pentium4)	659.0	588.0	Jan-02	31%	45%
348	AMD Asus A7M266-D MP2400+ / 2000 MHz	656.0	766.0	Nov-02	31%	58%
349	HP i2000 Itanium/800 2ML3	655.0	-	May-01	31%	-
350	HP 9000 Model J6750 / PA8700+/875	651.0	676.0	Oct-02	31%	52%
351	AMD MSI K7D M-L MP2600+ / 2133 MHz	650.0	781.0	Feb-03	31%	60%
352	Dell Prec. WorkSt. 730/Itanium/800 2ML3	645.0	314.0	Jul-01	31%	24%
353	PrimePower650 (SPARC64 GP/810MHz)	644.0	624.0	Jul-02	31%	48%
354	Fujitsu PrimePower650 (SPARC64 GP/810MHz)	644.0	624.0	Jul-02	31%	48%
355	API UP2000 833 MHz	644.0	533.0	Jun-00	31%	41%
356	AMD MSI K7D M-L MP2200+ / 1800 MHz	644.0	726.0	Aug-02	31%	55%
357	Intel D850EMV2 (1.5 GHz, Pentium4)	643.0	562.0	Aug-02	31%	43%
358	AMD Epox 8KHA+ XP2000+ / 1667 MHz	642.0	724.0	Jan-02	30%	55%
359	AMD MSI K7D M-L MP2100+ / 1733 MHz	637.0	713.0	Jun-02	30%	54%
360	Intel D850GB (1.6 GHz Pentium4)	637.0	565.0	Aug-01	30%	43%
361	Intel D850MD (1.6 GHz, Pentium4)	637.0	565.0	Jan-02	30%	43%
362	AMD Epox 8KHA+ XP1800 / 1600 MHz	634.0	701.0	Oct-01	30%	53%
363	Dell Prec. WorkSt. 330/1.8 GHz P4	631.0	619.0	Jul-01	30%	47%
364	Dell Prec. WorkSt. 340/1.5 GHz P4	629.0	547.0	Oct-01	30%	42%
365	Dell Prec. WorkSt. 530/1.5 GHz P4	629.0	545.0	Oct-01	30%	42%
366	HP i2000 Itanium/733 2ML3	623.0	-	May-01	30%	-
367	HP SERVER RX4610 Itanium/733 2ML3	623.0	-	May-01	30%	-
368	Intel D850MD (1.5 GHz, Pentium4)	616.0	539.0	Jan-02	29%	41%
369	Intel D850GB (1.5 GHz Pentium4)	615.0	539.0	Aug-01	29%	41%
370	AMD Epox 8KHA+ XP1800 / 1533 MHz	615.0	671.0	Oct-01	29%	51%
371	PrimePower850 (SPARC64 GP/810MHz)	613.0	617.0	Jul-02	29%	47%
372	Fujitsu PrimePower850 (SPARC64 GP/810MHz)	613.0	617.0	Jul-02	29%	47%
373	HP i2000 Itanium/800 4MB L3	610.0	365.0	Sep-01	29%	28%
374	Dell Prec. WorkSt. 330/1.7 GHz P4	607.0	591.0	May-01	29%	45%
375	AMD Epox 8KHA+ XP1700 / 1466 MHz	604.0	656.0	Oct-01	29%	50%
376	AMD Asus A7M266-D MP2000+ / 1667 MHz	596.0	662.0	Mar-02	28%	50%
377	Intel D850GB (1.4 GHz Pentium4)	590.0	512.0	Aug-01	28%	39%
378	AMD GA-7DX Motherboard, XP1800+/1533 M	588.0	644.0	Oct-01	28%	49%
379	Compaq AlphaServer DS20E Model 6/667	582.0	455.0	Aug-01	28%	35%
380	HP 9000 Model J6700 / PA8700-750	581.0	603.0	Apr-01	28%	46%
381	AMD GA-7DX Motherboard, XP1700+/1466 M	580.0	629.0	Oct-01	28%	48%
382	HP i2000 Itanium/733 4MB L3	577.0	-	Sep-01	27%	-
383	HP 9000 Model C3700 / PA8700-750	576.0	604.0	Apr-01	27%	46%
384	Intel D850GB (1.3 GHz Pentium4)	565.0	486.0	Aug-01	27%	37%
385	Dell Prec. WorkSt. 330/1.5 GHz P4	564.0	543.0	May-01	27%	41%
386	Compaq AlphaServer ES40 Model 6/667	562.0	433.0	Nov-99	27%	33%
387	Fujitsu CELSIUS 460/1.5 GHz P4 (400MHz Bus)	561.0	535.0	Dec-00	27%	41%
388	HP ProLiant BL10e G2 1GHz,ULV PentiumM	552.0	687.0	Jul-03	26%	52%
389	Fujitsu PrimePower800/1000/2000 (788MHz)	551.0	591.0	Jul-02	26%	45%
390	PrimePower800/1000/2000 (788MHz)	551.0	591.0	Jul-02	26%	45%
391	AMD Tyan Thunder K7 MP1800+/1533 MHz	547.0	609.0	Oct-01	26%	46%
392	AMD GA-7DX Motherboard, XP1600+/1400 M	547.0	595.0	Oct-01	26%	45%
393	HP C3650/PA8700-625	542.0	508.0	Dec-01	26%	39%
394	Dell Prec. WorkSt. 330/1.4 GHz P4	541.0	517.0	May-01	26%	39%
395	Dell Prec. WorkSt. 530/1.4 GHz P4	540.0	516.0	May-01	26%	39%
396	Compaq Evo Workst. W6000/1.4 GHz P4 Xeon	538.0	509.0	Sep-01	26%	39%
397	AMD GA-7DX Motherboard, XP1500+/1333 M	536.0	577.0	Oct-01	25%	44%
398	API UP2000 750 MHz	533.0	456.0	Jun-00	25%	35%
399	Compaq AlphaStation XP1000 Model 6/667	532.0	403.0	Sep-01	25%	31%
400	SGI Origin 3200 1X 600MHz R14k	529.0	500.0	Feb-02	25%	38%
401	AMD Tyan Thunder K7 MP1600+/1400 MHz	526.0	571.0	Oct-01	25%	44%
402	HP SERVER RP7400/PA8700-750	524.0	551.0	Sep-01	25%	42%
403	HP SERVER RP5470/PA8700-750	522.0	549.0	Oct-01	25%	42%
404	AMD Tyan Thunder K7 MP1500+/1333 MHz	516.0	554.0	Oct-01	24%	42%
405	Dell Prec. WorkSt. 330/1.3 GHz P4	516.0	490.0	May-01	24%	37%
406	Fujitsu PrimePower650 (SPARC64 GP/675MHz)	509.0	478.0	Oct-01	24%	36%
407	PrimePower650 (SPARC64 GP/675MHz)	509.0	478.0	Oct-01	24%	36%
408	PrimePower400 (SPARC64 GP/700MHz)	506.0	521.0	Jul-02	24%	40%
409	Fujitsu PrimePower400 (SPARC64 GP/700MHz)	506.0	521.0	Jul-02	24%	40%
410	PrimePower200 SPARC64 GP/700MHz-8MB	505.0	521.0	Jul-02	24%	40%
411	SGI Origin 300 1X 600MHz R14kA	495.0	483.0	Jul-02	23%	37%
412	PrimePower850 (SPARC64 GP/675MHz)	493.0	476.0	Oct-01	23%	36%
413	Fujitsu PrimePower850 (SPARC64 GP/675MHz)	493.0	476.0	Oct-01	23%	36%
414	HP SERVER RP7400/PA8700-650	489.0	493.0	Oct-01	23%	38%
415	HP SERVER RP5470/PA8700-650	488.0	491.0	Oct-01	23%	37%
416	AMD Tyan Thunder K7 1.2GHz Athlon MP	481.0	522.0	May-01	23%	40%
417	PrimePower200 SPARC64 GP/700MHz-4MB	476.0	491.0	Jul-02	23%	37%
418	Fujitsu PrimePower200 (SPARC64 GP/700MHz)	476.0	-	Jul-02	23%	-
419	SGI Origin 3200 1X 500MHz R14k	463.0	427.0	Apr-01	22%	33%
420	AMD Gigabyte GA-7DX 1.4 GHz Athlon	458.0	554.0	May-01	22%	42%
421	Dell PowerEdge 1500SC/1.4 GHz PIII	456.0	664.0	Jan-02	22%	51%
422	PrimePower800/1000/2000 (675MHz)	456.0	475.0	Sep-01	22%	36%
423	Fujitsu PrimePower800/1000/2000 (675MHz)	456.0	475.0	Sep-01	22%	36%
424	AMD Gigabyte GA-7DX 1.33GHz Athlon	445.0	539.0	Mar-01	21%	41%
425	Compaq AlphaServer GS320 Model 6/731	444.0	397.0	Jun-00	21%	30%
426	Compaq AlphaServer GS80 Model 6/731	444.0	397.0	Dec-00	21%	30%
427	Compaq AlphaServer GS160 Model 6/731	444.0	397.0	May-00	21%	30%
428	HP 9000 Model B2600 / PA8600-500	440.0	403.0	Apr-01	21%	31%
429	Dell PowerEdge 1500SC/1.26GHz PIII	440.0	623.0	Jan-02	21%	47%
430	RLX 1200i (1.20 GHz Pentium III)	437.0	575.0	Sep-02	21%	44%
431	IBM RS/6000 44P-170 (450 MHz)	434.0	346.0	Jun-02	21%	26%
432	IBM eServer pSeries 610/6E1 450MHz	433.0	334.0	Oct-01	21%	25%
433	HP 9000 model J6000 / PA8600-552 MHz	433.0	441.0	Oct-00	21%	34%
434	HP 9000 model C3600 / PA8600-552 MHz	433.0	432.0	Oct-00	21%	33%
435	IBM RS/6000 44P-270 (450MHz, 8MBL2)	433.0	334.0	Oct-01	21%	25%
436	IBM RS/6000 SP-450MHz T/W (1 CPU)	433.0	334.0	Jan-02	21%	25%
437	IBM eServer pSeries 640/B80 450MHz	433.0	334.0	Oct-01	21%	25%
438	IBM eServer pSeries 610/6C1 450MHz	433.0	334.0	Oct-01	21%	25%
439	PrimePower400 (SPARC64 GP/600MHz)	428.0	424.0	Aug-01	20%	32%
440	Fujitsu PrimePower400 (SPARC64 GP/600MHz)	428.0	424.0	Aug-01	20%	32%
441	Compaq AlphaServer DS20 Model 6/500	422.0	313.0	Nov-99	20%	24%
442	Compaq AlphaServer ES40 Model 6/500	419.0	311.0	Nov-00	20%	24%
443	AMD Gigabyte GA-7DX 1.2GHz Athlon	417.0	496.0	Feb-01	20%	38%
444	PrimePower600 (SPARC64 GP/600MHz)	416.0	420.0	Aug-01	20%	32%
445	Fujitsu PrimePower600 (SPARC64 GP/600MHz)	416.0	420.0	Aug-01	20%	32%
446	Sun Netra 20 / 750 MHz	415.0	417.0	Nov-01	20%	32%
447	HP 9000 Model A500 / PA8600-550 MHz	414.0	422.0	Nov-00	20%	32%
448	Compaq AlphaServer DS10 6/600	411.0	364.0	Jun-00	19%	28%
449	IBM eServer pSeries 620/6F1 750 MHz	410.0	458.0	Apr-02	19%	35%
450	IBM eServer pSeries 660/6H0 750 MHz	410.0	458.0	Apr-02	19%	35%
451	IBM eServer pSeries 660/6H1 750 MHz	410.0	458.0	Apr-02	19%	35%
452	IBM eServer pSeries 620/6F0 750 MHz	410.0	458.0	Apr-02	19%	35%
453	Sun Blade 1000 Model 1900 / 900MHz	410.0	466.0	Jul-01	19%	36%
454	SGI Origin 3200 400MHz R12k	407.0	353.0	Sep-00	19%	27%
455	Fujitsu PrimePower200 (SPARC64 GP/600MHz)	403.0	406.0	Aug-01	19%	31%
456	PrimePower200 (SPARC64 GP/600MHz)	403.0	406.0	Aug-01	19%	31%
457	Dell PowerEdge 2550/1.13GHz PIII	402.0	568.0	Jan-02	19%	43%
458	IBM IntelliStation 265 (450MHz, 4MBL2)	401.0	318.0	Jan-02	19%	24%
459	Sun Blade 1000 Model 1750 / 750MHz	395.0	396.0	Jun-01	19%	30%
460	Sun Fire 280R / 750 MHz	392.0	394.0	Jun-01	19%	30%
461	SGI 2200 1X 500MHz R14k	386.0	412.0	Jul-01	18%	31%
462	IBM RS/6000 SP-375MHz T/W (1 CPU)	382.0	260.0	Jan-00	18%	20%
463	IBM eServer pSeries 640 (8 MB L2)	378.0	273.0	Sep-00	18%	21%
464	Sun Fire V880 / 750 MHz	378.0	390.0	Nov-01	18%	30%
465	SGI Origin 300 1X 500MHz R14k	378.0	379.0	Oct-01	18%	29%
466	IBM RS/6000 44P-270 (1 CPU, 8MB L2)	378.0	273.0	Sep-00	18%	21%
467	HP 9000 Model L3000 / PA8600-550	376.0	388.0	Dec-00	18%	30%
468	IBM eServer pSeries 660/6M1 750 MHz	376.0	439.0	Aug-01	18%	33%
469	IBM IBM RS/6000 Model M80 (750 MHz)	376.0	439.0	Aug-01	18%	33%
470	AMD ASUS A7V 1.3 GHz Athlon	374.0	491.0	Mar-01	18%	37%
471	IBM eServer pSeries 610/6E1 375MHz	372.0	277.0	Oct-01	18%	21%
472	IBM eServer pSeries 610/6C1 375MHz	372.0	277.0	Oct-01	18%	21%
473	HP 9000 Model N4000 / PA8600-552 MHz	369.0	379.0	Mar-00	18%	29%
474	IBM RS/6000 44P-270 (1 CPU, 4MB L2)	366.0	262.0	Sep-00	17%	20%
475	IBM eServer pSeries 640 (4 MB L2)	366.0	262.0	Sep-00	17%	20%
476	RLX 800i (800 MHz Pentium III)	363.0	435.0	Sep-02	17%	33%
477	IBM RS/6000 44P-270 (1 CPU)	360.0	251.0	Jan-00	17%	19%
478	IBM RS/6000 44P-170 (400 MHz)	359.0	280.0	Jun-02	17%	21%
479	HP 9000 model B2000 / PA8500-400 MHz	357.0	332.0	Oct-00	17%	25%
480	AMD ASUS A7V 1.2 GHz Athlon	352.0	458.0	Feb-01	17%	35%
481	SGI 2200 2X 400MHz R12k	343.0	347.0	May-00	16%	26%
482	AMD Gigabyte GA-7ZM motherboard 1.2 GH	342.0	-	Oct-00	16%	-
483	Dell Prec. WorkSt. 420/1.0 GHz P3	340.0	462.0	Nov-00	16%	35%
484	IBM RS/6000 SP-375MHz High Node(1 CPU)	337.0	252.0	Jun-00	16%	19%
485	Intel OR840 (1.0 GHz Pentium III)	335.0	442.0	Jul-00	16%	34%
486	IBM eServer pSeries 610/6C1 333MHz	333.0	241.0	Feb-02	16%	18%
487	IBM eServer pSeries 610/6E1 333MHz	333.0	241.0	Feb-02	16%	18%
488	AMD Gigabyte GA-7ZM motherboard 1.1 GH	331.0	-	Aug-00	16%	-
489	Intel VC820 (1.13GHz Pentium III)	331.0	464.0	Aug-00	16%	35%
490	Dell Prec. WorkSt. 420/933MHz P3	328.0	440.0	Nov-00	16%	34%
491	AMD ASUS A7V 1.0 GHz Athlon	321.0	-	Feb-01	15%	-
492	Intel VC820 (1.0BGHz, Pentium III)	318.0	448.0	Nov-00	15%	34%
493	Dell Prec. WorkSt. 420/866MHz P3	317.0	417.0	Nov-00	15%	32%
494	Dell PowerEdge 4400/1.0 GHz PIII Xeon	317.0	451.0	Jan-02	15%	34%
495	Intel VC820 (1.0 GHz Pentium III)	314.0	428.0	Jul-00	15%	33%
496	Sun Blade 1000 Model 1750 / 750MHz	312.0	396.0	Oct-01	15%	30%
497	Intel D815EEA2 (1.0BGHz Pentium III)	310.0	457.0	Aug-01	15%	35%
498	Intel VC820 (933 MHz, Pentium III)	309.0	429.0	Nov-00	15%	33%
499	Dell Prec. WorkSt. 420/800MHz P3	306.0	399.0	Nov-00	15%	30%
500	Intel VC820 (933 MHz Pentium III)	305.0	410.0	Jul-00	14%	31%
501	Intel D815EEA2 (933 MHz Pentium III)	300.0	428.0	Aug-01	14%	33%
502	Intel VC820 (866 MHz, Pentium III)	297.0	407.0	Nov-00	14%	31%
503	HP 9000 Model N4000 / PA8600-440 MHz	297.0	-	Nov-99	14%	-
504	Intel VC820 (866 MHz Pentium III)	294.0	390.0	Jul-00	14%	30%
505	Dell PowerEdge 6400/PIII Xeon 700	294.0	438.0	Apr-01	14%	33%
506	SGI 2100 1X 350MHz R12k	293.0	289.0	Mar-01	14%	22%
507	Intel D815EEA2 (866 MHz Pentium III)	292.0	417.0	Aug-01	14%	32%
508	Intel D815EEA2 1.0B GHz Pentium III)	292.0	454.0	Nov-04	14%	35%
509	Sun Enterprise 450 / 480MHz	291.0	234.0	Aug-00	14%	18%
510	Dell Prec. WorkSt. 420/733MHz P3	290.0	374.0	Nov-00	14%	29%
511	SGI Origin200 360MHz R12k	290.0	298.0	Jun-00	14%	23%
512	Sun Blade 1000 Model 1600 / 600MHz	288.0	311.0	Jul-01	14%	24%
513	Intel VC820 (800EB MHz, Pentium III)	286.0	386.0	Nov-00	14%	29%
514	Intel VC820 (800 MHz Pentium III)	282.0	371.0	Jul-00	13%	28%
515	Intel D815EEA2 (800 MHz Pentium III)	280.0	386.0	Aug-01	13%	29%
516	Sun Fire 6800 / 750 MHz	278.0	360.0	Oct-01	13%	27%
517	IBM RS/6000 44P-170 (333 MHz)	277.0	202.0	Jun-02	13%	15%
518	Sun Blade 150 (UltraSPARC IIi/650)	276.0	246.0	Aug-02	13%	19%
519	Intel VC820 (850 MHz Pentium III)	268.0	367.0	Jul-00	13%	28%
520	Intel D815EEA2 (1.1 GHz Pentium III)	268.0	427.0	Aug-01	13%	33%
521	Sun Enterp. 420R (UltraSPARC-II/450)	265.0	214.0	Mar-01	13%	16%
522	Intel D815EEA2 (1.0 GHz Pentium III)	264.0	408.0	Aug-01	13%	31%
523	Sun Enterprise 3500/4500 - 400 MHz	261.0	212.0	Oct-00	12%	16%
524	Dell PowerEdge 8450/PIII Xeon 700	260.0	422.0	Apr-01	12%	32%
525	Intel VC820 (800EB MHz Pentium III)	256.0	355.0	Mar-00	12%	27%
526	Sun Blade 150 (UltraSPARC IIi/550)	254.0	217.0	Aug-02	12%	17%
527	IBM eServer pSeries 620/6F1 600 MHz	252.0	310.0	Apr-01	12%	24%
528	IBM eServer pSeries 660/6H1 600 MHz	252.0	310.0	Apr-01	12%	24%
529	IBM eServer pSeries 620/6F0 600 MHz	252.0	310.0	May-01	12%	24%
530	IBM eServer pSeries 660/6H0 600 MHz	252.0	310.0	May-01	12%	24%
531	IBM RS/6000 Model 7026-M80 (1 CPU)	250.0	275.0	May-00	12%	21%
532	Intel VC820 (733 MHz Pentium III)	244.0	335.0	Mar-00	12%	26%
533	Intel OR840 (733MHz Pentium III)	243.0	336.0	Nov-99	12%	26%
534	Fujitsu CELSIUS 650 / 733 MHz Pentium III	242.0	337.0	Jan-00	11%	26%
535	Intel SE440BX-2 (800 MHz Pentium III	237.0	344.0	Mar-00	11%	26%
536	Intel VC820 (667 MHz Pentium III)	233.0	314.0	Mar-00	11%	24%
537	Intel SE440BX-2 (750 MHz Pentium III	230.0	330.0	Mar-00	11%	25%
538	Intel SE440BX-2 (700 MHz Pentium III	223.0	315.0	Mar-00	11%	24%
539	Intel SE440BX-2 (650 MHz Pentium III	215.0	299.0	Mar-00	10%	23%
540	IBM RS/6000 Model 7026-H80 (1 CPU)	210.0	234.0	May-00	10%	18%
541	IBM RS/6000 Model 7026-F80 (1 CPU)	210.0	234.0	May-00	10%	18%
542	Dell Prec. WorkSt. 410/700MHz PIII	205.0	307.0	Nov-99	10%	23%
543	Intel SE440BX-2 (500 MHz Pentium III	193.0	233.0	Feb-02	9%	18%
544	Sun Blade 100 (UltraSPARC-IIe/500)	182.0	174.0	Mar-01	9%	13%
545	IBM RS/6000 7043-260 (1 CPU)	180.0	-	Nov-99	9%	-
546	Intel SE440BX-2 (450 MHz Pentium III	178.0	213.0	Mar-02	8%	16%
547	Compaq AlphaServer 4100 5/533	176.0	176.0	Nov-99	8%	13%
548	Compaq DIGITAL PW 500au	158.0	161.0	Nov-99	7%	12%
549	Compaq AlphaStation 500/500	153.0	163.0	Nov-99	7%	12%
550	Ultra 10 333MHz	126.0	133.0	Nov-99	6%	10%
551	IBM RS/6000 43P-150 (250MHz,xlf90)	90.8	105.0	Jun-00	4%	8%
552	IBM RS/6000 43P-150 (250MHz,xlf)	84.4	93.7	May-00	4%	7%

(*) Values derived under earlier releases of OS/compilers/libraries
(+) SPECfp_base2000 value




Table 2. SPEC CPU2000 - SPECfp_base and SPECint. Absolute and Relative Values.

SPEC CPU2000 - SPECfp_base and SPECint. Absolute Values and Values Relative to the HP RX5670 Itanium2/1500-H.

Rank	Machine	SPECfp	SPEC	Test	Relative Values (%)
		_base	int	Date	SPECfp	SPEC
					_base	int
1	HP Integrity RX4640/Itanium2 1.6G/9MB	2712.0	1590.0	Nov-04	129%	121%
2	HP Integrity RX1620/Itanium2 1.6G/3MB	2692.0	1452.0	Nov-04	128%	111%
3	HP Integrity RX2620/Itanium2 1.6G/6MB	2675.0	1535.0	Dec-04	127%	117%
4	SGI Altix 3700 Bx2/Itanium2 1.6G/9MB	2647.0	1441.0	Nov-04	126%	110%
5	SGI Altix 3700 Bx2/Itanium2 1.6G/6MB	2600.0	1410.0	Nov-04	123%	107%
6	IBM eServer p5 570 (1900 MHz) 2CC	2576.0	1452.0	Jul-04	122%	111%
7	Bull NovaScale 4040 Itanium2/1.6Hz 9MBL3	2566.0	1457.0	Nov-04	122%	111%
8	HP Integrity RX2620/Itanium2 1.6G/3MB	2553.0	1408.0	Nov-04	121%	107%
9	Dell PowerEdge 7250 Itanium2/1600 9MB L3	2545.0	1403.0	Nov-04	121%	107%
10	HP Integrity RX4640/Itanium2 1.5G/4MB	2502.0	1372.0	Nov-04	119%	105%
11	SGI Altix 3700 Bx2/Itanium2 1.5G/4MB	2361.0	1242.0	Dec-04	112%	95%
12	HP Integrity RX2620/Itanium2 1.3G/3MB	2229.0	1170.0	Dec-04	106%	89%
13	IBM eServer p5 590 (1650 MHz, 1 CPU)	2223.0	1259.0	Oct-04	105%	96%
14	HP Integrity RX1620/Itanium2 1.3G/3MB	2214.0	1178.0	Dec-04	105%	90%
15	HP Integrity RX4640/Itanium2,1.5GHz	2161.0	1404.0	Jan-04	103%	107%
16	SGI Altix 3000 (1.5GHz, Itanium 2)	2130.0	1243.0	Mar-04	101%	95%
17	IBM eServer i5 550 (1650 MHz) 2CC	2121.0	1248.0	Aug-04	101%	95%
18	IBM eServer p5 550 (1650 MHz) 2CC	2121.0	1248.0	Jul-04	101%	95%
19	HP Integrity RX2600/Itanium2,1.5GHz	2119.0	1408.0	Jun-03	101%	107%
20	HP Integrity RX5670/Itanium2,1.5GHz	2108.0	1312.0	Jun-03	100%	100%
21	HP ZX6000/Itanium2, 1.5GHz,6MB-L3, H	2106.0	1315.0	Jun-03	100%	100%
22	ION Computers I2X4 (1.5GHz Itanium2)	2038.0	1232.0	May-04	97%	94%
23	IBM eServer i5 520 (1650 MHz) 2CC	2034.0	1248.0	Aug-04	96%	95%
24	IBM eServer p5 520 (1650 MHz) 2CC	2034.0	1248.0	Jul-04	96%	95%
25	Bull NovaScale 4040 Itanium2/1.5GHz	1994.0	1232.0	Jul-03	95%	94%
26	ION Computers I2X2 (1.6GHz Itanium2)	1984.0	1178.0	May-04	94%	90%
27	IBM eServer p5 550 Express 1500MHz	1932.0	-	Oct-04	92%	-
28	SGI Altix 350/Itanium2 1.4GHz/3MB	1931.0	1078.0	May-04	92%	82%
29	IBM eServer p5 520 Express 1500MHz	1922.0	-	Oct-04	91%	-
30	HP Integrity RX4640/Itanium2,1.3GHz	1891.0	1132.0	Jan-04	90%	86%
31	Dell PowerEdge3250/Itanium2 1.5GHz/6ML3	1875.0	1099.0	Jul-03	89%	84%
32	Dell PowerEdge3250/Itanium2 1.4GHz/3ML3	1868.0	1022.0	Mar-04	89%	78%
33	IBM eServer OpenPower 720 (1650 MHz)	1865.0	1138.0	Oct-04	88%	87%
34	SGI Altix 3000 (1.3GHz, Itanium 2)	1848.0	1019.0	Mar-04	88%	78%
35	Intel D925XECV2 3.80 GHz P4	1839.0	1671.0	Nov-04	87%	127%
36	Supermicro SS6113M-i(1.5GHz/6MB, Itanium 2)	1830.0	1090.0	Dec-03	87%	83%
37	ION Computers I2X2 (1.4GHz Itanium2)	1817.0	926.0	Jul-03	86%	71%
38	HP Integrity RX5670/Itanium2,1.3GHz	1814.0	1066.0	Nov-03	86%	81%
39	HP Integrity RX2600/Itanium2,1.3GHz	1808.0	1073.0	Nov-03	86%	82%
40	AMD MSI K8N Neo2 Athlon64/2.6GHz FX-55	1741.0	1854.0	Sep-04	83%	141%
41	Intel D925XECV2 3.46 GHz P4	1719.0	1772.0	Nov-04	82%	135%
42	Fujitsu CELSIUS V810 / 2.4GHz Opteron 250 L	1706.0	1619.0	Nov-04	81%	123%
43	IBM xSeries 336 (3.6GHz Xeon)	1700.0	-	Sep-04	81%	-
44	AMD MSI K8N Neo2 Athlon64/2.6GHz FX-55	1695.0	1735.0	Nov-04	80%	132%
45	SGI Altix 350/Itanium2 1.4GHz/1.5MB	1668.0	986.0	Dec-03	79%	75%
46	IBM eServer 326 2.4 GHz Opteron EKO	1648.0	1584.0	Oct-04	78%	121%
47	IBM eServer pSeries655 (1.7GHz/1CPU)	1642.0	1158.0	Jan-04	78%	88%
48	Sun Java Wrkst. W1100z Opteron2.4GHz	1637.0	1582.0	Jul-04	78%	121%
49	Supermicro Super P8SC8 (3.80 GHz Pentium 4)	1633.0	1639.0	Nov-04	77%	125%
50	Intel D925XCV (AA-400 3.6GHz,P4)	1627.0	1575.0	Jun-04	77%	120%
51	IBM xSeries 336 (3.6GHz,800MHz FSB)	1625.0	-	Aug-04	77%	-
52	IBM eServer pSeries 690Turbo 1.7 GHz	1598.0	1113.0	May-03	76%	85%
53	AMD MSI K8T Neo2-FIR Athlon64 2.4GHz	1595.0	1717.0	Sep-04	76%	131%
54	HP ProLiant BL20p G3 3.6GHz Xeon	1593.0	1543.0	Nov-04	76%	118%
55	AMD ASUS SK8V/Opteron150 2.4GHz SES	1591.0	1487.0	May-04	75%	113%
56	Intel D925XCV (AA-400 3.4GHz,P4)	1578.0	1705.0	Jun-04	75%	130%
57	IBM eServer pSeries 655 1700 MHz	1576.0	1103.0	May-03	75%	84%
58	Intel D925XECV2 3.0E GHz P4	1568.0	1369.0	Nov-04	74%	104%
59	Sun Fire V20z (Opteron250/2.4GHz)	1555.0	1521.0	Jul-04	74%	116%
60	PathScale ASUS SK8N/Opteron248 2.2G	1553.0	1452.0	Apr-04	74%	111%
61	Sun Fire V40z (Opteron850/2.4GHz)	1551.0	1515.0	Jul-04	74%	115%
62	Intel D875PBZ (AA-301 3.4GHz,P4)	1548.0	-	Feb-04	73%	-
63	Dell Prec. Workst. 360/3.4GHz P4 Extreme	1544.0	1689.0	Jun-04	73%	129%
64	Intel D875PBZ (AA-206 3.4GHz P4)	1544.0	1705.0	Feb-04	73%	130%
65	Supermicro P8SAA ( 3.6 GHz Pentium 4, 1MBL2)	1537.0	1560.0	Nov-04	73%	119%
66	HP ProLiant DL145 Opteron250/2.4GHz	1534.0	1561.0	Oct-04	73%	119%
67	ION Computers SR1325TP (3.4GHz Pentium 4 H	1524.0	1671.0	May-04	72%	127%
68	Fujitsu CELSIUS M420/3.4EGHz P4(800MHz Bus)	1519.0	1501.0	Jun-04	72%	114%
69	PrimePower900 (SPARC64 GP/1890MHz)	1510.0	1345.0	May-04	72%	103%
70	PrimePower650 (SPARC64 GP/1890MHz)	1510.0	1345.0	May-04	72%	103%
71	Fujitsu CELSIUS M430/3.4FGHz P4(800MHz Bus)	1507.0	1484.0	Aug-04	71%	113%
72	AMD ASUS SK8N/Opteron148 2.2GHz SES	1505.0	1409.0	Nov-03	71%	107%
73	HP ProLiant ML370 G4 3.6GHz Xeon	1504.0	1556.0	Oct-04	71%	119%
74	HP ProLiant ML350 G4 3.4GHz Xeon	1503.0	1452.0	Nov-04	71%	111%
75	Dell PowerEdge 1850/3.6GHz P4Xeon	1502.0	1538.0	Aug-04	71%	117%
76	Intel D925XECV2 2.8E GHz P4	1496.0	1290.0	Nov-04	71%	98%
77	HP ProLiant DL380 G4 3.6GHz Xeon,L3	1495.0	1553.0	Jul-04	71%	118%
78	AMD Tyan K8QS Opteron850/2.4GHz	1492.0	1663.0	Sep-04	71%	127%
79	Dell Prec. Workst. 360/3.40E GHz P4	1489.0	1484.0	Jun-04	71%	113%
80	Dell PowerEdgeSC 1425 (3.6 GHz Xeon)	1485.0	1541.0	Nov-04	70%	117%
81	Dell PowerEdge 1855/3.6GHz Xeon	1485.0	1540.0	Nov-04	70%	117%
82	Dell Prec. Workst. 360/3.2GHz P4 Extreme	1482.0	1601.0	Feb-04	70%	122%
83	Intel D875PBZ (AA-301:3.4EGHz,P4)	1481.0	1491.0	Apr-04	70%	114%
84	Intel D875PBZ (AA-206 3.2GHz P4)	1474.0	1620.0	Nov-03	70%	123%
85	HP ProLiant ML370 G4 3.4GHz Xeon	1470.0	1489.0	Aug-04	70%	113%
86	Intel D925XCV (AA-400 3.0EGHz,P4)	1458.0	1353.0	Jun-04	69%	103%
87	Supermicro SUPER P4SCT+II 3.4GHz P4,800MHz bus	1454.0	1493.0	Jun-04	69%	114%
88	Dell PowerEdge 2850/3.6 GHz P4Xeon	1450.0	1522.0	Aug-04	69%	116%
89	Dell PowerEdge3250/Itanium2 1.4GHz/1.5ML3	1444.0	824.0	Sep-03	69%	63%
90	AMD ASUS SK8V/Opteron150 2.4GHz WS	1439.0	1655.0	May-04	68%	126%
91	Dell Prec. Workst. 360/3.20E GHz P4	1438.0	1404.0	Jan-04	68%	107%
92	Intel D875PBZ (AA-301 3.2EGHz,P4)	1434.0	1421.0	Feb-04	68%	108%
93	IBM eServer pSeries 670 1.5 GHz	1432.0	981.0	May-03	68%	75%
94	HP RX5670/Itanium2, 1GHz,3MB-L3, L	1431.0	-	Nov-02	68%	-
95	IBM eServer 325 2.4GHz Opteron 64-PGI	1428.0	1472.0	May-04	68%	112%
96	HP RX2600/Itanium2, 1GHz,3MB-L3, L	1427.0	-	Nov-02	68%	-
97	HP ZX6000/Itanium2, 1GHz,3MB-L3, L	1422.0	-	Nov-02	67%	-
98	Fujitsu CELSIUS V810 / 2.2GHz Opteron 248 L	1408.0	1354.0	Dec-03	67%	103%
99	HP ProLiant DL380 G4 3.4GHz Xeon,L3	1407.0	1477.0	Jul-04	67%	113%
100	IBM eServer pSeries655 1700MHz/LPR/L	1405.0	1023.0	Jul-03	67%	78%
101	IBM eServer pSeries655 1700MHz/L	1401.0	1031.0	Sep-03	66%	79%
102	IBM eServer pSeries 655 1500 MHz	1398.0	970.0	May-03	66%	74%
103	SGI Altix 3000 (1GHz, Itanium 2)	1396.0	683.0	May-03	66%	52%
104	AMD ASUS SK8N/Opteron148 2.2GHz WS	1393.0	1477.0	Nov-03	66%	113%
105	Bull NovaScale 4040 Itanium2/1.0GHz	1391.0	719.0	Mar-03	66%	55%
106	Dell Prec. Workst. 360/3.00E GHz P4	1382.0	1329.0	Jan-04	66%	101%
107	HP Integrity RX1600/Itanium2,1.0GHz	1382.0	837.0	Feb-04	66%	64%
108	Intel D875PBZ (AA-301 3.0EGHz,P4)	1381.0	1345.0	Feb-04	66%	103%
109	AMD ASUS SK8N Athlon FX-51/2.2GHz(L)	1371.0	1395.0	Sep-03	65%	106%
110	IBM x306 (3.0EGHz,Pentium4,1CC)	1369.0	1336.0	Mar-04	65%	102%
111	HP ProLiant DL360 G4 3.6GHz Xeon	1364.0	1467.0	Oct-04	65%	112%
112	SGI Altix 350 (1.0GHz, Itanium 2)	1357.0	743.0	Feb-04	64%	57%
113	HP ProLiant BL30p (3.2GHz Xeon,2MB L3)	1348.0	1550.0	Nov-04	64%	118%
114	Dell Prec. WorkSt.650/3.2GHz P4Xeon 2ML3	1338.0	1563.0	Jan-04	63%	119%
115	Dell PowerEdge 2650/3.2 GHz P4Xeon,2ML3	1330.0	1420.0	Feb-04	63%	108%
116	AMD ASUS SK8N Athlon FX-51/2.2GHz(W)	1329.0	1447.0	Sep-03	63%	110%
117	HP ProLiant DL360 G4 3.4GHz Xeon,L3	1327.0	1402.0	Jul-04	63%	107%
118	Intel D875PBZ (AA-301:3.4GHz,Pentium	1326.0	-	Feb-04	63%	-
119	HP ProLiant DL360 G3 3.2GHz/Xeon,2ML3 1	1326.0	1541.0	Apr-04	63%	117%
120	Dell Prec. Workst. 360/2.80E GHz P4	1325.0	1254.0	Jan-04	63%	96%
121	Intel D875PBZ (AA-301 2.8EGHz,P4)	1325.0	1268.0	Feb-04	63%	97%
122	HP ProLiant DL380 G3 3.2GHz/Xeon,2ML3 1	1318.0	1534.0	Apr-04	63%	117%
123	Fujitsu CELSIUS V810 / 2.2GHz Opteron 248 W	1309.0	1429.0	Dec-03	62%	109%
124	HP ProLiant ML370 G3 (3.2GHz/2MB L3	1308.0	1530.0	Jun-04	62%	117%
125	AMD ASUS SK8N/Opteron146 2.0GHz HE	1306.0	1354.0	Feb-04	62%	103%
126	Intel D875PBZ (3.4 GHz, Pentium4)	1300.0	1393.0	Feb-04	62%	106%
127	Dell Prec. Workst. 360/3.40 GHz P4	1295.0	1369.0	Jan-04	61%	104%
128	AMD ASUS SK8N Opteron146 2.0GHz-I,SES	1291.0	1262.0	Sep-03	61%	96%
129	IBM x335 (3.2GHz, 533MHz FSB)	1286.0	1517.0	Feb-04	61%	116%
130	HP AlphaServer GS1280 7/1300	1279.0	994.0	Jun-04	61%	76%
131	Fujitsu Celsius M420 / 3.2GHz P4(800MHz bus)	1271.0	1261.0	Jul-03	60%	96%
132	Dell PowerEdge 6600/3.0 GHz Xeon MP	1270.0	1408.0	Feb-04	60%	107%
133	Dell Prec. Workst. 360/3.2 GHz P4 DDR400	1267.0	1249.0	Jun-03	60%	95%
134	AMD ASUS SK8N Opteron146 2.0GHz-I,WS	1250.0	1335.0	Sep-03	59%	102%
135	Intel D875PBZ (3.2 GHz, Pentium4)	1245.0	1329.0	Nov-03	59%	101%
136	ION Computers BX21 (3.2GHz Xeon/2ML3 1C)	1241.0	1495.0	May-04	59%	114%
137	Dell Prec. Workst. 360/3.0 GHz P4 DDR400	1229.0	1193.0	Jun-03	58%	91%
138	ION Computers SR2300WV2 (3.2GHz Xeon/2ML3)	1224.0	1455.0	Feb-04	58%	111%
139	IBM eServer pSeries 650/6M2 1450MHz	1221.0	935.0	Nov-02	58%	71%
140	IBM eServer 325 2.0GHz Opteron 64-gcc	1217.0	1239.0	Sep-03	58%	94%
141	Intel D875PBZ (3.0 GHz, Pentium4)	1213.0	1265.0	May-03	58%	96%
142	HP ProLiant DL580 G2 3.0GHz XeonMP 1CC	1212.0	1491.0	Mar-04	57%	114%
143	Sun Fire V440/UltraSPARC-IIi 1.6GHz	1209.0	845.0	Nov-04	57%	64%
144	Dell Prec. Workst. 360/2.80E GHz DDR533	1209.0	1211.0	Jan-04	57%	92%
145	AMD Rioworks HDAMA Opteron 246 / 2.0 G	1209.0	1317.0	Jul-03	57%	100%
146	Dell PowerEdge 2650/3.2 GHz P4Xeon,1ML3	1207.0	1306.0	Oct-03	57%	100%
147	IBM eServer pSeries 690Turbo 1.3 GHz	1202.0	839.0	Apr-02	57%	64%
148	Dell Prec. WorkSt.650/3.2GHz P4 Xeon L3	1200.0	1333.0	Sep-03	57%	102%
149	IBM eServer pSeries 655/651 1300 MHz	1200.0	848.0	Nov-02	57%	65%
150	Dell Prec. WorkSt.650/3.2GHz P4Xeon 1ML3	1194.0	1415.0	Jan-04	57%	108%
151	ION Computers BX41 (3GHz Xeon MP/4ML3 1C)	1192.0	1486.0	May-04	57%	113%
152	Dell Prec. Workst. 360/2.8 GHz P4 DDR400	1186.0	1129.0	Jun-03	56%	86%
153	Dell PowerEdge 2650/3.06 GHz P4Xeon,1ML3	1183.0	1263.0	Oct-03	56%	96%
154	AMD MSI K8 Neo Athlon64 3200+/2 GHz	1180.0	1335.0	Sep-03	56%	102%
155	Dell PowerEdge 1750/3.2 GHz P4Xeon,1ML3	1179.0	1289.0	Oct-03	56%	98%
156	IBM BladeCenter JS20 (PPC-970,2.2GHz)	1178.0	1040.0	Oct-04	56%	79%
157	HP RX5670/Itanium2, 1GHz,3MB-L3, H	1178.0	807.0	Jul-02	56%	62%
158	HP ProLiant BL20p G2 3.2GHz Xeon,1ML3	1177.0	1321.0	Oct-03	56%	101%
159	HP RX2600/Itanium2, 1GHz,3MB-L3, H	1174.0	810.0	Jul-02	56%	62%
160	Dell Prec. WorkSt.650/3.06GHz P4 Xeon L3	1173.0	1294.0	Jun-03	56%	99%
161	HP ProLiant DL380 G3 3.2GHz Xeon,L3	1172.0	1313.0	Oct-03	56%	100%
162	IBM eServer 325 2.0GHz Opteron 64-PGI	1172.0	1226.0	Jul-03	56%	93%
163	HP ProLiant DL360 G3 3.2GHz Xeon,L3	1171.0	1319.0	Oct-03	56%	101%
164	Dell Prec. WorkSt.650/3.06GHz P4Xeon 1ML3	1168.0	1368.0	Feb-04	55%	104%
165	HP ProLiant BL20p G2 3.06GHzXeon,1ML3	1160.0	1256.0	Aug-03	55%	96%
166	HP ProLiant DL360 G3 3.06GHz Xeon,L3	1154.0	1258.0	Jun-03	55%	96%
167	Dell Prec. Workst. 360/3.2 GHz P4 DDR333	1153.0	1201.0	Jun-03	55%	92%
168	HP RX5670/Itanium2,900MHz,1.5MB-L3,L	1151.0	-	Nov-02	55%	-
169	HP RX2600/Itanium2,900MHz,1.5MB-L3,L	1151.0	-	Nov-02	55%	-
170	HP ProLiant DL380 G3 3.06GHz Xeon,L3	1150.0	1258.0	Jun-03	55%	96%
171	Dell Prec. Workst. 360/2.6 GHz P4 DDR400	1141.0	1067.0	Jun-03	54%	81%
172	Dell PowerEdge 1750/3.06 GHz Xeon L3	1140.0	1243.0	Jul-03	54%	95%
173	HP ZX6000/Itanium2,900MHz,1.5MB-L3,L	1139.0	-	Nov-02	54%	-
174	HP AlphaServer ES47 7/1150	1134.0	882.0	Jun-04	54%	67%
175	HP AlphaServer ES80 7/1150	1134.0	882.0	Jun-04	54%	67%
176	IBM IntelliStation POWER 275 1.45GHz	1129.0	978.0	Jun-03	54%	75%
177	HP AlphaServer GS1280 7/1150	1124.0	877.0	Jan-03	53%	67%
178	AMD Einox A4800/Opteron144 1.8GHz-I,WS	1122.0	1170.0	Apr-03	53%	89%
179	Dell Prec. Workst. 360/3.0 GHz P4 DDR333	1122.0	1149.0	Jun-03	53%	88%
180	ION Computers SR2300WV2 (3.2GHz Xeon/1ML3,	1107.0	1330.0	Feb-04	53%	101%
181	Dell PowerEdge 6650/2.8 GHz Xeon MP	1103.0	1234.0	Jun-03	52%	94%
182	IBM eServer pSeries 630/6E4 1450MHz	1097.0	910.0	Feb-03	52%	69%
183	IBM eServer pSeries 630/6C4 1450MHz	1097.0	910.0	Feb-03	52%	69%
184	PrimePower650 (SPARC64 GP/1350MHz)	1096.0	905.0	Jun-03	52%	69%
185	AMD Einox A4800/Opteron144 1.8GHz-I,SE	1093.0	1100.0	Apr-03	52%	84%
186	Dell Prec. WorkSt. 350/3.06 GHz P4 Xeon	1092.0	1130.0	Nov-02	52%	86%
187	Dell Prec. Workst. 360/2.4 GHz P4 DDR400	1092.0	1003.0	Jun-03	52%	76%
188	Dell Prec. WorkSt.650/2.8GHz P4Xeon 1ML3	1091.0	1270.0	Jan-04	52%	97%
189	IBM eServer pSeries 650/6M2-1 1450MH	1091.0	886.0	Apr-03	52%	68%
190	Dell Prec. Workst. 360/2.8 GHz P4 DDR333	1089.0	1093.0	Jun-03	52%	83%
191	Dell Prec. Mob. Workst. M60 Pentium M 755	1087.0	1541.0	Jun-04	52%	117%
192	HP ZX2000/Itanium2,900MHz,1.5MB-L3,L	1086.0	-	Nov-02	52%	-
193	SGI Altix 3000 (900MHz, Itanium 2)	1082.0	-	Feb-03	51%	-
194	HP ProLiant BL40p (2.8GHz Xeon MP)	1081.0	1249.0	Sep-03	51%	95%
195	Intel D850EMVR(3.06GHz,Pentium4,MT)	1077.0	1107.0	Dec-02	51%	84%
196	HP ProLiant DL560 (2.8GHz Xeon MP)	1075.0	1247.0	Sep-03	51%	95%
197	PrimePower450 (SPARC64 GP/1320MHz)	1074.0	882.0	Dec-03	51%	67%
198	AMD Einox A4800/Opteron144 1.8GHz-I,SE	1070.0	1100.0	Apr-03	51%	84%
199	HP ProLiant DL580 G2 2.8GHz Xeon MP	1063.0	1233.0	Sep-03	50%	94%
200	Intel D850EMVR(3.06 GHz,Pentium4)	1059.0	1098.0	Apr-03	50%	84%
201	Dell Prec. WorkSt.650/3.06 GHz P4 Xeon	1053.0	1138.0	Apr-03	50%	87%
202	Dell Prec. Workst. 360/2.6 GHz P4 DDR333	1052.0	1034.0	Jun-03	50%	79%
203	PrimePower250 (SPARC64 GP/1320MHz)	1051.0	881.0	Dec-03	50%	67%
204	Dell Prec. WorkSt. 350/2.8 GHz P4 Xeon	1048.0	1061.0	Nov-02	50%	81%
205	PrimePower900 (SPARC64 GP/1350MHz)	1047.0	892.0	May-03	50%	68%
206	HP ProLiant BL20p G2 3.06GHz Xeon	1044.0	1075.0	Aug-03	50%	82%
207	HP ProLiant DL360 G3 3.06GHz Xeon	1038.0	1070.0	Jun-03	49%	82%
208	IBM eServer pSeries 655/651 1100 MHz	1037.0	722.0	Nov-02	49%	55%
209	HP ProLiant DL380 G3 3.06GHz Xeon	1033.0	1068.0	Jun-03	49%	81%
210	Intel D850EMVR (2.8 GHz, Pentium4)	1031.0	1041.0	Apr-03	49%	79%
211	Dell PowerEdge 1750/3.06 GHz Xeon	1030.0	1067.0	Jul-03	49%	81%
212	AMD Einox A4800/Opteron142 1.6GHz-I,WS	1029.0	1053.0	Apr-03	49%	80%
213	ION Computers SR2300WV2 (3.066GHz Xeon/512	1027.0	1139.0	Feb-04	49%	87%
214	Dell Prec. WorkSt.650/2.4GHz P4Xeon 1ML3	1026.0	1123.0	Jan-04	49%	86%
215	Dell Prec. WorkSt. 350/2.66 GHz P4 Xeon	1024.0	1026.0	Nov-02	49%	78%
216	HP AlphaServer ES45 68/1250	1019.0	928.0	Nov-02	48%	71%
217	IBM eServer pSeries 670 Pwr4/1.1 GHz	1017.0	708.0	Apr-02	48%	54%
218	Supermicro X5DPA-GG (3.2GHz, 533MHz bus)	1014.0	1270.0	Jan-04	48%	97%
219	AMD ASUS SK8N/Opteron140 1.4GHz EE	1012.0	990.0	Feb-04	48%	75%
220	Intel D850EMVR (2.67 GHz, Pentium4)	1012.0	1005.0	Oct-02	48%	77%
221	Dell Prec. Workst. 360/2.4 GHz P4 DDR333	1011.0	972.0	Jun-03	48%	74%
222	Dell Prec. WorkSt.650/2.8 GHz P4 Xeon	1010.0	1068.0	Apr-03	48%	81%
223	Sun Fire V65x / (3.06 GHz Xeon)	1010.0	1066.0	May-03	48%	81%
224	Dell PowerEdge 2650/3.06 GHz P4 Xeon	996.0	1056.0	Apr-03	47%	80%
225	Dell Prec. WorkSt. 350/2.53 GHz P4 Xeon	992.0	986.0	Nov-02	47%	75%
226	Dell Prec. WorkSt.650/2.66 GHz P4 Xeon	991.0	1033.0	Apr-03	47%	79%
227	IBM eServer pSeries 630/6E4-1 1450MH	984.0	856.0	Apr-03	47%	65%
228	IBM eServer pSeries 630/6C4-1 1.45GH	984.0	856.0	Apr-03	47%	65%
229	HP AlphaServer ES80 7/1000	975.0	761.0	Jan-03	46%	58%
230	HP AlphaServer ES47 7/1000	975.0	761.0	Jan-03	46%	58%
231	Intel D850EMVR (2.53 GHz, Pentium4)	973.0	956.0	Apr-03	46%	73%
232	Dell Prec. WorkSt. 340/3.06 GHz P4 Xeon	972.0	1074.0	Nov-02	46%	82%
233	ION Computers BX21 (2.8GHz Xeon/512kB L2,1	966.0	1050.0	May-04	46%	80%
234	IBM eServer pSeries 615/6C3 1.2 GHz	966.0	822.0	May-03	46%	63%
235	IBM eServer pSeries 615/6E3 1.2 GHz	966.0	822.0	May-03	46%	63%
236	Dell Prec. WorkSt. 350/2.4 GHz P4 Xeon	965.0	951.0	Nov-02	46%	72%
237	HP ProLiant DL140 (3.2GHz Xeon,1ML3)	962.0	1258.0	Nov-03	46%	96%
238	IBM eServer pSeries 630/6C4 1200MHz	961.0	767.0	Apr-03	46%	58%
239	IBM eServer pSeries 630/6E4 1200MHz	961.0	767.0	Apr-03	46%	58%
240	Sun Fire V250 / 1280MHz Cu	954.0	702.0	Sep-03	45%	54%
241	Sun Blade Model 2000 / 1200 MHz	953.0	722.0	Mar-03	45%	55%
242	Sun Fire V240 / 1.28GHz Cu	953.0	704.0	Jan-04	45%	54%
243	Intel D850EMVR (2.4 GHz, Pentium4)	951.0	927.0	Oct-02	45%	71%
244	Dell Prec. Workst. 360/2.26 GHz P4 Xeon	951.0	904.0	Jun-03	45%	69%
245	Sun Netra 20 / 1200 MHz	945.0	712.0	Mar-03	45%	54%
246	Sun Fire 280R / 1200 MHz Cu	945.0	712.0	Mar-03	45%	54%
247	Dell Prec. WorkSt. 340/2.8 GHz P4 Xeon	938.0	1010.0	Sep-02	44%	77%
248	Dell Prec. WorkSt. 350/2.26 GHz P4 Xeon	936.0	909.0	Nov-02	44%	69%
249	AMD Einox A4800/Opteron140 1.4GHz-I,WS	934.0	933.0	Apr-03	44%	71%
250	Dell Prec. WorkSt.650/2.4 GHz P4 Xeon	934.0	955.0	Apr-03	44%	73%
251	Sun Fire V880 / 1200 MHz Cu	923.0	700.0	Sep-03	44%	53%
252	Intel D850EMVR (2.26 GHz, Pentium4)	922.0	888.0	Oct-02	44%	68%
253	Sun Fire V480 / 1200MHz Cu	921.0	702.0	Oct-03	44%	54%
254	Supermicro X5DPA-GG (3.06GHz, 533MHz bus)	920.0	1056.0	Jan-04	44%	80%
255	Dell Prec. WorkSt. 340/2.66 GHz P4 Xeon	919.0	978.0	Sep-02	44%	75%
256	Intel D850EMV2 (2.8 GHz, Pentium4)	915.0	984.0	Aug-02	43%	75%
257	Fujitsu CELSIUS R610 / 2.8GHz Pentium4 Xeon	914.0	1016.0	Feb-03	43%	77%
258	PrimePower650 (SPARC64 GP/1080MHz)	912.0	737.0	May-03	43%	56%
259	Intel D850EMV2 (2.67 GHz, Pentium4)	898.0	952.0	Aug-02	43%	73%
260	Dell Prec. WorkSt. 340/2.53 GHz P4 Xeon	898.0	941.0	Sep-02	43%	72%
261	PrimePower450 (SPARC64 GP/1100MHz)	884.0	682.0	May-03	42%	52%
262	Dell Prec. WorkSt. 530/2.8 GHz P4 Xeon	878.0	957.0	Sep-02	42%	73%
263	Sun Blade Model 2500 / 1.28 GHz	877.0	696.0	Aug-03	42%	53%
264	Intel D850EMV2 (2.53GHz, Pentium4)	875.0	914.0	Aug-02	42%	70%
265	Sun Fire V1280 / 1200MHz Cu	873.0	676.0	Jan-04	41%	52%
266	AMD ASUS A7N8X XP 3200+ / 2200 MHz	873.0	1080.0	May-03	41%	82%
267	Dell Prec. WorkSt. 340/2.4 GHz P4 Xeon	872.0	905.0	Sep-02	41%	69%
268	PrimePower250 (SPARC64 GP/1100MHz)	866.0	680.0	May-03	41%	52%
269	IBM IntelliStation POWER 275 1.0GHz	862.0	683.0	Jun-03	41%	52%
270	Dell Prec. WorkSt.650/2.0 GHz P4 Xeon	858.0	837.0	Apr-03	41%	64%
271	Intel D850EMV2 (2.4 GHz, Pentium4)	858.0	882.0	Aug-02	41%	67%
272	IBM eServer pSeries 630/6C4-1 1.2GHz	852.0	720.0	Apr-03	40%	55%
273	Dell Prec. WorkSt. 530/2.6 GHz P4 Xeon	852.0	911.0	Sep-02	40%	69%
274	IBM eServer pSeries 630/6E4-1 1200MH	852.0	720.0	Apr-03	40%	55%
275	Sun Fire V880 / 1050 MHz Cu	845.0	626.0	Jun-03	40%	48%
276	IBM eServer pSeries 630/6C4 1.0GHz	843.0	639.0	Jul-02	40%	49%
277	IBM eServer pSeries 630/6E4 1.0GHz	843.0	639.0	Jul-02	40%	49%
278	Intel D850EMV2 (2.26GHz, Pentium4)	833.0	847.0	Aug-02	40%	65%
279	Sun Fire V480 / 1050MHz Cu	826.0	619.0	Jul-03	39%	47%
280	Dell Prec. WorkSt. 530/2.4 GHz P4 Xeon	821.0	859.0	Sep-02	39%	65%
281	Dell Prec. WorkSt. 340/2.2 GHz P4	813.0	833.0	Sep-02	39%	63%
282	HP AlphaServer DS25 68/1000	812.0	678.0	Aug-02	39%	52%
283	Intel D850MD (2.4 GHz, Pentium4)	806.0	833.0	Apr-02	38%	63%
284	Intel D850EMV2 (2.2 GHz, Pentium4)	798.0	811.0	Aug-02	38%	62%
285	Dell Prec. WorkSt. 530/2.2 GHz P4	790.0	810.0	Sep-02	37%	62%
286	AMD ASUS A7N8X XP 2800+ / 2250 MHz	782.0	933.0	Sep-02	37%	71%
287	Compaq AlphaServer ES45 Model 68/1000	776.0	679.0	Oct-01	37%	52%
288	AMD ASUS A7N8X XP 3000+ / 2167 MHz	776.0	995.0	Feb-03	37%	76%
289	Dell PowerEdge 2650/2.8 GHz P4 Xeon	774.0	907.0	Oct-02	37%	69%
290	Dell Prec. WorkSt. 340/2.0 GHz P4	773.0	779.0	Sep-02	37%	59%
291	AMD ASUS A7N8X XP 2700+ / 2166 MHz	772.0	913.0	Sep-02	37%	70%
292	Intel D850EMVR (2.0A GHz, Pentium4)	766.0	828.0	Apr-03	36%	63%
293	Intel D850MD (2.2 GHz, Pentium4)	766.0	784.0	Jan-02	36%	60%
294	Intel D850EMV2 (2.0A GHz, Pentium4)	764.0	820.0	Aug-02	36%	63%
295	Sun Blade 1500 / 1.062GHz	757.0	589.0	Jun-03	36%	45%
296	Dell Prec. WorkSt. 530/2.0 GHz P4	753.0	757.0	Sep-02	36%	58%
297	HP AlphaServer GS160 68/1224	747.0	833.0	Nov-02	35%	63%
298	HP AlphaServer GS320 68/1224	747.0	833.0	Nov-02	35%	63%
299	HP AlphaServer GS80 68/1224	747.0	833.0	Nov-02	35%	63%
300	Intel D850EMV2 (2.0 GHz, Pentium4)	735.0	684.0	Aug-02	35%	52%
301	Sun Fire V1280 / 900MHz Cu	733.0	535.0	Feb-03	35%	41%
302	Intel D850MD (2.0AGHz, Pentium4)	732.0	735.0	Jan-02	35%	56%
303	Dell PowerEdge 2650/2.4 GHz P4 Xeon	726.0	815.0	Apr-02	34%	62%
304	Sun Fire V210 / 1002MHz Cu	722.0	555.0	Mar-03	34%	42%
305	Sun Fire V240 / 1002MHz Cu	718.0	553.0	Mar-03	34%	42%
306	Intel D850EMV2 (1.9 GHz, Pentium4)	716.0	660.0	Aug-02	34%	50%
307	Dell Prec. WorkSt. 530/1.8 GHz P4	715.0	703.0	Sep-02	34%	54%
308	Sun Blade Model 2000 / 1050 MHz	711.0	-	Apr-02	34%	-
309	Dell Prec. WorkSt. 340/1.8 GHz P4	709.0	650.0	Sep-02	34%	50%
310	Intel D850GB (2.0 GHz Pentium4)	704.0	656.0	Aug-01	33%	50%
311	Dell PowerEdge 7150 Itanium/800 4MB L3	703.0	-	Jun-01	33%	-
312	Intel D850MD (2.0 GHz, Pentium4)	702.0	653.0	Jan-02	33%	50%
313	HP SERVER RX4610 Itanium/800 4MB L3	701.0	379.0	May-01	33%	29%
314	Sun Blade Model 2050 / 1050 MHz	701.0	610.0	Nov-01	33%	46%
315	Intel D850EMV2 (1.8 GHz, Pentium4)	699.0	638.0	Aug-02	33%	49%
316	Intel D850MD (1.9 GHz, Pentium4)	688.0	633.0	Jan-02	33%	48%
317	Sun Fire 280R / 1015 MHz Cu	688.0	574.0	Nov-02	33%	44%
318	Intel D850GB (1.9 GHz Pentium4)	687.0	634.0	Aug-01	33%	48%
319	Sun Blade Model 2000 / 1015 MHz	682.0	576.0	Oct-02	32%	44%
320	HP AlphaServer DS15/1000	682.0	631.0	Jun-04	32%	48%
321	Intel D850EMV2 (1.7 GHz, Pentium4)	677.0	613.0	Aug-02	32%	47%
322	Dell PowerEdge 6650/2.0 GHz Xeon MP	677.0	842.0	Oct-02	32%	64%
323	Dell PowerEdge 4600/2.0 GHz P4 Xeon	672.0	701.0	Jan-02	32%	53%
324	Intel D850GB (1.8 GHz Pentium4)	669.0	613.0	Aug-01	32%	47%
325	Intel D850MD (1.8 GHz, Pentium4)	667.0	612.0	Jan-02	32%	47%
326	Dell Prec. WorkSt. 340/1.7 GHz P4	659.0	596.0	Oct-01	31%	45%
327	Dell Prec. WorkSt. 530/1.7 GHz P4	656.0	592.0	Oct-01	31%	45%
328	Intel D850EMV2 (1.6 GHz, Pentium4)	656.0	588.0	Aug-02	31%	45%
329	AMD Epox 8KHA+ XP2600+ / 2133 MHz	655.0	839.0	Aug-02	31%	64%
330	HP i2000 Itanium/800 2ML3	655.0	-	May-01	31%	-
331	Intel D850GB (1.7 GHz Pentium4)	650.0	587.0	Aug-01	31%	45%
332	Dell Prec. WorkSt. 330/2.0 GHz P4	649.0	659.0	Aug-01	31%	50%
333	Intel D850MD (1.7 GHz, Pentium4)	648.0	588.0	Jan-02	31%	45%
334	Dell Prec. WorkSt. 730/Itanium/800 2ML3	645.0	314.0	Jul-01	31%	24%
335	Compaq AlphaServer DS20E Model 68/833	643.0	571.0	Jun-01	31%	44%
336	AMD Epox 8KHA+ XP2400+ / 2000 MHz	641.0	808.0	Aug-02	30%	62%
337	Sun Fire V480 / 900 MHz Cu	637.0	531.0	Jun-02	30%	40%
338	Intel D850EMV2 (1.5 GHz, Pentium4)	634.0	562.0	Aug-02	30%	43%
339	Sun Netra 20 / 900 MHz Cu	634.0	533.0	May-02	30%	41%
340	Sun Blade 1000 Model 900 / 900MHz Cu	629.0	533.0	Nov-01	30%	41%
341	Intel D850GB (1.6 GHz Pentium4)	628.0	565.0	Aug-01	30%	43%
342	Intel D850MD (1.6 GHz, Pentium4)	626.0	565.0	Jan-02	30%	43%
343	AMD Epox 8KHA+ XP2200+ / 1800 MHz	624.0	765.0	Jul-02	30%	58%
344	HP i2000 Itanium/733 2ML3	623.0	-	May-01	30%	-
345	HP SERVER RX4610 Itanium/733 2ML3	623.0	-	May-01	30%	-
346	Compaq AlphaServer ES40 Model 6/833	621.0	565.0	Jun-01	29%	43%
347	Dell Prec. WorkSt. 530/1.5 GHz P4	615.0	545.0	Oct-01	29%	42%
348	Dell Prec. WorkSt. 340/1.5 GHz P4	615.0	547.0	Oct-01	29%	42%
349	Dell Prec. WorkSt. 330/1.8 GHz P4	615.0	619.0	Jul-01	29%	47%
350	Sun Fire V880 / 900 MHz Cu	614.0	507.0	May-02	29%	39%
351	AMD Epox 8KHA+ XP2100+ / 1733 MHz	613.0	749.0	Apr-02	29%	57%
352	Intel D850GB (1.5 GHz Pentium4)	606.0	539.0	Aug-01	29%	41%
353	Intel D850MD (1.5 GHz, Pentium4)	605.0	539.0	Jan-02	29%	41%
354	AMD Asus A7M266-D MP2400+ / 2000 MHz	605.0	766.0	Nov-02	29%	58%
355	AMD MSI K7D M-L MP2600+ / 2133 MHz	602.0	781.0	Feb-03	29%	60%
356	HP 9000 Model C3750 / PA8700+/875	600.0	678.0	Oct-02	28%	52%
357	AMD Epox 8KHA+ XP2000+ / 1667 MHz	596.0	724.0	Jan-02	28%	55%
358	AMD MSI K7D M-L MP2200+ / 1800 MHz	592.0	726.0	Aug-02	28%	55%
359	Sun Fire 280R / 900 MHz Cu	592.0	529.0	Nov-01	28%	40%
360	Dell Prec. WorkSt. 330/1.7 GHz P4	591.0	591.0	May-01	28%	45%
361	AMD Epox 8KHA+ XP1800 / 1600 MHz	588.0	701.0	Oct-01	28%	53%
362	Compaq AlphaServer GS320 Model 32 68/1001	585.0	621.0	Jun-01	28%	47%
363	Compaq AlphaServer GS80 Model 8 68/1001	585.0	621.0	Jun-01	28%	47%
364	Compaq AlphaServer GS160 Model 16 68/1001	585.0	621.0	Jun-01	28%	47%
365	AMD MSI K7D M-L MP2100+ / 1733 MHz	584.0	713.0	Jun-02	28%	54%
366	Compaq AlphaServer DS20L Model 68/833	583.0	537.0	Apr-02	28%	41%
367	Intel D850GB (1.4 GHz Pentium4)	581.0	512.0	Aug-01	28%	39%
368	HP 9000 Model J6750 / PA8700+/875	580.0	676.0	Oct-02	28%	52%
369	AMD Epox 8KHA+ XP1800 / 1533 MHz	572.0	671.0	Oct-01	27%	51%
370	API UP2000 833 MHz	571.0	533.0	Jun-00	27%	41%
371	AMD Epox 8KHA+ XP1700 / 1466 MHz	561.0	656.0	Oct-01	27%	50%
372	Intel D850GB (1.3 GHz Pentium4)	557.0	486.0	Aug-01	26%	37%
373	Fujitsu CELSIUS 460/1.5 GHz P4 (400MHz Bus)	550.0	535.0	Dec-00	26%	41%
374	Dell Prec. WorkSt. 330/1.5 GHz P4	549.0	543.0	May-01	26%	41%
375	AMD Asus A7M266-D MP2000+ / 1667 MHz	548.0	662.0	Mar-02	26%	50%
376	HP ProLiant BL10e G2 1GHz,ULV PentiumM	547.0	687.0	Jul-03	26%	52%
377	AMD GA-7DX Motherboard, XP1800+/1533 M	542.0	644.0	Oct-01	26%	49%
378	Compaq AlphaServer DS20E Model 6/667	535.0	455.0	Aug-01	25%	35%
379	AMD GA-7DX Motherboard, XP1700+/1466 M	535.0	629.0	Oct-01	25%	48%
380	Compaq Evo Workst. W6000/1.4 GHz P4 Xeon	530.0	509.0	Sep-01	25%	39%
381	Dell Prec. WorkSt. 330/1.4 GHz P4	527.0	517.0	May-01	25%	39%
382	HP i2000 Itanium/800 4MB L3	526.0	365.0	Sep-01	25%	28%
383	HP 9000 Model J6700 / PA8700-750	526.0	603.0	Apr-01	25%	46%
384	HP 9000 Model C3700 / PA8700-750	526.0	604.0	Apr-01	25%	46%
385	Dell Prec. WorkSt. 530/1.4 GHz P4	526.0	516.0	May-01	25%	39%
386	AMD GA-7DX Motherboard, XP1600+/1400 M	504.0	595.0	Oct-01	24%	45%
387	AMD Tyan Thunder K7 MP1800+/1533 MHz	504.0	609.0	Oct-01	24%	46%
388	Dell Prec. WorkSt. 330/1.3 GHz P4	502.0	490.0	May-01	24%	37%
389	Compaq AlphaServer ES40 Model 6/667	500.0	433.0	Nov-99	24%	33%
390	SGI Origin 3200 1X 600MHz R14k	499.0	500.0	Feb-02	24%	38%
391	HP i2000 Itanium/733 4MB L3	494.0	-	Sep-01	23%	-
392	AMD GA-7DX Motherboard, XP1500+/1333 M	494.0	577.0	Oct-01	23%	44%
393	AMD Tyan Thunder K7 MP1600+/1400 MHz	483.0	571.0	Oct-01	23%	44%
394	Fujitsu PrimePower650 (SPARC64 GP/810MHz)	483.0	624.0	Jul-02	23%	48%
395	PrimePower650 (SPARC64 GP/810MHz)	483.0	624.0	Jul-02	23%	48%
396	HP C3650/PA8700-625	482.0	508.0	Dec-01	23%	39%
397	API UP2000 750 MHz	478.0	456.0	Jun-00	23%	35%
398	AMD Tyan Thunder K7 MP1500+/1333 MHz	473.0	554.0	Oct-01	22%	42%
399	SGI Origin 300 1X 600MHz R14kA	472.0	483.0	Jul-02	22%	37%
400	HP SERVER RP7400/PA8700-750	464.0	551.0	Sep-01	22%	42%
401	HP SERVER RP5470/PA8700-750	462.0	549.0	Oct-01	22%	42%
402	Fujitsu PrimePower850 (SPARC64 GP/810MHz)	458.0	617.0	Jul-02	22%	47%
403	PrimePower850 (SPARC64 GP/810MHz)	458.0	617.0	Jul-02	22%	47%
404	Compaq AlphaStation XP1000 Model 6/667	452.0	403.0	Sep-01	21%	31%
405	Dell PowerEdge 1500SC/1.4 GHz PIII	437.0	664.0	Jan-02	21%	51%
406	SGI Origin 3200 1X 500MHz R14k	436.0	427.0	Apr-01	21%	33%
407	AMD Tyan Thunder K7 1.2GHz Athlon MP	433.0	522.0	May-01	21%	40%
408	HP SERVER RP5470/PA8700-650	432.0	491.0	Oct-01	20%	37%
409	HP SERVER RP7400/PA8700-650	430.0	493.0	Oct-01	20%	38%
410	IBM eServer pSeries 640/B80 450MHz	426.0	334.0	Oct-01	20%	25%
411	IBM eServer pSeries 610/6E1 450MHz	426.0	334.0	Oct-01	20%	25%
412	AMD Gigabyte GA-7DX 1.4 GHz Athlon	426.0	554.0	May-01	20%	42%
413	IBM RS/6000 44P-170 (450 MHz)	426.0	346.0	Jun-02	20%	26%
414	IBM RS/6000 44P-270 (450MHz, 8MBL2)	426.0	334.0	Oct-01	20%	25%
415	IBM RS/6000 SP-450MHz T/W (1 CPU)	426.0	334.0	Jan-02	20%	25%
416	IBM eServer pSeries 610/6C1 450MHz	426.0	334.0	Oct-01	20%	25%
417	Dell PowerEdge 1500SC/1.26GHz PIII	422.0	623.0	Jan-02	20%	47%
418	Fujitsu PrimePower800/1000/2000 (788MHz)	420.0	591.0	Jul-02	20%	45%
419	RLX 1200i (1.20 GHz Pentium III)	420.0	575.0	Sep-02	20%	44%
420	PrimePower800/1000/2000 (788MHz)	420.0	591.0	Jul-02	20%	45%
421	AMD Gigabyte GA-7DX 1.33GHz Athlon	414.0	539.0	Mar-01	20%	41%
422	Compaq AlphaServer GS320 Model 6/731	405.0	397.0	Jun-00	19%	30%
423	Compaq AlphaServer GS80 Model 6/731	405.0	397.0	Dec-00	19%	30%
424	Compaq AlphaServer GS160 Model 6/731	405.0	397.0	May-00	19%	30%
425	HP 9000 model J6000 / PA8600-552 MHz	400.0	441.0	Oct-00	19%	34%
426	HP 9000 Model B2600 / PA8600-500	397.0	403.0	Apr-01	19%	31%
427	IBM eServer pSeries 660/6H1 750 MHz	396.0	458.0	Apr-02	19%	35%
428	IBM eServer pSeries 620/6F1 750 MHz	396.0	458.0	Apr-02	19%	35%
429	IBM IntelliStation 265 (450MHz, 4MBL2)	396.0	318.0	Jan-02	19%	24%
430	IBM eServer pSeries 620/6F0 750 MHz	396.0	458.0	Apr-02	19%	35%
431	IBM eServer pSeries 660/6H0 750 MHz	396.0	458.0	Apr-02	19%	35%
432	PrimePower400 (SPARC64 GP/700MHz)	394.0	521.0	Jul-02	19%	40%
433	Fujitsu PrimePower400 (SPARC64 GP/700MHz)	394.0	521.0	Jul-02	19%	40%
434	PrimePower200 SPARC64 GP/700MHz-8MB	393.0	521.0	Jul-02	19%	40%
435	HP 9000 model C3600 / PA8600-552 MHz	392.0	432.0	Oct-00	19%	33%
436	HP 9000 Model A500 / PA8600-550 MHz	389.0	422.0	Nov-00	18%	32%
437	AMD Gigabyte GA-7DX 1.2GHz Athlon	387.0	496.0	Feb-01	18%	38%
438	Compaq AlphaServer DS20 Model 6/500	383.0	313.0	Nov-99	18%	24%
439	Compaq AlphaServer ES40 Model 6/500	382.0	311.0	Nov-00	18%	24%
440	SGI Origin 3200 400MHz R12k	382.0	353.0	Sep-00	18%	27%
441	Compaq AlphaServer DS10 6/600	377.0	364.0	Jun-00	18%	28%
442	Dell PowerEdge 2550/1.13GHz PIII	377.0	568.0	Jan-02	18%	43%
443	Sun Netra 20 / 750 MHz	376.0	417.0	Nov-01	18%	32%
444	Fujitsu PrimePower200 (SPARC64 GP/700MHz)	374.0	-	Jul-02	18%	-
445	PrimePower200 SPARC64 GP/700MHz-4MB	374.0	491.0	Jul-02	18%	37%
446	Fujitsu PrimePower650 (SPARC64 GP/675MHz)	371.0	478.0	Oct-01	18%	36%
447	PrimePower650 (SPARC64 GP/675MHz)	371.0	478.0	Oct-01	18%	36%
448	Sun Blade 1000 Model 1900 / 900MHz	369.0	466.0	Jul-01	18%	36%
449	IBM eServer pSeries 610/6E1 375MHz	368.0	277.0	Oct-01	17%	21%
450	IBM eServer pSeries 610/6C1 375MHz	368.0	277.0	Oct-01	17%	21%
451	SGI 2200 1X 500MHz R14k	362.0	412.0	Jul-01	17%	31%
452	IBM eServer pSeries 660/6M1 750 MHz	359.0	439.0	Aug-01	17%	33%
453	PrimePower850 (SPARC64 GP/675MHz)	359.0	476.0	Oct-01	17%	36%
454	Fujitsu PrimePower850 (SPARC64 GP/675MHz)	359.0	476.0	Oct-01	17%	36%
455	IBM IBM RS/6000 Model M80 (750 MHz)	359.0	439.0	Aug-01	17%	33%
456	SGI Origin 300 1X 500MHz R14k	356.0	379.0	Oct-01	17%	29%
457	IBM RS/6000 44P-170 (400 MHz)	355.0	280.0	Jun-02	17%	21%
458	Sun Fire V880 / 750 MHz	354.0	390.0	Nov-01	17%	30%
459	RLX 800i (800 MHz Pentium III)	350.0	435.0	Sep-02	17%	33%
460	AMD ASUS A7V 1.3 GHz Athlon	348.0	491.0	Mar-01	17%	37%
461	HP 9000 Model L3000 / PA8600-550	345.0	388.0	Dec-00	16%	30%
462	HP 9000 Model N4000 / PA8600-552 MHz	338.0	379.0	Mar-00	16%	29%
463	PrimePower800/1000/2000 (675MHz)	332.0	475.0	Sep-01	16%	36%
464	Fujitsu PrimePower800/1000/2000 (675MHz)	332.0	475.0	Sep-01	16%	36%
465	IBM RS/6000 SP-375MHz T/W (1 CPU)	330.0	260.0	Jan-00	16%	20%
466	IBM eServer pSeries 610/6C1 333MHz	329.0	241.0	Feb-02	16%	18%
467	IBM eServer pSeries 610/6E1 333MHz	329.0	241.0	Feb-02	16%	18%
468	Dell Prec. WorkSt. 420/1.0 GHz P3	329.0	462.0	Nov-00	16%	35%
469	AMD ASUS A7V 1.2 GHz Athlon	328.0	458.0	Feb-01	16%	35%
470	Intel OR840 (1.0 GHz Pentium III)	327.0	442.0	Jul-00	16%	34%
471	IBM RS/6000 44P-270 (1 CPU, 8MB L2)	327.0	273.0	Sep-00	16%	21%
472	IBM eServer pSeries 640 (8 MB L2)	327.0	273.0	Sep-00	16%	21%
473	Sun Blade 1000 Model 1750 / 750MHz	326.0	396.0	Jun-01	15%	30%
474	Sun Fire 280R / 750 MHz	324.0	394.0	Jun-01	15%	30%
475	IBM RS/6000 SP-375MHz High Node(1 CPU)	322.0	252.0	Jun-00	15%	19%
476	HP 9000 model B2000 / PA8500-400 MHz	321.0	332.0	Oct-00	15%	25%
477	Intel VC820 (1.13GHz Pentium III)	320.0	464.0	Aug-00	15%	35%
478	SGI 2200 2X 400MHz R12k	319.0	347.0	May-00	15%	26%
479	Dell Prec. WorkSt. 420/933MHz P3	318.0	440.0	Nov-00	15%	34%
480	PrimePower400 (SPARC64 GP/600MHz)	314.0	424.0	Aug-01	15%	32%
481	Fujitsu PrimePower400 (SPARC64 GP/600MHz)	314.0	424.0	Aug-01	15%	32%
482	IBM eServer pSeries 640 (4 MB L2)	313.0	262.0	Sep-00	15%	20%
483	IBM RS/6000 44P-270 (1 CPU, 4MB L2)	313.0	262.0	Sep-00	15%	20%
484	Sun Blade 1000 Model 1750 / 750MHz	312.0	396.0	Oct-01	15%	30%
485	AMD Gigabyte GA-7ZM motherboard 1.1 GH	311.0	-	Aug-00	15%	-
486	IBM RS/6000 44P-270 (1 CPU)	309.0	251.0	Jan-00	15%	19%
487	Dell Prec. WorkSt. 420/866MHz P3	306.0	417.0	Nov-00	15%	32%
488	Intel VC820 (1.0 GHz Pentium III)	304.0	428.0	Jul-00	14%	33%
489	Fujitsu PrimePower600 (SPARC64 GP/600MHz)	304.0	420.0	Aug-01	14%	32%
490	PrimePower600 (SPARC64 GP/600MHz)	304.0	420.0	Aug-01	14%	32%
491	AMD Gigabyte GA-7ZM motherboard 1.2 GH	304.0	-	Oct-00	14%	-
492	Intel VC820 (1.0BGHz, Pentium III)	304.0	448.0	Nov-00	14%	34%
493	Intel D815EEA2 1.0B GHz Pentium III)	303.0	454.0	Nov-04	14%	35%
494	Dell PowerEdge 4400/1.0 GHz PIII Xeon	302.0	451.0	Jan-02	14%	34%
495	AMD ASUS A7V 1.0 GHz Athlon	298.0	-	Feb-01	14%	-
496	Fujitsu PrimePower200 (SPARC64 GP/600MHz)	298.0	406.0	Aug-01	14%	31%
497	PrimePower200 (SPARC64 GP/600MHz)	298.0	406.0	Aug-01	14%	31%
498	Intel D815EEA2 (1.0BGHz Pentium III)	297.0	457.0	Aug-01	14%	35%
499	Dell Prec. WorkSt. 420/800MHz P3	295.0	399.0	Nov-00	14%	30%
500	Intel VC820 (933 MHz Pentium III)	295.0	410.0	Jul-00	14%	31%
501	Intel VC820 (933 MHz, Pentium III)	295.0	429.0	Nov-00	14%	33%
502	Intel D815EEA2 (933 MHz Pentium III)	288.0	428.0	Aug-01	14%	33%
503	Intel VC820 (866 MHz, Pentium III)	285.0	407.0	Nov-00	14%	31%
504	Intel VC820 (866 MHz Pentium III)	284.0	390.0	Jul-00	13%	30%
505	Dell Prec. WorkSt. 420/733MHz P3	281.0	374.0	Nov-00	13%	29%
506	Intel D815EEA2 (866 MHz Pentium III)	280.0	417.0	Aug-01	13%	32%
507	SGI 2100 1X 350MHz R12k	278.0	289.0	Mar-01	13%	22%
508	Sun Fire 6800 / 750 MHz	278.0	360.0	Oct-01	13%	27%
509	Intel VC820 (800EB MHz, Pentium III)	274.0	386.0	Nov-00	13%	29%
510	IBM RS/6000 44P-170 (333 MHz)	274.0	202.0	Jun-02	13%	15%
511	Sun Enterprise 450 / 480MHz	274.0	234.0	Aug-00	13%	18%
512	SGI Origin200 360MHz R12k	274.0	298.0	Jun-00	13%	23%
513	Intel VC820 (800 MHz Pentium III)	273.0	371.0	Jul-00	13%	28%
514	Dell PowerEdge 6400/PIII Xeon 700	271.0	438.0	Apr-01	13%	33%
515	Intel D815EEA2 (800 MHz Pentium III)	269.0	386.0	Aug-01	13%	29%
516	HP 9000 Model N4000 / PA8600-440 MHz	266.0	-	Nov-99	13%	-
517	Sun Blade 150 (UltraSPARC IIi/650)	261.0	246.0	Aug-02	12%	19%
518	Sun Blade 1000 Model 1600 / 600MHz	260.0	311.0	Jul-01	12%	24%
519	Intel VC820 (850 MHz Pentium III)	259.0	367.0	Jul-00	12%	28%
520	Intel D815EEA2 (1.1 GHz Pentium III)	258.0	427.0	Aug-01	12%	33%
521	Intel D815EEA2 (1.0 GHz Pentium III)	254.0	408.0	Aug-01	12%	31%
522	Sun Enterp. 420R (UltraSPARC-II/450)	251.0	214.0	Mar-01	12%	16%
523	Sun Enterprise 3500/4500 - 400 MHz	246.0	212.0	Oct-00	12%	16%
524	IBM eServer pSeries 620/6F1 600 MHz	245.0	310.0	Apr-01	12%	24%
525	IBM eServer pSeries 660/6H1 600 MHz	245.0	310.0	Apr-01	12%	24%
526	IBM eServer pSeries 620/6F0 600 MHz	245.0	310.0	May-01	12%	24%
527	Intel VC820 (800EB MHz Pentium III)	245.0	355.0	Mar-00	12%	27%
528	IBM eServer pSeries 660/6H0 600 MHz	245.0	310.0	May-01	12%	24%
529	IBM RS/6000 Model 7026-M80 (1 CPU)	243.0	275.0	May-00	12%	21%
530	Intel OR840 (733MHz Pentium III)	243.0	336.0	Nov-99	12%	26%
531	Fujitsu CELSIUS 650 / 733 MHz Pentium III	242.0	337.0	Jan-00	11%	26%
532	Sun Blade 150 (UltraSPARC IIi/550)	240.0	217.0	Aug-02	11%	17%
533	Dell PowerEdge 8450/PIII Xeon 700	234.0	422.0	Apr-01	11%	32%
534	Intel VC820 (733 MHz Pentium III)	234.0	335.0	Mar-00	11%	26%
535	Intel SE440BX-2 (800 MHz Pentium III	226.0	344.0	Mar-00	11%	26%
536	Intel VC820 (667 MHz Pentium III)	222.0	314.0	Mar-00	11%	24%
537	Intel SE440BX-2 (750 MHz Pentium III	219.0	330.0	Mar-00	10%	25%
538	Intel SE440BX-2 (700 MHz Pentium III	213.0	315.0	Mar-00	10%	24%
539	IBM RS/6000 Model 7026-F80 (1 CPU)	205.0	234.0	May-00	10%	18%
540	IBM RS/6000 Model 7026-H80 (1 CPU)	205.0	234.0	May-00	10%	18%
541	Dell Prec. WorkSt. 410/700MHz PIII	205.0	307.0	Nov-99	10%	23%
542	Intel SE440BX-2 (650 MHz Pentium III	204.0	299.0	Mar-00	10%	23%
543	Intel SE440BX-2 (500 MHz Pentium III	191.0	233.0	Feb-02	9%	18%
544	IBM RS/6000 7043-260 (1 CPU)	180.0	-	Nov-99	9%	-
545	Intel SE440BX-2 (450 MHz Pentium III	177.0	213.0	Mar-02	8%	16%
546	Compaq AlphaServer 4100 5/533	176.0	176.0	Nov-99	8%	13%
547	Sun Blade 100 (UltraSPARC-IIe/500)	163.0	174.0	Mar-01	8%	13%
548	Compaq DIGITAL PW 500au	158.0	161.0	Nov-99	7%	12%
549	Compaq AlphaStation 500/500	153.0	163.0	Nov-99	7%	12%
550	Ultra 10 333MHz	126.0	133.0	Nov-99	6%	10%
551	IBM RS/6000 43P-150 (250MHz,xlf90)	90.8	105.0	Jun-00	4%	8%
552	IBM RS/6000 43P-150 (250MHz,xlf)	84.4	93.7	May-00	4%	7%

(*) Values derived under earlier releases of OS/compilers/libraries




Table 3. SPEC CPU2000 - SPECfp_rate and SPECint_rate. Absolute and Relative Values.

SPEC CPU2000 - SPECfp_rate and SPECint_rate. Absolute Values and Values Relative to the HP RX5670 Itanium2/1500-H.

		SPEC	SPEC	Test	Relative Values (%)
Machine	NCPU	fp	int	Date	SPECfp	SPECint
SGI Origin 3800 256X 600MHz R14kA	256	1215.0	1402.0	Aug-02	4959%	9224%
SGI Origin 3800 256X 500MHz R14k	256	1090.0	1189.0	Nov-01	4449%	7822%
128 CPUs
SGI Altix 3700 Bx2/Itanium2 1.5G/4MB	128	2437.0	1713.0	Dec-04	9947%	11270%
SGI Altix 3000 (1.5GHz, Itanium 2)	128	2410.0	1721.0	May-04	9837%	11322%
SGI Origin 3800 128X 600MHz R14k	128	638.0	714.0	Apr-02	2604%	4697%
SGI Origin 3800 128X 500MHz R14k	128	570.0	605.0	Dec-01	2327%	3980%
SGI Origin 3800 128X 400MHz R12k	128	525.0	511.0	Aug-00	2143%	3362%
Fujitsu PrimePower2000 (SPARC64 GP/675MHz)	128	517.0	571.0	Oct-01	2110%	3757%
PrimePower2000 (SPARC64 GP/675MHz)	128	517.0	571.0	Oct-01	2110%	3757%
SGI 2800 128X 400MHz R12k	128	407.0	477.0	May-00	1661%	3138%
64 CPUs
IBM eServer p5 595 (1900 MHz) 2 CC	64	1752.0	-	Nov-04	7151%	-
SGI Altix 3700 Bx2/Itanium2 1.6G/9MB	64	1596.0	1052.0	Nov-04	6514%	6921%
SGI Altix 3000 (1.5GHz, Itanium 2)	64	1257.0	854.0	Aug-03	5131%	5618%
HP AlphaServer GS1280 7/1300	64	1224.0	728.0	Aug-04	4996%	4789%
HP AlphaServer GS1280 Model 64	64	1068.0	632.0	Aug-03	4359%	4158%
SGI Altix 3000 (1.3GHz, Itanium 2)	64	1059.0	705.0	Jun-03	4322%	4638%
HP Integrity Superdome Itanium2/1.5G	64	928.0	904.0	Sep-03	3788%	5947%
SGI Altix 3000 (1GHz, Itanium 2)	64	862.0	-	Jan-03	3518%	-
SGI Altix 3000 (900MHz, Itanium 2)	64	693.0	389.0	Mar-03	2829%	2559%
Sun Fire 15K / 1200 MHz Cu	64	645.0	436.0	Mar-03	2633%	2868%
Compaq AlphaServer SC40 Cluster EV67/833	64	439.0	-	May-02	1792%	-
HP 9000 Superdome (1.0GHz PA-8800,16	64	389.0	541.0	Jan-04	1588%	3559%
PrimePower800/1000/2000 (788MHz)	64	329.0	368.0	Aug-02	1343%	2421%
Fujitsu PrimePower800/1000/2000 (788MHz)	64	329.0	368.0	Aug-02	1343%	2421%
SGI Origin 3800 64X 600MHz R14k	64	327.0	362.0	Feb-02	1335%	2382%
HP 9000 Superdome (1.0GHz PA-8800,8C	64	314.0	542.0	Jan-04	1282%	3566%
SGI Origin 3800 64X 500MHz R14k	64	293.0	307.0	May-01	1196%	2020%
HP9000 Superdome 64-way PA8700+/875	64	288.0	413.0	Jul-02	1176%	2717%
Fujitsu PrimePower2000 (SPARC64 GP/675MHz)	64	281.0	319.0	Oct-01	1147%	2099%
PrimePower2000 (SPARC64 GP/675MHz)	64	281.0	319.0	Oct-01	1147%	2099%
HP9000 Superdome 64-way /PA8700-750	64	267.0	377.0	Sep-01	1090%	2480%
SGI Origin 3800 64X 400MHz R12k	64	265.0	259.0	Jul-00	1082%	1704%
SGI 2400 64X 500MHz R14k	64	228.0	289.0	Aug-01	931%	1901%
HP9000 Superdome 64-way / PA8600-552	64	218.0	272.0	Apr-01	890%	1789%
32 CPUs
IBM eServer p5 590 (1650 MHz) 2 CC	32	870.0	-	Oct-04	3551%	-
SGI Altix 3700 Bx2/Itanium2 1.6G/9MB	32	813.0	527.0	Nov-04	3318%	3467%
SGI Altix 3700 Bx2/Itanium2 1.6G/6MB	32	797.0	510.0	Nov-04	3253%	3355%
SGI Altix 3000 (1.5GHz, Itanium 2)	32	647.0	447.0	Jun-03	2641%	2941%
HP AlphaServer GS1280 7/1300	32	619.0	366.0	Jun-04	2527%	2408%
SGI Altix 3000 (1.3GHz, Itanium 2)	32	544.0	367.0	Jun-03	2220%	2414%
HP AlphaServer GS1280 7/1150	32	536.0	313.0	Apr-03	2188%	2059%
Bull NovaScale 6320 Itanium2/1.6GHz 9MBL3	32	510.0	433.0	Nov-04	2082%	2849%
HP Integrity Superdome Itanium2/1.5G	32	470.0	453.0	Sep-03	1918%	2980%
Bull NovaScale 6320 Itanium2/1.5GHz	32	452.0	387.0	May-04	1845%	2546%
PrimePower1500 (SPARC64 GP/1890MHz)	32	449.0	426.0	May-04	1833%	2803%
SGI Altix 3000 (1GHz, Itanium 2)	32	443.0	245.0	Jan-03	1808%	1612%
IBM eServer pSeries 690Turbo 1.7GHz 2CC	32	372.0	339.0	May-03	1518%	2230%
SGI Altix 3000 (900MHz, Itanium 2)	32	352.0	199.0	Feb-03	1437%	1309%
PrimePower1500 (SPARC64 GP/1350MHz)	32	346.0	288.0	Jun-03	1412%	1895%
Sun Fire 12K / 1200 MHz Cu	32	338.0	232.0	Mar-03	1380%	1526%
HP AlphaServer GS320 68/1224	32	320.0	296.0	Nov-02	1306%	1947%
TX7/i9010 (1.0GHz,Itanium 2)	32	303.0	-	Dec-02	1237%	-
IBM eServer pSeries 690Turbo 1.3GHz 2CC	32	260.0	249.0	Apr-02	1061%	1638%
Compaq AlphaServer GS320 Model 32 68/1001	32	242.0	218.0	Jun-01	988%	1434%
Sun Fire E6900 / s400 1.2 GHz 2CC	32	221.0	230.0	Apr-04	902%	1513%
Sun Fire E6900 / s400 1.05GHz 2CC	32	195.0	204.0	Apr-04	796%	1342%
Fujitsu PrimePower800/1000/2000 (788MHz)	32	182.0	205.0	Aug-02	743%	1349%
PrimePower800/1000/2000 (788MHz)	32	182.0	205.0	Aug-02	743%	1349%
HP 9000 RP8420-32/ PA8800-1.0GHz 2CC	32	179.0	278.0	Jan-04	731%	1829%
SGI Origin 3400 32X 600MHz R14k	32	166.0	183.0	Feb-02	678%	1204%
PrimePower2000 (SPARC64 GP/675MHz)	32	154.0	170.0	Sep-01	629%	1118%
Fujitsu PrimePower2000 (SPARC64 GP/675MHz)	32	154.0	170.0	Sep-01	629%	1118%
SGI Origin 300 32X 600MHz R14kA	32	153.0	168.0	Jun-02	624%	1105%
SGI Origin 3400 32X 500MHz R14k	32	149.0	155.0	May-01	608%	1020%
Compaq AlphaServer GS320 Model 6/731	32	148.0	142.0	May-00	604%	934%
HP9000 Superdome 32-way /PA8700-750	32	138.0	193.0	Sep-01	563%	1270%
SGI Origin 3400 32X 400MHz R12k	32	134.0	130.0	Aug-00	547%	855%
SGI 2400 32X 500MHz R14k	32	116.0	147.0	Aug-01	473%	967%
SGI Origin 300 32X 500MHz R14k	32	115.0	136.0	Dec-01	469%	895%
HP9000 Superdome 32-way / PA8600-552	32	112.0	140.0	Apr-01	457%	921%
SGI 2400 32X 400MHz R12k	32	106.0	125.0	May-00	433%	822%
24 CPUs
Sun Fire 6800 / 1200 MHz Cu	24	205.0	180.0	May-03	837%	1184%
Sun Fire E4900 / s400 1.2 GHz 2CC	24	182.0	175.0	Apr-04	743%	1151%
Sun Fire E2900 (12 processor,2CC)	24	175.0	171.0	Mar-04	714%	1125%
Sun Fire E4900 / s400 1.05GHz 2CC	24	167.0	155.0	Apr-04	682%	1020%
Sun Fire 6800 / 1050 MHz Cu	24	140.0	151.0	May-02	571%	993%
Sun Fire 6800 / 750 MHz	24	67.4	101.0	Oct-01	275%	664%
16 CPUs
IBM eServer p5 570 (1900 MHz) 2CC	16	460.0	294.0	Jul-04	1878%	1934%
SGI Altix 3700 Bx2/Itanium2 1.6G/9MB	16	410.0	265.0	Nov-04	1673%	1743%
SGI Altix 3700 Bx2/Itanium2 1.6G/6MB	16	403.0	257.0	Nov-04	1645%	1691%
Bull NovaScale 6320 HPC/Itanium2 1.6G/9MB	16	373.0	249.0	Nov-04	1522%	1638%
SGI Altix 3000 (1.5GHz, Itanium 2)	16	329.0	227.0	Jun-03	1343%	1493%
HP AlphaServer GS1280 7/1300	16	310.0	183.0	Jun-04	1265%	1204%
SGI Altix 350/Itanium2 1.4GHz/3MB	16	300.0	193.0	May-04	1224%	1270%
SGI Altix 3000 (1.3GHz, Itanium 2)	16	279.0	186.0	Jun-03	1139%	1224%
HP AlphaServer GS1280 7/1150	16	274.0	162.0	Jan-03	1118%	1066%
Bull NovaScale 6160 Itanium2/1.6GHz 9MBL3	16	266.0	222.0	Nov-04	1086%	1461%
PrimePower900 (SPARC64 GP/1890MHz)	16	252.0	233.0	May-04	1029%	1533%
PrimePower1500 (SPARC64 GP/1890MHz)	16	252.0	216.0	May-04	1029%	1421%
PrimePower850 (SPARC64 GP/1890MHz)	16	252.0	233.0	May-04	1029%	1533%
SGI Altix 350/Itanium2 1.4GHz/1.5MB	16	243.0	173.0	Jan-04	992%	1138%
HP Integrity Superdome Itanium2/1.5G	16	236.0	229.0	Sep-03	963%	1507%
HP Integrity RX8620/Itanium2,1.5GHz	16	234.0	232.0	Oct-03	955%	1526%
SGI Altix 3000 (1GHz, Itanium 2)	16	227.0	124.0	Jan-03	927%	816%
Unisys ES7000 Aries420/Itanium2,1.5GHz	16	222.0	-	Nov-03	906%	-
Bull NovaScale 5160 Itanium2/1.5GHz	16	215.0	198.0	Jul-03	878%	1303%
HP Integrity RX8620/Itanium2,1.3GHz	16	207.0	183.0	Nov-03	845%	1204%
Bull NovaScale 5160 Itanium2/1.4GHz	16	203.0	158.0	Jul-03	829%	1039%
Unisys ES7000 Aries420/Itanium2,1.3GHz	16	201.0	-	Nov-03	820%	-
PrimePower850 (SPARC64 GP/1350MHz)	16	194.0	154.0	Jun-03	792%	1013%
PrimePower900 (SPARC64 GP/1350MHz)	16	194.0	154.0	May-03	792%	1013%
PrimePower1500 (SPARC64 GP/1350MHz)	16	193.0	149.0	Jun-03	788%	980%
IBM eServer pSeries 690 HPC 1.3GHz 1CC	16	187.0	149.0	Jul-02	763%	980%
IBM eServer pSeries 670 1.5GHz 2CC	16	187.0	156.0	May-03	763%	1026%
Unisys ES7000 Aries 130 Server	16	181.0	-	Mar-03	739%	-
SGI Altix 3000 (900MHz, Itanium 2)	16	178.0	99.7	Feb-03	727%	656%
Bull NovaScale 5160 HPC Itanium2/1.0GHz	16	177.0	117.0	Mar-03	722%	770%
Sun Fire 12K / 1200 MHz Cu	16	174.0	119.0	Mar-03	710%	783%
PrimePower850 (SPARC64 GP/1080MHz)	16	171.0	128.0	Jun-03	698%	842%
PrimePower900 (SPARC64 GP/1350MHz)*	16	169.0	147.0	Mar-03	690%	967%
HP AlphaServer GS160 68/1224	16	161.0	148.0	Nov-02	657%	974%
Unisys ES7000 Aries130/Itanium2,1.0GHz	16	159.0	-	Dec-02	649%	-
TX7/i9010 (1.0GHz,Itanium 2)	16	159.0	-	Dec-02	649%	-
Sun Fire 6800 / 1200 MHz Cu	16	153.0	122.0	May-03	624%	803%
Sun Fire V890 /UltraSPARC-4 1200MHz	16	151.0	118.0	Aug-04	616%	776%
IBM eServer pSeries 690Turbo 1.3GHz 2CC	16	145.0	131.0	Jul-02	592%	862%
Sun Fire E4900 / s400 1.2 GHz 2CC	16	134.0	118.0	Apr-04	547%	776%
Compaq AlphaServer GS160 Model 16 68/1001	16	122.0	111.0	Jun-01	498%	730%
Sun Fire E4900 / s400 1.05GHz 2CC	16	122.0	104.0	Apr-04	498%	684%
HP 9000 RP8420-32/ PA8800-1.0GHz 2CC	16	110.0	142.0	Jan-04	449%	934%
HP 9000 RP7420-16/ PA8800-1.0GHz 2CC	16	107.0	-	Jan-04	437%	-
HP 9000 RP8420-32/ PA8800-1.0GHz 2CC	16	105.0	142.0	Jan-04	429%	934%
Fujitsu PrimePower800/1000/2000 (788MHz)	16	93.9	105.0	Aug-02	383%	691%
PrimePower800/1000/2000 (788MHz)	16	93.9	105.0	Aug-02	383%	691%
PrimePower850 (SPARC64 GP/675MHz)	16	87.0	88.0	Oct-01	355%	579%
Fujitsu PrimePower850 (SPARC64 GP/675MHz)	16	87.0	88.0	Oct-01	355%	579%
SGI Origin 3400 16X 600MHz R14k	16	83.2	91.6	Feb-02	340%	603%
Fujitsu PrimePower800/1000/2000 (675MHz)	16	78.6	85.9	Sep-01	321%	565%
PrimePower800/1000/2000 (675MHz)	16	78.6	85.9	Sep-01	321%	565%
SGI Origin 300 16X 600MHz R14kA	16	76.8	84.1	Jun-02	313%	553%
SGI Origin 3400 16X 500MHz R14k	16	74.8	77.5	Apr-01	305%	510%
Compaq AlphaServer GS160 Model 6/731	16	73.3	69.9	May-00	299%	460%
HP9000 Superdome 16-way /PA8700-750	16	72.9	99.4	Sep-01	298%	654%
HP SERVER RP8400/PA8700-750	16	71.0	98.2	Sep-01	290%	646%
SGI Origin 3400 16X 400MHz R12k	16	66.9	65.3	Aug-00	273%	430%
HP9000 Superdome 16-way / PA8600-552	16	59.0	72.1	Apr-01	241%	474%
SGI 2400 16X 500MHz R14k	16	58.2	74.3	Jul-01	238%	489%
SGI Origin 300 16X 500MHz R14k	16	57.9	68.3	Dec-01	236%	449%
IBM RS/6000 SP-375MHz High Node(16 CPU)	16	51.7	46.0	Jun-00	211%	303%
14 CPUs
Sun Enterprise 4500	14	33.5	34.5	Oct-00	137%	227%
12 CPUs
Sun Fire 6800 / 1200 MHz Cu	12	123.0	-	Jan-04	502%	-
Sun Fire 4800 / 1200 MHz Cu	12	122.0	91.8	May-03	498%	604%
Sun Fire V1280 / 900MHz Cu	12	104.0	72.9	Feb-03	424%	480%
IBM RS/6000 SP-375MHz High Node(12 CPU)	12	41.1	34.6	Jun-00	168%	228%
8 CPUs
IBM eServer p5 570 (1900 MHz) 2CC	8	249.0	147.0	Jul-04	1016%	967%
SGI Altix 3700 Bx2/Itanium2 1.6G/9MB	8	206.0	133.0	Nov-04	841%	875%
SGI Altix 3700 Bx2/Itanium2 1.6G/6MB	8	203.0	129.0	Nov-04	829%	849%
Bull NovaScale 6160 HPC/Itanium2 1.6G/9MB	8	189.0	125.0	Nov-04	771%	822%
SGI Altix 3000 (1.5GHz, Itanium 2)	8	165.0	114.0	Jun-03	673%	750%
HP AlphaServer GS1280 7/1300	8	156.0	91.9	Jun-04	637%	605%
Unisys ES7000 Aries420/Itanium2,1.5GHz	8	154.0	-	Nov-03	629%	-
Bull NovaScale 5160 HPC Itanium2/1.5GHz	8	152.0	93.3	Jul-03	620%	614%
SGI Altix 350/Itanium2 1.4GHz/3MB	8	148.0	98.2	May-04	604%	646%
PrimePower850 (SPARC64 GP/1890MHz)	8	146.0	118.0	May-04	596%	776%
PrimePower900 (SPARC64 GP/1890MHz)	8	146.0	118.0	May-04	596%	776%
SGI Altix 3000 (1.3GHz, Itanium 2)	8	142.0	93.1	Jun-03	580%	613%
HP Integrity RX8620/Itanium2,1.5GHz*	8	142.0	116.0	Oct-03	580%	763%
HP AlphaServer ES80 7/1150	8	136.0	80.5	Jun-04	555%	530%
HP AlphaServer GS1280 7/1150	8	133.0	80.7	Jan-03	543%	531%
PrimePower650 (SPARC64 GP/1890MHz)	8	127.0	119.0	May-04	518%	783%
SGI Altix 350/Itanium2 1.4GHz/1.5MB	8	127.0	88.3	Dec-03	518%	581%
HP Integrity RX8620/Itanium2,1.3GHz*	8	126.0	92.7	Nov-03	514%	610%
Bull NovaScale 5080 Itanium2/1.5GHz	8	125.0	109.0	Jul-03	510%	717%
HP Integrity RX8620/Itanium2,1.5GHz	8	118.0	117.0	Oct-03	482%	770%
HP Integrity RX7620/Itanium2,1.5GHz	8	118.0	116.0	Oct-03	482%	763%
HP AlphaServer ES80 7/1000	8	117.0	68.9	Jan-03	478%	453%
SGI Altix 3000 (1GHz, Itanium 2)	8	116.0	62.1	May-03	473%	409%
Bull NovaScale 5160 HPC Itanium2/1.0GHz	8	113.0	61.4	Mar-03	461%	404%
IBM eServer pSeries655 1.7GHz 2CC	8	111.0	92.0	Jan-04	453%	605%
PrimePower850 (SPARC64 GP/1350MHz)	8	111.0	79.2	Jun-03	453%	521%
PrimePower900 (SPARC64 GP/1350MHz)	8	110.0	79.2	May-03	449%	521%
HP Integrity RX7620/Itanium2,1.3GHz	8	104.0	91.9	Nov-03	424%	605%
HP Integrity RX8620/Itanium2,1.3GHz	8	104.0	91.9	Nov-03	424%	605%
PrimePower900 (SPARC64 GP/1350MHz)*	8	99.9	74.9	Mar-03	408%	493%
PrimePower650 (SPARC64 GP/1350MHz)	8	99.6	78.5	Jun-03	407%	516%
PrimePower850 (SPARC64 GP/1080MHz)	8	96.1	65.8	Jun-03	392%	433%
IBM eServer pSeries655 1.5GHz 2CC	8	92.8	77.5	May-03	379%	510%
Unisys ES7000 Aries130/Itanium2,1.0GHz	8	91.0	-	Dec-02	371%	-
SGI Altix 3000 (900MHz, Itanium 2)	8	89.8	50.3	Feb-03	367%	331%
PrimePower650 (SPARC64 GP/1080MHz)	8	87.8	65.4	May-03	358%	430%
Sun Fire 4800 / 1200 MHz Cu	8	86.3	61.4	May-03	352%	404%
Sun Fire V880 / 1200 MHz Cu	8	82.8	62.8	Sep-03	338%	413%
IBM eServer pSeries 650/6M2 1450MHz 2CC	8	82.4	75.5	Nov-02	336%	497%
Sun Fire V890 /UltraSPARC-4 1200MHz	8	81.3	59.8	Aug-04	332%	393%
HP AlphaServer GS80 68/1224	8	81.0	76.4	Nov-02	331%	503%
Sun Fire V880 / 1050 MHz Cu	8	73.7	55.9	Jun-03	301%	368%
Sun Fire V490 /UltraSPARC-4 1050MHz	8	73.4	52.6	Oct-04	300%	346%
IBM eServer pSeries 655/651 1.1GHz 2CC	8	68.1	56.9	Nov-02	278%	374%
HP 9000 RP7420-16/ PA8800-1.0GHz 2CC	8	62.5	71.7	Jan-04	255%	472%
Compaq AlphaServer GS80 Model 8 68/1001	8	60.0	56.0	Jun-01	245%	368%
HP 9000 RP7420-16/ PA8800-1.0GHz 2CC	8	57.0	71.7	Jan-04	233%	472%
HP 9000 RP4440-8 / PA8800-1.0GHz 2CC	8	55.4	73.2	Jan-04	226%	482%
PrimePower650 (SPARC64 GP/810MHz)	8	54.4	56.5	Aug-02	222%	372%
Fujitsu PrimePower650 (SPARC64 GP/810MHz)	8	54.4	56.5	Aug-02	222%	372%
Sun Fire V880 / 900 MHz Cu	8	53.3	47.6	May-02	218%	313%
PrimePower650/850 (SPARC64 GP/675MHz	8	44.0	43.9	Oct-01	180%	289%
Fujitsu PrimePower650/850 (SPARC64 GP/675MHz	8	44.0	43.9	Oct-01	180%	289%
SGI Origin 3200 8X 600MHz R14k	8	41.7	45.9	Feb-02	170%	302%
HP ProLiant DL740 (3.0GHz Xeon MP 8C)	8	41.1	86.2	Mar-04	168%	567%
SGI Origin 300 8X 600MHz R14kA	8	38.7	43.3	May-02	158%	285%
Compaq AlphaServer GS80 Model 6/731	8	38.5	36.0	Dec-00	157%	237%
SGI Origin 3200 8X 500MHz R14k	8	37.7	38.5	Apr-01	154%	253%
HP SERVER RP7410/PA8700-750	8	36.8	49.9	Feb-02	150%	328%
HP SERVER RP8400/PA8700-750	8	35.9	49.9	Sep-01	147%	328%
SGI Origin 3200 8X 400MHz R12k	8	33.8	32.6	Sep-00	138%	214%
HP SERVER RP7400/PA8700-750	8	32.1	46.7	Sep-01	131%	307%
PrimePower600 (SPARC64 GP/600MHz)	8	31.5	37.2	Sep-01	129%	245%
Fujitsu PrimePower600 (SPARC64 GP/600MHz)	8	31.5	37.2	Sep-01	129%	245%
Sun Fire V880 / 750 MHz	8	31.2	36.2	Nov-01	127%	238%
HP SERVER RP7400/PA8700-650	8	30.5	42.6	Oct-01	124%	280%
IBM IBM RS/6000 Model M80 (750 MHz)	8	30.0	38.5	Aug-01	122%	253%
IBM eServer pSeries 660/6M1 750MHz	8	30.0	38.5	Aug-01	122%	253%
SGI 2200 8X 500MHz R14k	8	29.3	37.3	Jul-01	120%	245%
SGI Origin 300 8X 500MHz R14k	8	28.9	34.0	Nov-01	118%	224%
IBM RS/6000 SP-375MHz High Node(8 CPU)	8	28.0	23.1	Jun-00	114%	152%
SGI 2200 8X 400MHz R12k	8	26.2	30.5	May-00	107%	201%
HP 9000 Model N4000 / PA8600-552 MHz	8	23.0	32.7	Mar-00	94%	215%
SGI 2100 8X 350MHz R12k	8	22.8	25.1	May-00	93%	165%
IBM RS/6000 Model 7026-M80 (8 CPU)	8	21.1	25.1	May-00	86%	165%
Sun Enterprise 3500/4500 - 400 MHz	8	20.9	19.5	Oct-00	85%	128%
Dell PowerEdge 8450/PIII Xeon 700	8	9.2	25.7	Apr-01	38%	169%
6 CPUs
IBM eServer pSeries 620/6F1 668MHz	6	17.3	26.5	Apr-01	71%	174%
IBM eServer pSeries 660/6H1 668MHz	6	17.3	26.5	Apr-01	71%	174%
IBM RS/6000 Model 7026-F80 (6 CPU)	6	13.2	18.0	May-00	54%	118%
IBM RS/6000 Model 7026-H80 (6 CPU)	6	13.2	18.0	May-00	54%	118%
4 CPUs
IBM eServer p5 570 (1900 MHz) 2CC	4	130.0	76.3	Jul-04	531%	502%
SGI Altix 3700 Bx2/Itanium2 1.6G/9MB	4	104.0	66.5	Nov-04	424%	438%
SGI Altix 3700 Bx2/Itanium2 1.6G/6MB	4	102.0	64.6	Nov-04	416%	425%
IBM eServer p5 550 (1650 MHz) 2CC	4	84.8	65.5	Jul-04	346%	431%
IBM eServer i5 520 (1650 MHz) 2CC	4	84.8	-	Aug-04	346%	-
IBM eServer p5 550 Express (1500MHz) 2CC	4	83.2	-	Oct-04	340%	-
SGI Altix 3000 (1.5GHz, Itanium 2)	4	82.7	-	Jun-03	338%	-
IBM eServer OpenPower 720 (1650 MHz)	4	80.8	59.8	Oct-04	330%	393%
AMD Tyan K8QS Opteron850/2.4GHz SES8/P	4	79.2	70.2	Aug-04	323%	462%
Bull NovaScale 4040 Itanium2/1.6Hz 9MBL3	4	78.8	66.4	Nov-04	322%	437%
AMD Tyan K8QS Pro Opteron850/2.4GHz	4	78.7	-	Nov-04	321%	-
Dell PowerEdge 7250 Itanium2/1600 9MB L3	4	78.2	64.0	Nov-04	319%	421%
HP AlphaServer GS1280 7/1300	4	78.0	46.1	Jun-04	318%	303%
HP Integrity RX4640/Itanium2 1.6G/9MB	4	77.9	72.5	Nov-04	318%	477%
SGI Altix 350/Itanium2 1.4GHz/3MB	4	75.6	49.2	May-04	309%	324%
HP ProLiant DL585 Opteron850/2.4GHz	4	72.8	67.6	Sep-04	297%	445%
HP Integrity RX7620/Itanium2,1.5GHz*	4	71.4	58.6	Oct-03	291%	386%
SGI Altix 3000 (1.3GHz, Itanium 2)	4	71.2	46.7	Jan-04	291%	307%
HP Integrity RX4640/Itanium2 1.5G/4MB	4	70.5	62.2	Nov-04	288%	409%
IBM eServer pSeries655 1.7GHz/4-LPAR 1CC	4	70.2	52.4	Oct-03	287%	345%
IBM eServer pSeries655 1.7GHz 1CC	4	70.1	52.5	Oct-03	286%	345%
Sun Fire V40z (Opteron850/2.4GHz)	4	70.0	63.5	Jul-04	286%	418%
IBM eServer pSeries 655 1.7GHz 1CC	4	68.5	-	May-03	280%	-
HP AlphaServer ES80 7/1150	4	68.4	40.3	Jun-04	279%	265%
HP AlphaServer ES47 7/1150	4	68.4	40.3	Jun-04	279%	265%
HP AlphaServer GS1280 7/1150	4	68.2	40.6	Jan-03	278%	267%
HP Integrity RX5670/Itanium2,1.5GHz	4	66.4	60.0	Jun-03	271%	395%
HP Integrity RX4640/Itanium2,1.5GHz	4	65.6	64.2	Jan-04	268%	422%
Bull NovaScale 4040 Itanium2/1.5GHz	4	64.6	56.1	Jul-03	264%	369%
SGI Altix 350/Itanium2 1.4GHz/1.5MB	4	64.5	44.4	Dec-03	263%	292%
ION Computers I2X4 (1.5GHz Itanium2)	4	64.3	56.1	May-04	262%	369%
HP Integrity RX7620/Itanium2,1.3GHz	4	63.2	46.0	Nov-03	258%	303%
AMD Tyan K8QS Opteron850/2.4GHz W/PGI	4	61.6	-	Sep-04	251%	-
HP AlphaServer ES80 7/1000	4	58.9	34.6	Jan-03	240%	228%
HP AlphaServer ES47 7/1000	4	58.9	34.6	Jan-03	240%	228%
HP Integrity RX7620/Itanium2,1.5GHz	4	58.9	58.3	Oct-03	240%	384%
SGI Altix 3000 (1GHz, Itanium 2)	4	58.4	31.0	May-03	238%	204%
Sun Fire V440/UltraSPARC-IIi 1.6GHz	4	57.8	38.7	Nov-04	236%	255%
HP Integrity RX4640/Itanium2,1.3GHz	4	57.4	51.4	Jan-04	234%	338%
HP Integrity RX5670/Itanium2,1.3GHz	4	57.2	48.6	Nov-03	233%	320%
SGI Altix 350 (1.0GHz, Itanium 2)	4	55.0	33.8	Feb-04	224%	222%
PrimePower450 (SPARC64 GP/1320MHz)	4	53.9	39.8	Dec-03	220%	262%
AMD Appro 4144H/Opteron846 2.0GHz-I,WS	4	52.5	56.7	Sep-03	214%	373%
HP Integrity RX7620/Itanium2,1.3GHz*	4	52.2	46.4	Nov-03	213%	305%
IBM eServer pSeries 655/651 1.3GHz 1CC	4	51.7	38.3	Nov-02	211%	252%
AMD QuatreX-64 Opteron850/2.4GHz WS 1C	4	50.5	68.5	May-04	206%	451%
HP AlphaServer ES45 68/1250	4	50.0	42.0	Nov-02	204%	276%
AMD QuatreX-64 Opteron850/2.4GHz SES 1	4	49.8	57.0	May-04	203%	375%
HP RX5670/Itanium2, 1GHz,3MB-L3, L	4	49.3	-	Nov-02	201%	-
AMD Einox A4800/Opteron844 1.8GHz-I,WS	4	49.2	48.5	Apr-03	201%	319%
Bull NovaScale 4040 Itanium2/1.0GHz	4	48.8	32.9	Feb-03	199%	216%
Unisys ES7000 Aries130/Itanium2,1.0GHz	4	48.0	-	Dec-02	196%	-
Sun Fire V480 / 1200MHz Cu	4	45.5	31.7	Oct-03	186%	209%
AMD Einox A4800/Opteron842 1.6GHz-I,WS	4	45.0	45.1	Apr-03	184%	297%
SGI Altix 3000 (900MHz, Itanium 2)	4	45.0	25.3	Feb-03	184%	166%
HP RX5670/Itanium2, 1GHz,3MB-L3, H	4	43.7	36.8	Jul-02	178%	242%
Sun Fire V880 / 1050 MHz Cu	4	41.4	28.2	Jun-03	169%	186%
PrimePower450 (SPARC64 GP/1100MHz)	4	41.4	30.5	May-03	169%	201%
AMD Einox A4800/Opteron840 1.4GHz-I,WS	4	40.7	40.0	Apr-03	166%	263%
Sun Fire V480 / 1050MHz Cu	4	40.1	28.2	Jul-03	164%	186%
IBM eServer pSeries 630/6C4 1450MHz 2CC	4	38.8	37.0	Feb-03	158%	243%
IBM eServer pSeries 630/6E4 1450MHz 2CC	4	38.8	37.0	Feb-03	158%	243%
HP RX5670/Itanium2,900MHz,1.5MB-L3,L	4	38.7	-	Nov-02	158%	-
Sun Fire V490 /UltraSPARC-4 1050MHz	4	37.9	26.4	Oct-04	155%	174%
Compaq AlphaServer ES45 Model 68/1000	4	36.8	30.8	Oct-01	150%	203%
IBM eServer pSeries 630/6E4 1200MHz 2CC	4	35.1	31.4	Apr-03	143%	207%
IBM eServer pSeries 630/6C4 1200MHz 2CC	4	35.1	31.4	Apr-03	143%	207%
HP 9000 RP4440-8 / PA8800-1.0GHz 2CC	4	34.7	37.0	Jan-04	142%	243%
HP 9000 RP3440-4 / PA8800-1.0GHz 2CC	4	32.6	37.1	Feb-04	133%	244%
Dell PowerEdge 6600/3.0 GHz Xeon MP	4	32.4	57.2	Feb-04	132%	376%
ION Computers BX41 (3GHz Xeon MP/4ML3 4C)	4	31.0	60.0	May-04	127%	395%
HP ProLiant DL580 G2 3.0GHz XeonMP 1CC	4	30.7	61.6	Mar-04	125%	405%
Sun Fire V880 / 900 MHz Cu	4	30.6	23.9	May-02	125%	157%
Sun Fire V480 / 900 MHz Cu	4	29.9	24.1	Jun-02	122%	159%
Compaq AlphaServer ES40 Model 6/833	4	29.2	25.8	Jun-01	119%	170%
Compaq AlphaServer ES40 Model 6/833*	4	26.1	24.7	Feb-01	107%	163%
Dell PowerEdge 6650/2.8 GHz Xeon MP	4	25.4	47.4	Jun-03	104%	312%
HP ProLiant DL740 (3.0GHz Xeon MP 4C)	4	25.4	48.5	Mar-04	104%	319%
HP ProLiant DL580 G2 2.8GHz Xeon MP	4	24.6	47.6	Sep-03	100%	313%
Dell PowerEdge 7150 Itanium/800 4MB L3	4	23.0	-	Jun-01	94%	-
HP ProLiant BL40p (2.8GHz Xeon MP)	4	22.8	46.5	Sep-03	93%	306%
HP ProLiant DL560 (2.8GHz Xeon MP)	4	22.7	46.5	Sep-03	93%	306%
HP SERVER RX4610 Itanium/800 4MB L3	4	22.4	-	May-01	91%	-
Compaq AlphaServer ES40 Model 6/667	4	21.9	19.8	Nov-00	89%	130%
SGI Origin 3200 4X 600MHz R14k	4	21.0	23.0	Feb-02	86%	151%
HP SERVER RP7400/PA8700-750	4	20.5	24.6	Sep-01	84%	162%
HP SERVER RX4610 Itanium/800 4MB L3*	4	20.3	-	Aug-01	83%	-
Dell PowerEdge 6650/2.0 GHz Xeon MP	4	20.2	34.7	Oct-02	82%	228%
HP SERVER RX4610 Itanium/733 4MB L3	4	20.1	-	May-01	82%	-
HP SERVER RP5470/PA8700-750	4	19.9	24.5	Oct-01	81%	161%
SGI Origin 300 4X 600MHz R14kA	4	19.5	21.6	May-02	80%	142%
HP SERVER RP7400/PA8700-650	4	19.3	22.1	Oct-01	79%	145%
SGI Origin 3200 4X 500MHz R14k	4	19.0	19.6	May-01	78%	129%
HP SERVER RP7410/PA8700-750	4	18.9	25.3	Feb-02	77%	166%
HP SERVER RP8400/PA8700-750	4	18.8	25.7	Sep-01	77%	169%
HP SERVER RP5470/PA8700-650	4	18.7	22.0	Oct-01	76%	145%
Fujitsu PrimePower400 (SPARC64 GP/600MHz)	4	18.0	19.3	Sep-01	73%	127%
PrimePower400 (SPARC64 GP/600MHz)	4	18.0	19.3	Sep-01	73%	127%
PrimePower600 (SPARC64 GP/600MHz)	4	17.4	19.0	Sep-01	71%	125%
Fujitsu PrimePower600 (SPARC64 GP/600MHz)	4	17.4	19.0	Sep-01	71%	125%
SGI Origin 3200 4X 400MHz R12k	4	16.9	16.3	Sep-00	69%	107%
IBM eServer pSeries 660/6M1 750MHz	4	16.5	20.3	Aug-01	67%	134%
IBM IBM RS/6000 Model M80 (750 MHz)	4	16.5	20.3	Aug-01	67%	134%
IBM eServer pSeries 660/6H0 750MHz	4	15.6	20.5	Apr-02	64%	135%
IBM eServer pSeries 660/6H1 750MHz	4	15.6	20.5	Apr-02	64%	135%
Compaq AlphaServer ES40 Model 6/500	4	15.6	13.7	Nov-00	64%	90%
IBM eServer pSeries 620/6F0 750MHz	4	15.6	20.5	Apr-02	64%	135%
IBM eServer pSeries 620/6F1 750MHz	4	15.6	20.5	Apr-02	64%	135%
SGI 2200 4X 500MHz R14k	4	14.8	18.6	Jul-01	60%	122%
SGI Origin 300 4X 500MHz R14k	4	14.5	17.1	Oct-01	59%	113%
IBM RS/6000 SP-375MHz High Node(4 CPU)	4	14.5	11.6	Jun-00	59%	76%
HP 9000 Model L3000 / PA8600-550	4	14.5	17.4	Dec-00	59%	114%
HP 9000 Model N4000 / PA8600-552 MHz	4	14.4	17.0	Mar-00	59%	112%
IBM RS/6000 44P-270 (450MHz, 8MBL2)	4	14.1	15.2	Oct-01	58%	100%
IBM RS/6000 SP-450MHz T/W (4 CPU)	4	14.1	15.2	Jan-02	58%	100%
IBM eServer pSeries 640/B80 450MHz	4	14.1	15.2	Oct-01	58%	100%
SGI 2200 4X 400MHz R12k	4	13.2	15.4	May-00	54%	101%
IBM RS/6000 44P-270 (4 CPU, 8MBL2)	4	12.8	12.4	Sep-00	52%	82%
IBM eServer pSeries 640 (8 MB L2)	4	12.8	12.4	Sep-00	52%	82%
IBM RS/6000 44P-270 (4 CPU, 4MBL2)	4	11.9	11.7	Sep-00	49%	77%
IBM eServer pSeries 640 (4 MB L2)	4	11.9	11.7	Sep-00	49%	77%
IBM RS/6000 44P-270 (4 CPU)	4	11.8	11.2	Jan-00	48%	74%
SGI 2100 4X 350MHz R12k	4	11.5	12.8	Jun-00	47%	84%
IBM eServer pSeries 660/6H0 600MHz	4	11.1	15.0	May-01	45%	99%
IBM eServer pSeries 620/6F0 600MHz	4	11.1	15.0	May-01	45%	99%
IBM eServer pSeries 660/6H1 600MHz	4	11.1	15.0	Apr-01	45%	99%
Sun Enterprise 3500/4500 - 400 MHz	4	11.1	9.7	Oct-00	45%	64%
IBM eServer pSeries 620/6F1 600MHz	4	11.1	15.0	Apr-01	45%	99%
Sun Enterprise 450 / 480MHz	4	11.1	10.9	Aug-00	45%	72%
IBM RS/6000 Model 7026-H80 (4 CPU)	4	9.4	11.3	May-00	39%	74%
IBM RS/6000 Model 7026-F80 (4 CPU)	4	9.4	11.3	May-00	39%	74%
Dell PowerEdge 6400/PIII Xeon 700	4	8.9	17.9	Apr-01	36%	118%
2 CPUs
HP Integrity RX2620/Itanium2 1.6G/6MB	2	51.5	35.5	Dec-04	210%	234%
HP Integrity RX1620/Itanium2 1.6G/3MB	2	50.4	33.2	Nov-04	206%	218%
HP Integrity RX2620/Itanium2 1.6G/3MB	2	48.5	32.3	Nov-04	198%	212%
HP Integrity RX4640/Itanium2 1.5G/4MB	2	48.3	31.7	Nov-04	197%	209%
IBM eServer p5 520 Express (1500MHz) 2CC	2	44.0	-	Oct-04	180%	-
HP Integrity RX4640/Itanium2,1.5GHz	2	43.2	32.5	Jan-04	176%	214%
IBM eServer p5 520 (1650 MHz) 2CC	2	43.0	32.9	Jul-04	176%	216%
IBM eServer i5 520 (1650 MHz) 2CC	2	43.0	32.9	Aug-04	176%	216%
HP Integrity RX2620/Itanium2 1.3G/3MB	2	42.7	26.9	Dec-04	174%	177%
HP Integrity RX1620/Itanium2 1.3G/3MB	2	42.7	27.0	Dec-04	174%	178%
HP Integrity RX5670/Itanium2,1.5GHz	2	42.6	30.3	Jun-03	174%	199%
HP Integrity RX2600/Itanium2,1.5GHz	2	42.4	30.5	Jun-03	173%	201%
HP ZX6000/Itanium2, 1.5GHz,6MB-L3, H	2	42.4	-	Jun-03	173%	-
AMD Tyan K8QS Pro Opteron850/2.4GHz	2	42.2	-	Nov-04	172%	-
Fujitsu CELSIUS V810 / 2.4GHz Opteron 250 L	2	42.2	37.1	Nov-04	172%	244%
IBM eServer 326 2.4 GHz Opteron EKO	2	39.9	35.3	Oct-04	163%	232%
Sun Java Wrkst. W2100z Opteron2.4GHz	2	39.6	36.0	Jul-04	162%	237%
HP AlphaServer GS1280 7/1300	2	39.1	23.1	Jun-04	160%	152%
HP ProLiant DL145 Opteron250/2.4GHz	2	39.0	35.7	Oct-04	159%	235%
ION Computers I2X2 (1.6GHz Itanium2)	2	38.9	27.0	May-04	159%	178%
HP ProLiant DL585 Opteron850/2.4GHz	2	38.8	35.8	Sep-04	158%	236%
Fujitsu CELSIUS V810 / 2.4GHz Opteron 250 L*	2	38.8	-	Aug-04	158%	-
AMD Tyan K8QS Opteron850/2.4GHz SES8/P	2	38.7	37.2	Aug-04	158%	245%
SGI Altix 350/Itanium2 1.4GHz/3MB	2	38.3	24.7	May-04	156%	163%
HP Integrity RX4640/Itanium2,1.3GHz	2	37.9	25.8	Jan-04	155%	170%
Dell PowerEdge3250/Itanium2 1.5GHz/6ML3	2	37.3	25.4	Jul-03	152%	167%
HP Integrity RX5670/Itanium2,1.3GHz	2	37.3	24.5	Nov-03	152%	161%
Dell PowerEdge 3250/Itanium2 1.4GHz/3ML3	2	37.2	23.6	Mar-04	152%	155%
Sun Fire V20z (Opteron250/2.4GHz)	2	37.2	33.8	Jul-04	152%	222%
Sun Fire V40z (Opteron850/2.4GHz)	2	37.2	33.7	Jul-04	152%	222%
HP Integrity RX2600/Itanium2,1.3GHz	2	36.9	24.8	Nov-03	151%	163%
ION Computers I2X2 (1.4GHz Itanium2)	2	36.8	21.3	Jul-03	150%	140%
AMD Tyan K8W Opteron250/2.4GHz WS 1CC	2	36.3	37.7	May-04	148%	248%
Supermicro SS6113M-i(1.5GHz/6MB, Itanium 2)	2	35.8	25.3	Dec-03	146%	166%
HP AlphaServer ES47 7/1150	2	34.2	20.2	Jun-04	140%	133%
HP AlphaServer ES80 7/1150	2	34.2	20.2	Jun-04	140%	133%
IBM eServer 325 2.4GHz Opteron 64-PGI	2	34.1	32.9	May-04	139%	216%
HP AlphaServer GS1280 7/1150	2	33.9	20.3	Jan-03	138%	134%
AMD Tyan K8QS Opteron850/2.4GHz W/PGI	2	33.8	-	Sep-04	138%	-
Supermicro SS6113M-i(1.4GHz/4MB, Itanium 2)	2	33.6	22.3	Dec-03	137%	147%
AMD Tyan K8W Opteron250/2.4GHz SES 1CC	2	32.9	34.0	May-04	134%	224%
AMD RioworksHDAMA Opteron 248/2.2GHz S	2	32.8	30.9	Nov-03	134%	203%
AMD RioworksHDAMA Opteron 248/2.2GHz W	2	31.6	33.1	Nov-03	129%	218%
Sun Fire V440/UltraSPARC-IIi 1.6GHz	2	31.2	19.5	Nov-04	127%	128%
Fujitsu CELSIUS V810 (Opteron 248 2.2 GHz)	2	31.1	29.9	Feb-04	127%	197%
HP RX5670/Itanium2, 1GHz,3MB-L3, L	2	30.7	-	Nov-02	125%	-
HP ZX6000/Itanium2, 1GHz,3MB-L3, L	2	30.0	-	Nov-02	122%	-
AMD Rioworks HDAMA Opteron 246 / 2.0 G	2	29.9	30.5	Jul-03	122%	201%
HP RX2600/Itanium2, 1GHz,3MB-L3, L	2	29.9	-	Nov-02	122%	-
HP AlphaServer ES47 7/1000	2	29.6	17.4	Jan-03	121%	114%
HP AlphaServer ES80 7/1000	2	29.6	17.4	Jan-03	121%	114%
HP AlphaServer ES45 68/1250	2	29.0	21.3	Nov-02	118%	140%
IBM xSeries 336 (3.6GHz Xeon)	2	28.6	-	Sep-04	117%	-
Dell PowerEdge3250/Itanium2 1.4GHz/1.5ML3	2	27.8	18.8	Oct-03	113%	124%
HP Integrity RX1600/Itanium2,1.0GHz	2	27.6	19.1	Feb-04	113%	126%
IBM xSeries 336 (3.6GHz,800MHz FSB)	2	27.4	-	Aug-04	112%	-
PrimePower250 (SPARC64 GP/1320MHz	2	27.3	20.0	Dec-03	111%	132%
HP RX5670/Itanium2, 1GHz,3MB-L3, H	2	27.0	18.6	Jul-02	110%	122%
IBM eServer 325 2.0GHz Opteron 64-PGI	2	27.0	26.6	Jul-03	110%	175%
AMD Einox A4800/Opteron244 1.8GHz-I,WS	2	26.7	26.8	Apr-03	109%	176%
HP ProLiant BL20p G3 3.6GHz Xeon	2	26.5	33.7	Nov-04	108%	222%
HP RX2600/Itanium2, 1GHz,3MB-L3, H	2	26.5	18.7	Jul-02	108%	123%
HP ProLiant ML350 G4 3.4GHz Xeon	2	25.9	31.6	Nov-04	106%	208%
HP ProLiant ML370 G4 3.6GHz Xeon	2	25.8	33.8	Oct-04	105%	222%
HP ProLiant ML370 G4 3.4GHz Xeon	2	25.4	32.4	Aug-04	104%	213%
Dell PowerEdge 1855/3.6GHz Xeon	2	25.2	33.4	Nov-04	103%	220%
Dell PowerEdgeSC 1425 (3.6 GHz Xeon)	2	25.2	33.5	Nov-04	103%	220%
AMD Einox A4800/Opteron242 1.6GHz-I,WS	2	25.1	24.0	Apr-03	102%	158%
Dell PowerEdge 2850/3.6 GHz P4Xeon	2	24.9	33.2	Aug-04	102%	218%
Dell PowerEdge 1850/3.6GHz P4Xeon	2	24.9	33.2	Aug-04	102%	218%
AMD Einox A4800/Opteron144 1.8GHz-I,SE	2	24.9	-	Apr-03	102%	-
HP ProLiant DL380 G4 3.6GHz Xeon,L3	2	24.8	32.9	Jul-04	101%	216%
HP ProLiant BL30p (3.2GHz Xeon,2MB L3)	2	24.7	34.3	Nov-04	101%	226%
HP RX5670/Itanium2,900MHz,1.5MB-L3,L	2	24.5	-	Nov-02	100%	-
HP ProLiant DL380 G4 3.4GHz Xeon,L3	2	24.3	31.5	Jul-04	99%	207%
Dell PowerEdge 2650/3.2 GHz P4Xeon,2ML3	2	24.3	31.5	Feb-04	99%	207%
RSN-1164/op (1.8 GHz Opteron)	2	24.2	25.1	Apr-03	99%	165%
HP ProLiant DL360 G3 3.2GHz/Xeon,2ML3 1	2	24.0	34.0	Apr-04	98%	224%
HP ProLiant ML370 G3 (3.2GHz/2MB L3	2	24.0	33.8	Jun-04	98%	222%
HP RX2600/Itanium2,900MHz,1.5MB-L3,L	2	23.9	-	Nov-02	98%	-
Sun Blade Model 2000 / 1200 MHz	2	23.9	16.4	Apr-03	98%	108%
HP ZX6000/Itanium2,900MHz,1.5MB-L3,L	2	23.9	-	Nov-02	98%	-
AMD Rioworks HDAMA Opteron 240EE/1.4GH	2	23.6	22.3	Feb-04	96%	147%
Sun Netra 20 / 1200 MHz Cu	2	23.6	16.3	Mar-03	96%	107%
Sun Fire 280R / 1200 MHz Cu	2	23.6	16.3	Mar-03	96%	107%
HP ProLiant DL380 G3 3.2GHz Xeon 1CC	2	23.5	33.6	Apr-04	96%	221%
Supermicro X6DH8-G2 (3.40 GHz Xeon, 1MBL2)	2	23.4	-	Jul-04	96%	-
Supermicro X6DHT-G (3.6GHz, Intel Xeon )	2	23.3	31.2	Nov-04	95%	205%
Dell PowerEdge 6600/3.0 GHz Xeon MP	2	23.1	31.4	Feb-04	94%	207%
IBM x335 (3.2GHz, 533MHz FSB)	2	22.9	33.8	Feb-04	93%	222%
HP ProLiant DL360 G4 3.6GHz Xeon,L3	2	22.8	31.7	Oct-04	93%	209%
AMD Einox A4800/Opteron240 1.4GHz-I,WS	2	22.7	21.2	Apr-03	93%	139%
Supermicro X6DVL-EG (3.60 GHz Xeon 1MBL2)	2	22.6	30.8	Sep-04	92%	203%
Supermicro X6DHP-8G2 (3.6GHz, Intel Xeon )	2	22.5	26.3	Nov-04	92%	173%
Supermicro X6DH8-G (3.6GHz, Intel Xeon)	2	22.4	31.1	Nov-04	91%	205%
Supermicro X6DHP-8G (3.60 GHz Xeon, 1MBL2)	2	22.2	29.3	Sep-04	91%	193%
HP ProLiant DL360 G4 3.4GHz Xeon,L3	2	22.1	30.4	Jul-04	90%	200%
Prec. WorkSt.650/3.2GHz P4Xeon 2ML3	2	22.1	33.6	Jan-04	90%	221%
HP ProLiant DL580 G2 3.0GHz XeonMP 1CC	2	22.1	33.4	Mar-04	90%	220%
ION Computers BX21 (3.2GHz Xeon/2ML3 2C)	2	21.9	32.3	May-04	89%	212%
Supermicro X6DA8-G2 (3.40 GHz Xeon 1MBL3)	2	21.8	26.3	Jun-04	89%	173%
AMD Einox A4800/Opteron244 1.8GHz-I,SE	2	21.6	24.4	Apr-03	88%	161%
Sun Fire V880 / 1050 MHz Cu	2	21.6	14.1	Jun-03	88%	93%
HP AlphaServer DS25 68/1000	2	21.5	15.5	Aug-02	88%	102%
PrimePower250 (SPARC64 GP/1100MHz	2	21.3	15.3	May-03	87%	101%
Supermicro X6DH8-XG2 (3.40 GHz Xeon, 1MBL2)	2	21.3	26.5	Jul-04	87%	174%
Compaq AlphaServer ES45 Model 68/1000	2	21.1	15.6	Oct-01	86%	103%
Dell PowerEdge 2650/3.2 GHz P4Xeon,1ML3	2	21.1	28.3	Oct-03	86%	186%
Sun Fire V480 / 1050MHz Cu	2	20.8	14.2	Jul-03	85%	93%
Dell PowerEdge 2650/3.06 GHz P4Xeon,1ML3	2	20.8	27.4	Oct-03	85%	180%
Dell PowerEdge 1750/3.2 GHz P4Xeon,1ML3	2	20.7	28.0	Oct-03	84%	184%
ION Computers SR2300WV2 (3.2GHz Xeon/2ML3,	2	20.6	31.7	Feb-04	84%	209%
Supermicro X6DHR-TG (3.6GHz, Intel Xeon )	2	20.5	28.1	Nov-04	84%	185%
HP ProLiant DL360 G3 3.2GHz Xeon,L3	2	20.1	28.2	Oct-03	82%	186%
IBM BladeCenter JS20 (PPC-970,2.2GHz)	2	20.0	21.5	Oct-04	82%	141%
IBM IntelliStation POWER 275 1.45GHz 2CC	2	19.9	20.0	Jun-03	81%	132%
ION Computers SR2300WV2 (3.2GHz Xeon/2ML3)	2	19.9	30.7	Feb-04	81%	202%
HP ProLiant DL360 G3 3.06GHz Xeon,L3	2	19.9	27.1	Jun-03	81%	178%
HP ProLiant BL20p G2 3.06GHzXeon,1ML3	2	19.9	26.8	Aug-03	81%	176%
HP ProLiant DL380 G3 3.06GHz Xeon,L3	2	19.9	26.7	Jun-03	81%	176%
Dell PowerEdge 1750/3.06 GHz Xeon L3	2	19.8	27.0	Jul-03	81%	178%
HP ProLiant BL20p G2 3.2GHz Xeon,1ML3	2	19.6	28.0	Oct-03	80%	184%
HP ProLiant DL380 G3 3.2GHz Xeon,L3	2	19.6	28.0	Oct-03	80%	184%
Supermicro X6DAR-8G (3.40 GHz Xeon, 1MBL3)	2	19.4	25.8	Jul-04	79%	170%
Sun Fire V490 /UltraSPARC-4 1050MHz	2	19.4	13.3	Oct-04	79%	88%
Dell PowerEdge 6650/2.8 GHz Xeon MP	2	19.3	27.0	Jun-03	79%	178%
HP 9000 RP4440-8 / PA8800-1.0GHz 2CC	2	19.3	18.6	Jan-04	79%	122%
HP 9000 RP3440-4 / PA8800-1.0GHz 2CC	2	19.2	18.7	Feb-04	78%	123%
HP ProLiant DL580 G2 2.8GHz Xeon MP	2	19.0	26.9	Sep-03	78%	177%
Dell Prec. WorkSt.650/3.06GHz P4Xeon 1ML3	2	18.7	28.5	Feb-04	76%	188%
Prec. WorkSt.650/3.2GHz P4Xeon 2ML3	2	18.6	33.6	Jan-04	76%	221%
Supermicro X6DHR-8XG (3.40 GHz Xeon 1MBL3)	2	18.6	26.2	Jun-04	76%	172%
Dell Prec. WorkSt.650/3.2GHz P4 Xeon L3	2	18.3	27.6	Sep-03	75%	182%
Sun Fire V250 / 1280MHz Cu	2	18.3	14.7	Sep-03	75%	97%
Sun Fire V240 / 1.28GHz Cu	2	18.2	14.7	Jan-04	74%	97%
HP ProLiant BL40p (2.8GHz Xeon MP)	2	18.2	26.9	Sep-03	74%	177%
Dell Prec. WorkSt.650/3.06GHz P4 Xeon L3	2	18.2	26.8	Jun-03	74%	176%
HP ProLiant DL560 (2.8GHz Xeon MP)	2	18.2	26.8	Sep-03	74%	176%
Dell Prec. WorkSt.650/2.8GHz P4Xeon 1ML3	2	18.0	26.7	Jan-04	73%	176%
IBM eServer pSeries 615/6C3 1.2 GHz 2CC	2	18.0	16.9	May-03	73%	111%
IBM eServer pSeries 615/6E3 1.2 GHz 2CC	2	18.0	16.9	May-03	73%	111%
Sun Blade Model 2500 / 1.28 GHz	2	17.7	14.4	Aug-03	72%	95%
HP ProLiant BL20p G2 3.06GHz Xeon	2	17.7	22.2	Aug-03	72%	146%
HP ProLiant DL380 G3 3.06GHz Xeon	2	17.6	22.4	Jun-03	72%	147%
ION Computers SR2300WV2 (3.2GHz Xeon/1ML3,	2	17.6	27.1	Feb-04	72%	178%
Dell PowerEdge 1750/3.06 GHz Xeon	2	17.4	22.5	Jul-03	71%	148%
HP ProLiant DL360 G3 3.06GHz Xeon	2	17.3	22.1	Jun-03	71%	145%
Dell PowerEdge 2650/3.06 GHz P4 Xeon	2	17.0	22.2	Apr-03	69%	146%
Dell Prec. WorkSt.650/2.4GHz P4Xeon 1ML3	2	17.0	23.8	Jan-04	69%	157%
Sun Blade Model 2000 / 1050 MHz	2	16.8	-	Apr-02	69%	-
Sun Fire 280R / 1015 MHz Cu	2	16.4	13.2	Oct-02	67%	87%
Compaq AlphaServer ES40 Model 6/833	2	16.4	13.0	Jun-01	67%	86%
ION Computers BX21 (2.8GHz Xeon/512kB L2,2	2	16.2	21.3	May-04	66%	140%
Sun Blade Model 2000 / 1015 MHz	2	16.1	13.2	Oct-02	66%	87%
Compaq AlphaServer DS20E Model 68/833	2	15.9	13.1	Jun-01	65%	86%
Sun Fire V880 / 900 MHz Cu	2	15.9	12.0	May-02	65%	79%
Dell Prec. WorkSt.650/3.06 GHz P4 Xeon	2	15.8	22.5	Apr-03	64%	148%
ION Computers SR2300WV2 (3.066GHz Xeon/512	2	15.5	22.0	Feb-04	63%	145%
Dell Prec. WorkSt.650/2.8 GHz P4 Xeon	2	15.5	21.3	Apr-03	63%	140%
Sun Fire V480 / 900 MHz Cu	2	15.5	12.2	Jun-02	63%	80%
Dell Prec. WorkSt.650/2.66 GHz P4 Xeon	2	15.3	20.7	Apr-03	62%	136%
HP ProLiant DL140 (3.2GHz Xeon,1ML3)	2	15.3	26.1	Nov-03	62%	172%
Sun Netra 20 / 900 MHz Cu	2	14.9	12.2	May-02	61%	80%
Sun Fire V240 / 1002MHz Cu	2	14.9	11.8	Mar-03	61%	78%
Sun Blade 1000 Model 2900 / 900 MHz	2	14.9	12.0	Nov-01	61%	79%
Dell Prec. WorkSt.650/2.4 GHz P4 Xeon	2	14.9	19.4	Apr-03	61%	128%
Sun Fire V210 / 1002MHz Cu	2	14.9	11.7	Mar-03	61%	77%
Dell Prec. WorkSt. 530/2.8 GHz P4 Xeon	2	14.7	19.6	Sep-02	60%	129%
Compaq AlphaServer ES40 Model 6/833*	2	14.4	12.6	Mar-01	59%	83%
Dell Prec. WorkSt. 530/2.6 GHz P4 Xeon	2	14.4	18.8	Sep-02	59%	124%
Sun Fire 280R / 900 MHz Cu	2	14.3	11.9	Nov-01	58%	78%
Compaq AlphaServer DS20L Model 68/833	2	14.3	12.2	May-02	58%	80%
Dell PowerEdge 7150 Itanium/800 4MB L3	2	14.2	-	Jun-01	58%	-
HP SERVER RX4610 Itanium/800 4MB L3	2	14.2	-	May-01	58%	-
Dell Prec. WorkSt. 530/2.4 GHz P4 Xeon*	2	14.2	17.7	Apr-02	58%	116%
Dell Prec. WorkSt. 530/2.4 GHz P4 Xeon	2	13.8	17.7	Sep-02	56%	116%
Dell PowerEdge 6650/2.0 GHz Xeon MP	2	13.8	18.7	Oct-02	56%	123%
Dell Prec. WorkSt.650/2.0 GHz P4 Xeon	2	13.8	17.2	Apr-03	56%	113%
Dell Prec. WorkSt. 530/2.2 GHz P4*	2	13.6	16.8	Jan-02	56%	111%
Dell Prec. WorkSt. 530/2.2 GHz P4	2	13.5	16.9	Sep-02	55%	111%
HP i2000 Itanium/800 2ML3	2	13.2	-	May-01	54%	-
Dell Prec. WorkSt. 530/2.0 GHz P4*	2	13.2	15.9	Jan-02	54%	105%
Dell PowerEdge 2650/2.8 GHz P4 Xeon	2	13.2	18.8	Oct-02	54%	124%
Dell Prec. WorkSt. 730/Itanium/800 2ML3	2	13.1	7.2	Jul-01	53%	47%
Dell Prec. WorkSt. 530/2.0 GHz P4	2	13.0	15.9	Sep-02	53%	105%
HP SERVER RX4610 Itanium/800 4MB L3*	2	13.0	-	Aug-01	53%	-
Dell PowerEdge 2650/2.4 GHz P4 Xeon	2	12.9	17.2	Apr-02	53%	113%
HP SERVER RX4610 Itanium/733 2ML3	2	12.7	-	May-01	52%	-
Dell Prec. WorkSt. 530/1.8 GHz P4	2	12.6	14.9	Sep-02	51%	98%
Dell Prec. WorkSt. 530/2.0 GHz P4**	2	12.4	15.7	Oct-01	51%	103%
Compaq AlphaServer DS20E Model 6/667	2	12.4	10.1	Aug-01	51%	66%
Supermicro X5DPA-GG (3.06GHz, 533MHz bus)	2	12.4	15.3	Jan-04	51%	101%
Compaq AlphaServer ES40 Model 6/667	2	12.2	10.3	Nov-00	50%	68%
Dell Prec. WorkSt. 530/1.7 GHz P4	2	11.9	12.2	Oct-01	49%	80%
HP SERVER RP5470/PA8700-750	2	11.5	12.5	Oct-01	47%	82%
HP 9000 Model J6750 / PA8700+/875	2	11.5	14.9	Oct-02	47%	98%
Dell Prec. WorkSt. 530/1.5 GHz P4	2	11.2	11.3	Oct-01	46%	74%
AMD Asus A7M266-D MP2400+ / 2000 MHz	2	11.2	15.4	Nov-02	46%	101%
AMD MSI K7D M-L MP2200+ / 1800 MHz	2	11.1	14.7	Aug-02	45%	97%
PrimePower200 SPARC64 GP/700MHz-8MB	2	11.1	11.5	Jul-02	45%	76%
AMD MSI K7D M-L MP2600+ / 2133 MHz	2	11.0	15.7	Feb-03	45%	103%
HP SERVER RP7400/PA8700-750	2	11.0	12.5	Sep-01	45%	82%
AMD MSI K7D M-L MP2100+ / 1733 MHz	2	11.0	14.4	Jun-02	45%	95%
Dell Prec. WorkSt. 530/1.7 GHz P4*	2	10.8	12.1	May-01	44%	80%
HP SERVER RP5470/PA8700-650	2	10.8	11.2	Oct-01	44%	74%
AMD Asus A7M266-D MP2000+ / 1667 MHz	2	10.6	13.9	Mar-02	43%	91%
Fujitsu PrimePower200 SPARC64 GP/700MHz-8MB	2	10.5	11.5	Jul-02	43%	76%
HP 9000 Model J6700 / PA8700-750	2	10.5	13.4	Apr-01	43%	88%
PrimePower200 SPARC64 GP/700MHz-4MB	2	10.5	10.9	Jul-02	43%	72%
HP SERVER RP7400/PA8700-650	2	10.4	11.3	Oct-01	42%	74%
Dell Prec. WorkSt. 530/1.5 GHz P4*	2	10.2	11.2	May-01	42%	74%
Sun Blade 1000 Model 2900 / 900MHz*	2	10.2	10.7	Sep-00	42%	70%
AMD Tyan Thunder K7 MP1800+/1533 MHz	2	10.1	13.0	Oct-01	41%	86%
Compaq Evo Workst. W6000/1.4 GHz P4 Xeon	2	10.0	10.6	Sep-01	41%	70%
Dell Prec. WorkSt. 530/1.4 GHz P4	2	10.0	10.7	May-01	41%	70%
AMD Tyan Thunder K7 MP1600+/1400 MHz	2	10.0	12.4	Oct-01	41%	82%
AMD Tyan Thunder K7 MP1500+/1333 MHz	2	9.8	12.1	Oct-01	40%	80%
Fujitsu PrimePower400 (SPARC64 GP/600MHz)	2	9.7	9.8	Sep-01	40%	64%
PrimePower400 (SPARC64 GP/600MHz)	2	9.7	9.8	Sep-01	40%	64%
PrimePower600 (SPARC64 GP/600MHz)	2	9.2	9.7	Sep-01	38%	64%
Fujitsu PrimePower600 (SPARC64 GP/600MHz)	2	9.2	9.7	Sep-01	38%	64%
AMD Tyan Thunder K7 Motherboard, 1.2 G	2	9.1	11.1	May-01	37%	73%
Compaq AlphaServer ES40 Model 6/500	2	9.0	7.1	Nov-00	37%	46%
Sun Blade 1000 Model 2750 / 750MHz*	2	9.0	9.0	Sep-00	37%	59%
IBM eServer pSeries 610/6E1 450MHz	2	9.0	7.7	Oct-01	37%	51%
IBM RS/6000 SP-450MHz T/W (2 CPU)	2	9.0	7.7	Jan-02	37%	51%
IBM eServer pSeries 610/6C1 450MHz	2	9.0	7.7	Oct-01	37%	51%
IBM RS/6000 44P-270 (450MHz, 8MBL2)	2	9.0	7.7	Oct-01	37%	51%
IBM eServer pSeries 640/B80 450MHz	2	9.0	7.7	Oct-01	37%	51%
Fujitsu PrimePower200 (SPARC64 GP/600MHz)	2	8.9	9.3	Sep-01	37%	61%
PrimePower200 (SPARC64 GP/600MHz)	2	8.9	9.3	Sep-01	37%	61%
Sun Netra 20 / 750 MHz	2	8.9	9.4	Nov-01	36%	62%
Sun Blade 1000 Model 2900 / 900MHz	2	8.9	10.7	Jul-01	36%	70%
Compaq AlphaServer DS20 Model 6/500	2	8.8	7.3	Mar-01	36%	48%
Sun Blade 1000 Model 2750 / 750MHz	2	8.6	9.1	Jun-01	35%	60%
Sun Fire 280R / 750 MHz*	2	8.5	9.1	Jun-01	35%	60%
HP 9000 Model L3000 / PA8600-550	2	8.3	8.9	Dec-00	34%	58%
IBM IntelliStation 265 (450MHz, 4MBL2)	2	8.2	7.3	Jan-02	33%	48%
HP 9000 model J6000 / PA8600-552	2	8.0	9.7	Oct-00	33%	64%
IBM RS/6000 44P-270 (2 CPU, 8MBL2)	2	8.0	6.2	Sep-00	33%	41%
IBM eServer pSeries 640 (8 MB L2)	2	8.0	6.2	Sep-00	33%	41%
Sun Fire 280R / 750 MHz	2	7.8	9.0	Jul-01	32%	59%
Sun Blade 1000 Model 2750 / 750MHz**	2	7.8	9.0	Jul-01	32%	59%
HP 9000 Model N4000 / PA8600-552 MHz	2	7.8	8.7	Mar-00	32%	57%
IBM eServer pSeries 610/6C1 375MHz	2	7.8	6.3	Oct-01	32%	42%
IBM eServer pSeries 610/6E1 375MHz	2	7.8	6.3	Oct-01	32%	42%
IBM eServer pSeries 640 (4 MB L2)	2	7.6	6.0	Sep-00	31%	39%
IBM RS/6000 44P-270 (2 CPU, 4MBL2)	2	7.6	6.0	Sep-00	31%	39%
HP 9000 Model A500 / PA8600-550 MHz	2	7.6	9.3	Nov-00	31%	61%
IBM RS/6000 44P-270 (2 CPU)	2	7.6	5.8	Jan-00	31%	38%
Sun Blade 1000 Model 2600 / 600MHz*	2	7.2	7.1	Sep-00	30%	47%
IBM eServer pSeries 610/6E1 333MHz	2	7.1	5.6	Feb-02	29%	37%
IBM eServer pSeries 610/6C1 333MHz	2	7.1	5.6	Feb-02	29%	37%
Dell PowerEdge 1500SC/1.4 GHz PIII	2	6.9	13.1	Jan-02	28%	86%
Dell PowerEdge 1500SC/1.26GHz PIII	2	6.8	12.4	Jan-02	28%	82%
SGI 2200 2X 400MHz R12k	2	6.7	7.8	May-00	27%	51%
Dell PowerEdge 2550/1.13GHz PIII	2	6.3	11.4	Jan-02	26%	75%
Sun Blade 1000 Model 2600 / 600MHz	2	6.3	7.1	Jul-01	26%	47%
Dell Prec. WorkSt. 420/1.0 GHz P3	2	6.2	9.6	Dec-00	25%	63%
Sun Enterprise 3500/4500 - 400 MHz	2	5.8	4.9	Oct-00	24%	32%
1 CPUs
HP Integrity RX1620/Itanium2 1.6G/3MB	1	31.2	16.8	Nov-04	127%	111%
HP Integrity RX2620/Itanium2 1.6G/6MB	1	31.0	17.8	Dec-04	127%	117%
HP Integrity RX2620/Itanium2 1.6G/3MB	1	29.6	16.3	Nov-04	121%	107%
HP Integrity RX4640/Itanium2 1.5G/4MB	1	29.0	15.9	Nov-04	118%	105%
HP Integrity RX2620/Itanium2 1.3G/3MB	1	25.9	13.6	Dec-04	106%	89%
HP Integrity RX1620/Itanium2 1.3G/3MB	1	25.7	13.7	Dec-04	105%	90%
HP Integrity RX4640/Itanium2,1.5GHz	1	25.1	16.3	Jan-04	102%	107%
HP Integrity RX2600/Itanium2,1.5GHz	1	24.6	15.3	Jun-03	100%	101%
HP Integrity RX5670/Itanium2,1.5GHz	1	24.5	15.2	Jun-03	100%	100%
HP ZX6000/Itanium2, 1.5GHz,6MB-L3, H	1	24.4	-	Jun-03	100%	-
HP Integrity RX4640/Itanium2,1.3GHz	1	21.9	13.1	Jan-04	89%	86%
Dell PowerEdge 3250/Itanium2 1.4GHz/3ML3	1	21.7	11.9	Mar-04	89%	78%
Dell PowerEdge3250/Itanium2 1.5GHz/6ML3	1	21.7	12.7	Jul-03	89%	84%
Fujitsu CELSIUS V810 / 2.4GHz Opteron 250 L	1	21.4	18.8	Nov-04	87%	124%
AMD Tyan K8QS Pro Opteron850/2.4GHz	1	21.4	-	Nov-04	87%	-
AMD Tyan K8QS Opteron850/2.4GHz SES8/P	1	21.4	18.7	Aug-04	87%	123%
ION Computers I2X2 (1.4GHz Itanium2)	1	21.1	10.8	Jul-03	86%	71%
HP Integrity RX5670/Itanium2,1.3GHz	1	21.0	12.4	Nov-03	86%	82%
HP Integrity RX2600/Itanium2,1.3GHz	1	21.0	12.4	Nov-03	86%	82%
Sun Java Wrkst. W1100z Opteron2.4GHz	1	20.7	18.4	Jul-04	84%	121%
IBM eServer 326 2.4 GHz Opteron EKO	1	20.5	18.4	Oct-04	84%	121%
AMD Tyan K8QS Opteron850/2.4GHz SES/P	1	19.9	-	Sep-04	81%	-
HP ProLiant DL145 Opteron250/2.4GHz	1	19.7	18.1	Oct-04	80%	119%
HP AlphaServer GS1280 7/1300	1	19.6	11.5	Jun-04	80%	76%
Sun Fire V20z (Opteron250/2.4GHz)	1	19.6	17.6	Jul-04	80%	116%
Sun Fire V40z (Opteron850/2.4GHz)	1	19.6	17.6	Jul-04	80%	116%
Supermicro Super P8SC8 (3.80 GHz Pentium 4)	1	19.5	19.9	Nov-04	80%	131%
AMD ASUS SK8V/Opteron150 2.4GHz SES	1	19.1	17.2	May-04	78%	113%
Supermicro P8SAA ( 3.6 GHz Pentium 4, 1MBL2)	1	18.7	19.0	Nov-04	76%	125%
HP ProLiant BL20p G3 3.6GHz Xeon	1	18.5	17.9	Nov-04	76%	118%
AMD Tyan K8QS Opteron850/2.4GHz W/PGI	1	17.8	-	Sep-04	73%	-
HP ProLiant ML350 G4 3.4GHz Xeon	1	17.8	16.8	Nov-04	73%	111%
HP ProLiant ML370 G4 3.6GHz Xeon	1	17.7	18.1	Oct-04	72%	119%
AMD ASUS SK8V/Opteron150 2.4GHz WS	1	17.7	19.2	May-04	72%	126%
AMD ASUS SK8N/Opteron148 2.2GHz SES	1	17.6	16.3	Nov-03	72%	107%
HP ProLiant DL380 G4 3.6GHz Xeon,L3	1	17.6	18.0	Jul-04	72%	118%
AMD ASUS SK8N/Opteron148 2.2GHz WS	1	17.3	17.1	Nov-03	71%	113%
HP AlphaServer ES80 7/1150	1	17.3	10.2	Jun-04	71%	67%
HP ProLiant ML370 G4 3.4GHz Xeon	1	17.3	17.3	Aug-04	71%	114%
HP AlphaServer ES47 7/1150	1	17.3	10.2	Jun-04	71%	67%
HP AlphaServer GS1280 7/1150	1	17.2	10.2	Jan-03	70%	67%
HP ProLiant DL380 G4 3.4GHz Xeon,L3	1	16.9	17.1	Jul-04	69%	113%
HP RX5670/Itanium2, 1GHz,3MB-L3, L	1	16.6	-	Nov-02	68%	-
HP RX2600/Itanium2, 1GHz,3MB-L3, L	1	16.6	-	Nov-02	68%	-
HP ZX6000/Itanium2, 1GHz,3MB-L3, L	1	16.5	-	Nov-02	67%	-
Fujitsu CELSIUS V810 (Opteron 248 2.2 GHz)	1	16.4	15.7	Feb-04	67%	103%
IBM x306 (3.0EGHz,Pentium4,1CC)	1	16.4	16.4	Mar-04	67%	108%
Dell PowerEdge3250/Itanium2 1.4GHz/1.5ML3	1	16.2	9.6	Oct-03	66%	63%
HP ProLiant DL360 G4 3.6GHz Xeon,L3	1	16.1	17.0	Oct-04	66%	112%
IBM x335 (3.2GHz, 533MHz FSB)	1	16.0	18.6	Feb-04	65%	122%
HP Integrity RX1600/Itanium2,1.0GHz	1	16.0	9.7	Feb-04	65%	64%
HP AlphaServer ES45 68/1250	1	15.8	10.8	Nov-02	64%	71%
HP ProLiant DL360 G4 3.4GHz Xeon,L3	1	15.7	16.3	Jul-04	64%	107%
HP ProLiant BL30p (3.2GHz Xeon,2MB L3)	1	15.7	18.0	Nov-04	64%	118%
HP ProLiant DL360 G3 3.2GHz/Xeon,2ML3 1	1	15.7	17.9	Apr-04	64%	118%
HP ProLiant DL380 G3 3.2GHz Xeon 1CC	1	15.6	17.8	Apr-04	64%	117%
Dell PowerEdge 2650/3.2 GHz P4Xeon,2ML3	1	15.6	16.5	Feb-04	64%	109%
Prec. WorkSt.650/3.2GHz P4Xeon 2ML3	1	15.6	18.1	Jan-04	64%	119%
HP ProLiant ML370 G3 (3.2GHz/2MB L3	1	15.3	17.7	Jun-04	62%	116%
HP RX5670/Itanium2, 1GHz,3MB-L3, H	1	15.1	9.4	Jul-02	62%	62%
HP RX2600/Itanium2, 1GHz,3MB-L3, H	1	15.1	9.4	Jul-02	62%	62%
HP AlphaServer ES80 7/1000	1	14.9	8.8	Jan-03	61%	58%
HP AlphaServer ES47 7/1000	1	14.9	8.8	Jan-03	61%	58%
Dell PowerEdge 6600/3.0 GHz Xeon MP	1	14.9	16.3	Feb-04	61%	107%
IBM eServer 325 2.0GHz Opteron 64-PGI	1	14.3	14.2	Sep-03	58%	93%
Dell PowerEdge 2650/3.2 GHz P4Xeon,1ML3	1	14.1	15.1	Oct-03	58%	99%
Dell Prec. WorkSt.650/3.2GHz P4 Xeon L3	1	14.1	15.5	Sep-03	58%	102%
AMD Einox A4800/Opteron144 1.8GHz-I,WS	1	14.1	13.6	Apr-03	58%	89%
HP ProLiant DL580 G2 3.0GHz XeonMP 1CC	1	14.0	17.3	Mar-04	57%	114%
HP ProLiant DL360 G3 3.2GHz Xeon,L3	1	14.0	15.3	Oct-03	57%	101%
Dell PowerEdge 2650/3.06 GHz P4Xeon,1ML3	1	13.9	14.6	Oct-03	57%	96%
HP ProLiant BL20p G2 3.2GHz Xeon,1ML3	1	13.9	15.3	Oct-03	57%	101%
HP ProLiant DL380 G3 3.2GHz Xeon,L3	1	13.9	15.2	Oct-03	57%	100%
Dell Prec. WorkSt.650/3.2GHz P4Xeon 2ML3	1	13.9	18.1	Jan-04	57%	119%
HP ProLiant BL20p G2 3.06GHzXeon,1ML3	1	13.8	14.6	Aug-03	56%	96%
Dell Prec. WorkSt.650/3.06GHz P4 Xeon L3	1	13.8	15.0	Jun-03	56%	99%
Dell PowerEdge 1750/3.2 GHz P4Xeon,1ML3	1	13.8	14.9	Oct-03	56%	98%
Intel D850EMVR(3.06GHz,Pentium4,MT)	1	13.7	14.1	Dec-02	56%	93%
HP ProLiant DL360 G3 3.06GHz Xeon,L3	1	13.7	14.5	Jun-03	56%	95%
HP ProLiant DL380 G3 3.06GHz Xeon,L3	1	13.7	14.5	Jun-03	56%	95%
IBM eServer 325 2.0GHz Opteron 64-PGI*	1	13.7	-	Aug-03	56%	-
Dell Prec. WorkSt.650/3.06GHz P4Xeon 1ML3	1	13.6	15.9	Feb-04	56%	105%
AMD Einox A4800/Opteron144 1.8GHz-I,SE	1	13.5	12.9	Apr-03	55%	85%
HP RX2600/Itanium2,900MHz,1.5MB-L3,L	1	13.4	-	Nov-02	55%	-
AMD Einox A4800/Opteron144 1.8GHz-I,SE	1	13.4	12.8	Apr-03	55%	84%
Dell PowerEdge 1750/3.06 GHz Xeon L3	1	13.4	14.4	Jul-03	55%	95%
HP RX5670/Itanium2,900MHz,1.5MB-L3,L	1	13.3	-	Nov-02	54%	-
HP ZX6000/Itanium2,900MHz,1.5MB-L3,L	1	13.2	-	Nov-02	54%	-
AMD Einox A4800/Opteron142 1.6GHz-I,WS	1	13.0	12.2	Apr-03	53%	80%
Dell PowerEdge 6650/2.8 GHz Xeon MP	1	13.0	14.3	Jun-03	53%	94%
Sun Blade Model 2000 / 1200 MHz	1	13.0	8.4	Apr-03	53%	55%
HP ProLiant BL40p (2.8GHz Xeon MP)	1	12.8	14.5	Sep-03	52%	95%
HP ProLiant DL560 (2.8GHz Xeon MP)	1	12.8	14.4	Sep-03	52%	95%
Dell Prec. WorkSt. 350/3.06 GHz P4 Xeon	1	12.8	13.1	Nov-02	52%	86%
Dell Prec. WorkSt.650/2.8GHz P4Xeon 1ML3	1	12.8	14.7	Jan-04	52%	97%
Intel D850EMVR(3.06GHz,Pentium4,MT)*	1	12.7	12.8	Dec-02	52%	84%
HP ProLiant DL580 G2 2.8GHz Xeon MP	1	12.7	14.3	Sep-03	52%	94%
HP ZX2000/Itanium2,900MHz,1.5MB-L3,L	1	12.6	-	Nov-02	51%	-
HP ProLiant BL20p G2 3.06GHz Xeon	1	12.5	12.5	Aug-03	51%	82%
Dell Prec. WorkSt. 350/2.8 GHz P4 Xeon	1	12.3	12.3	Nov-02	50%	81%
Dell Prec. WorkSt.650/3.06 GHz P4 Xeon	1	12.3	13.2	Apr-03	50%	87%
HP ProLiant DL360 G3 3.06GHz Xeon	1	12.2	12.4	Jun-03	50%	82%
HP ProLiant DL380 G3 3.06GHz Xeon	1	12.1	12.4	Jun-03	49%	82%
Dell PowerEdge 1750/3.06 GHz Xeon	1	12.1	12.4	Jul-03	49%	82%
Dell Prec. WorkSt. 350/2.66 GHz P4 Xeon	1	12.0	11.9	Nov-02	49%	78%
Dell Prec. WorkSt.650/2.4GHz P4Xeon 1ML3	1	12.0	13.0	Jan-04	49%	86%
Dell Prec. WorkSt.650/2.8 GHz P4 Xeon	1	11.8	12.4	Apr-03	48%	82%
Sun Blade Model 2500 / 1.28 GHz	1	11.8	8.1	Aug-03	48%	53%
HP AlphaServer GS320 68/1224	1	11.7	9.7	Nov-02	48%	64%
AMD Einox A4800/Opteron140 1.4GHz-I,WS	1	11.7	10.8	Apr-03	48%	71%
HP AlphaServer GS80 68/1224	1	11.7	9.7	Nov-02	48%	64%
Dell Prec. WorkSt. 350/2.53 GHz P4 Xeon	1	11.7	11.4	Nov-02	48%	75%
HP AlphaServer GS160 68/1224	1	11.7	9.7	Nov-02	48%	64%
Dell PowerEdge 2650/3.06 GHz P4 Xeon	1	11.6	12.2	Apr-03	47%	80%
Dell Prec. WorkSt.650/2.66 GHz P4 Xeon	1	11.6	12.0	Apr-03	47%	79%
Dell Prec. WorkSt. 350/2.4 GHz P4 Xeon	1	11.4	11.0	Nov-02	47%	72%
Dell Prec. WorkSt. 340/3.06 GHz P4 Xeon	1	11.4	12.5	Nov-02	47%	82%
HP ProLiant DL140 (3.2GHz Xeon,1ML3)	1	11.4	14.6	Nov-03	47%	96%
HP AlphaServer DS25 68/1000	1	11.4	7.9	Aug-02	47%	52%
Compaq AlphaServer ES45 Model 68/1000	1	11.1	7.9	Oct-01	45%	52%
Dell Prec. WorkSt. 350/2.26 GHz P4 Xeon	1	11.0	10.5	Nov-02	45%	69%
Dell Prec. WorkSt. 340/2.8 GHz P4 Xeon	1	11.0	11.7	Sep-02	45%	77%
Dell Prec. WorkSt.650/2.4 GHz P4 Xeon	1	11.0	11.1	Apr-03	45%	73%
Dell Prec. WorkSt. 340/2.66 GHz P4 Xeon	1	10.8	11.3	Sep-02	44%	74%
Dell Prec. WorkSt. 340/2.53 GHz P4 Xeon	1	10.5	10.9	Sep-02	43%	72%
Dell Prec. WorkSt. 340/2.53 GHz P4 Xeon*	1	10.4	10.7	May-02	42%	70%
Sun Blade 1500 / 1.062GHz	1	10.3	6.8	Jun-03	42%	45%
Dell Prec. WorkSt. 530/2.8 GHz P4 Xeon	1	10.3	11.1	Sep-02	42%	73%
Dell Prec. WorkSt. 340/2.4 GHz P4 Xeon	1	10.2	10.5	Sep-02	42%	69%
Dell Prec. WorkSt. 340/2.4 GHz P4 Xeon*	1	10.2	10.3	Apr-02	42%	68%
HP AlphaServer DS15/1000	1	10.2	7.3	Jun-04	42%	48%
Dell Prec. WorkSt.650/2.0 GHz P4 Xeon	1	10.1	9.7	Apr-03	41%	64%
Dell Prec. WorkSt. 530/2.6 GHz P4 Xeon	1	10.0	10.6	Sep-02	41%	70%
Sun Fire V1280 / 900MHz Cu	1	9.8	6.2	Feb-03	40%	41%
AMD ASUS A7N8X XP 2800+ / 2250 MHz	1	9.8	10.8	Sep-02	40%	71%
AMD ASUS A7N8X XP 2700+ / 2166 MHz	1	9.6	10.6	Sep-02	39%	70%
Dell Prec. WorkSt. 530/2.4 GHz P4 Xeon	1	9.6	10.0	Sep-02	39%	66%
Dell Prec. WorkSt. 340/2.2 GHz P4	1	9.5	9.6	Sep-02	39%	63%
Dell PowerEdge 2650/2.8 GHz P4 Xeon	1	9.4	10.5	Oct-02	38%	69%
Dell Prec. WorkSt. 340/2.2 GHz P4*	1	9.3	9.4	Jan-02	38%	62%
Dell Prec. WorkSt. 530/2.2 GHz P4	1	9.2	9.4	Sep-02	38%	62%
Compaq AlphaServer DS20E Model 68/833	1	9.1	6.6	Jun-01	37%	44%
Dell Prec. WorkSt. 340/2.0 GHz P4	1	9.1	9.1	Sep-02	37%	60%
Compaq AlphaServer ES40 Model 6/833	1	9.0	6.5	Jun-01	37%	43%
Dell Prec. WorkSt. 340/2.0A GHz P4	1	8.9	-	Jan-02	36%	-
Dell Prec. WorkSt. 530/2.0 GHz P4	1	8.8	8.8	Sep-02	36%	58%
Compaq AlphaServer GS80 Model 8 68/1001	1	8.8	7.2	Jun-01	36%	47%
Compaq AlphaServer GS320 Model 32 68/1001	1	8.8	-	Jun-01	36%	-
Compaq AlphaServer GS160 Model 16 68/1001	1	8.8	7.2	Jun-01	36%	47%
Dell PowerEdge 2650/2.4 GHz P4 Xeon	1	8.7	9.5	Apr-02	35%	62%
Compaq AlphaServer DS20L Model 68/833	1	8.4	6.2	Apr-02	34%	41%
Dell Prec. WorkSt. 530/1.8 GHz P4	1	8.3	8.2	Sep-02	34%	54%
Dell Prec. WorkSt. 340/1.8 GHz P4	1	8.2	7.5	Sep-02	34%	50%
Dell PowerEdge 7150 Itanium/800 4MB L3	1	8.2	-	Jun-01	33%	-
HP SERVER RX4610 Itanium/800 4MB L3	1	8.1	-	May-01	33%	-
Dell PowerEdge 6650/2.0 GHz Xeon MP	1	8.1	9.8	Oct-02	33%	64%
Dell PowerEdge 4600/2.0 GHz P4 Xeon	1	8.0	8.1	Jan-02	33%	53%
Compaq AlphaServer ES40 Model 6/833*	1	7.6	6.3	Oct-00	31%	42%
AMD Asus A7M266-D MP2400+ / 2000 MHz	1	7.6	8.9	Nov-02	31%	58%
HP i2000 Itanium/800 2ML3	1	7.6	-	May-01	31%	-
HP SERVER RX4610 Itanium/800 4MB L3*	1	7.6	-	Aug-01	31%	-
AMD MSI K7D M-L MP2600+ / 2133 MHz	1	7.5	9.1	Feb-03	31%	60%
AMD MSI K7D M-L MP2200+ / 1800 MHz	1	7.5	8.4	Aug-02	30%	55%
AMD MSI K7D M-L MP2100+ / 1733 MHz	1	7.4	8.3	Jun-02	30%	54%
Dell Prec. WorkSt. 330/1.8 GHz P4	1	7.3	-	Jul-01	30%	-
HP SERVER RX4610 Itanium/733 2ML3	1	7.2	-	May-01	29%	-
HP i2000 Itanium/733 2ML3	1	7.2	-	May-01	29%	-
Dell Prec. WorkSt. 330/1.7 GHz P4	1	7.0	-	May-01	29%	-
AMD Asus A7M266-D MP2000+ / 1667 MHz	1	6.9	7.7	Mar-02	28%	51%
Compaq AlphaServer DS20E Model 6/667	1	6.8	5.3	Aug-01	28%	35%
HP 9000 Model C3700 / PA8700-750	1	6.7	7.0	Apr-01	27%	46%
Dell Prec. WorkSt. 330/1.5 GHz P4	1	6.5	6.1	May-01	27%	40%
Compaq AlphaServer ES40 Model 6/667	1	6.5	5.0	Nov-00	27%	33%
HP ProLiant BL10e G2 1GHz,ULV PentiumM	1	6.4	8.0	Jul-03	26%	53%
Dell Prec. WorkSt. 330/1.5 GHz P4*	1	6.4	-	Dec-00	26%	-
AMD Tyan Thunder K7 MP1800+/1533 MHz	1	6.3	7.1	Oct-01	26%	46%
HP C3650/PA8700-625	1	6.3	5.9	Dec-01	26%	39%
Dell Prec. WorkSt. 330/1.4 GHz P4	1	6.3	5.8	May-01	26%	38%
Compaq Evo Workst. W6000/1.4 GHz P4 Xeon	1	6.2	5.9	Sep-01	25%	39%
Compaq AlphaStation XP1000 Model 6/667	1	6.2	4.7	Sep-01	25%	31%
Dell Prec. WorkSt. 330/1.4 GHz P4*	1	6.2	-	Dec-00	25%	-
AMD Tyan Thunder K7 MP1600+/1400 MHz	1	6.1	6.6	Oct-01	25%	44%
HP SERVER RP7400/PA8700-750	1	6.1	6.4	Sep-01	25%	42%
HP SERVER RP5470/PA8700-750	1	6.0	6.4	Oct-01	25%	42%
AMD Tyan Thunder K7 MP1500+/1333 MHz	1	6.0	6.4	Oct-01	24%	42%
HP SERVER RP7400/PA8700-650	1	5.7	5.7	Oct-01	23%	38%
HP SERVER RP5470/PA8700-650	1	5.7	5.7	Oct-01	23%	37%
AMD Tyan Thunder K7 Motherboard, 1.2 G	1	5.6	6.0	May-01	23%	40%
Dell PowerEdge 1500SC/1.4 GHz PIII	1	5.3	7.7	Jan-02	22%	51%
Compaq AlphaServer GS320 Model 6/731	1	5.2	4.6	Jun-00	21%	30%
Compaq AlphaServer GS160 Model 6/731	1	5.2	4.6	Jun-00	21%	30%
Compaq AlphaServer GS80 Model 6/731	1	5.2	4.6	Mar-01	21%	30%
Dell PowerEdge 1500SC/1.26GHz PIII	1	5.1	7.2	Jan-02	21%	48%
HP 9000 Model B2600 / PA8600-500	1	5.1	4.7	Apr-01	21%	31%
HP 9000 model C3600 / PA8600-552	1	5.0	5.0	Oct-00	20%	33%
Compaq AlphaServer ES40 Model 6/500	1	4.9	3.6	Nov-00	20%	24%
Compaq AlphaServer DS20 Model 6/500	1	4.8	3.7	Mar-01	20%	24%
Compaq AlphaServer DS10 6/600	1	4.8	4.2	Jun-00	19%	28%
HP 9000 Model L3000 / PA8600-550	1	4.4	4.5	Dec-00	18%	30%
HP 9000 model B2000 / PA8500-400	1	4.2	3.9	Oct-00	17%	25%
Dell PowerEdge 4400/1.0 GHz PIII Xeon	1	3.7	5.2	Jan-02	15%	34%
Sun Enterprise 3500/4500 - 400 MHz	1	3.0	2.5	Oct-00	12%	16%

Table 4. Vector Whetstone-97 Benchmark.

Vector Whetstone-97 Benchmark. Total CPU times (seconds), Mflop Ratings (see text) and MWIPS Ratings

Rank	Machine	Mflop ratings (Vl=1024)			Total CPU	MWIPS
		N2	N3	N8	(seconds)
1	IBM eServer p5 570/1.9 GHz	1966	1966	1625	6.2	6219
2	SunFire V20 2.2GHz (PathScale)	1311	1298	1481	7.7	4496
3	HP DL380 Pentium4/3600 (EM64T)	1966	1720	607	8.4	4408
4	Pentium4/3400 (EM64T)	1966	1529	569	9.0	4071
5	Pentium4/3400 (EM64T,32-bit)	1966	1529	569	9.1	4071
6	Pentium 4/3066 (ifc)	1966	529	1347	9.2	4071
7	Pentium4/3200 (EM64T)	1966	1376	542	9.5	3899
8	HP Superdome Itanium2/1500	492	3441	2907	9.8	3826
9	HP RX5670 Itanium2/1500-H	655	3441	2907	9.8	3855
10	AMD Opteron850/2400 (pgi5.2)	1966	430	1841	9.8	3826
11	HP RX1620 Itanium2/1600-H	655	3441	3682	10.1	3741
12	Pentium 4/2666 (ifc)	1966	444	1201	10.4	3532
13	AMD Opteron848/2200 (pgi5.2)	1966	405	1674	10.5	3570
14	SunFire V40z 2.2GHz (pgi5.2)	1966	405	1726	10.6	3532
15	HP DL145 Opteron248/2200 (pgi5.2)	1966	405	1674	10.7	3508
16	SunFire V20 2.2GHz (pgi5.2)	1966	393	1674	10.7	3496
17	HP DL140 Pentium 4/3200	1966	1059	552	10.8	3425
18	IBM pSeries 690Turbo/1.7 GHz	1996	475	1841	10.8	3472
19	Compaq Alpha ES45/1250	1679	815	1925	10.9	3441
20	HP RX4640 Itanium2/1300-H	492	2753	2511	11.3	3324
21	IBM Regatta-HPC/1300	492	444	1454	11.5	3281
22	IBM pSeries 690Turbo/1.3 GHz	1996	353	1905	11.7	3260
23	AMD Opteron848/2200 (pgi5.1)	1966	1147	1255	11.8	3158
24	HP RX2600 Itanium2/1500-L	2876	2876	2719	12.2	3089
25	HP BL20p G2 Pentium 4/2800	1966	983	489	12.4	2999
26	Intel Tiger Itanium2/1500	1549	2472	1481	12.5	3014
27	HP RX1620 Itanium2/1600-L	491	3064	3428	12.6	2991
28	HP RX4640 Itanium2/1600-L	491	3064	3428	13.0	2993
29	AMD Opteron846/2000 (pgi5.1)	983	1059	1127	13.0	2864
30	Compaq Marvel EV7/1000	1340	842	1637	13.0	2894
31	SGI Altix3700/Itanium2-1500	458	2876	3251	13.4	2803
32	Compaq Alpha ES45/1000	1343	653	1538	13.6	2778
33	SGI Altix3700/Itanium2-1300	2876	2472	2347	14.1	2676
34	AMD Opteron244/1800 (pgi5.1)	1966	918	1023	14.3	2626
35	AMD Opteron848/2200 (pgi5.0)	983	1147	1042	14.5	2593
36	IBM pSeries 630/1000	983	270	1416	15.4	2461
37	HP RX2600 Itanium2/1000-H	492	2294	1973	15.9	2369
38	AMD Opteron846/2000 (pgi5.0)	983	1059	952	15.9	2352
39	SGI Onyx 300/R14k-600	359	399	-	16.4	2280
40	AMD MP2400+/2000 (pgi)	1966	860	1042	16.9	2209
41	AMD Opteron244/1800 (pgi5.0)	983	918	863	17.7	2112
42	Intel Tiger Itanium2/1200 (+)	1258	1985	1183	18.7	2006
43	SGI O300/R14k-500	631	208	-	19.6	1914
44	AMD Opteron842/1600 (pgi5.0)	983	860	767	19.9	1888
45	SGI O3800/R14k-500	624	158	-	20.2	1851
46	AMD MP2000+/1667 (pgi)	983	459	863	20.3	1814
47	Compaq Alpha ES40/833	959	547	767	21.1	1777
48	Pentium 4/2000 (ifc)	655	336	801	21.2	1772
49	AMD MP1800+/1533 (pgi)	983	765	789	21.9	1710
50	HP RX5670 Itanium2/1000-L	1060	1658	1164	22.2	1675
51	Intel Tiger Itanium2/1000	1060	1639	985	22.4	1670
52	AMD Opteron840/1400 (pgi5.0)	983	724	674	22.7	1654
53	Pentium 4/2666 (pgi)	1966	810	1347	22.8	1638
54	API UP2000 6/833	1006	524	369	23.4	1593
55	AMD MP2400+/2000	983	529	921	23.6	1584
56	HP ZX6000 Itanium2/900	959	1484	1488	24.1	1568
57	AMD Athlon K7/1400(pgi)	983	626	717	24.3	1541
58	SGI Octane2/R12k-400	240	310	-	24.4	1526
59	SGI O3800/R12k-400	500	122	-	24.4	1472
60	HP 9000/RP3440-1000	492	655	1201	25.0	1506
61	Compaq PW XP1000/667	775	324	603	26.1	1433
62	Compaq Alpha DS20E/667	806	437	613	26.3	1429
63	API UP2000 6/667	805	430	607	26.4	1411
64	SGI Troon IA64/666	575	434	559	28.0	1484
65	HP 9000/RP7410-875	393	551	986	28.6	1303
66	Compaq Alpha ES40/667	775	376	582	28.6	1307
67	AMD Athlon K7/1200(pgi)	983	529	526	28.9	1290
68	HP 9000/J6700-750	983	918	877	29.3	1275
69	AMD MP2000+/1667	983	306	757	30.0	1239
70	HP Itanium/733-2M L3	1966	1147	1534	30.7	1230
71	Pentium 4/2533 (pgi)	983	430	354	31.6	1180
72	SUN Fire 6800/900-Cu	257	1141	-	32.3	1150
73	AMD MP1800+/1533	983	626	690	32.5	1139
74	SUN Blade 2000/1056-Cu	496	492	457	32.7	1131
75	SGI O2000/R12k-400	240	311	-	32.8	1135
76	IBM Itanium/800-4M L3	959	792	432	33.0	1140
77	Compaq PW XP1000/500	576	243	461	34.8	1074
78	Pentium 4/3066	1007	514	428	34.9	1067
79	AMD Athlon K7/1000(pgi)	492	265	476	35.1	1055
80	Compaq Alpha DS20/500	576	326	459	35.6	1055
81	AMD Athlon K7/1400	983	430	600	35.9	1029
82	AlphaPC 264DP-500	559	239	447	36.2	1034
83	Compaq Alpha GS140	593	298	470	36.9	1008
84	Compaq Alpha ES40/500	517	289	440	37.6	985
85	Compaq XP1000 6/450	480	269	408	38.0	982
86	Compaq Alpha DS10/466	517	217	419	38.2	980
87	IBM RS/6000 44P-270	179	205	449	38.6	967
88	SUN Fire V880/900-Cu	379	398	391	39.5	938
89	HP 9000/J6000-552	983	655	642	39.8	939
90	Pentium 4/2000 (pgi)	655	336	279	39.8	934
91	AMD Athlon K7/850 (pgi)	492	382	386	39.9	936
92	SGI O2000/R12k-300	180	124	-	39.9	936
93	DEC Alpha 8400/6-575	610	342	514	40.3	927
94	Pentium 4/2666	983	459	373	40.7	913
95	Pentium III/1000-CM(pgi)	492	212	409	41.1	905
96	Pentium 4/2533	655	430	352	41.8	892
97	AMD Athlon K7/1200	655	444	511	42.1	887
98	SGI Octane/R12k-270	162	111	-	44.5	839
99	SGI O2 R12k/270	169	148	-	46.4	801
100	AMD Athlon K7/1000	492	382	422	47.3	790
101	Pentium III/866-CM(pgi)	393	184	329	47.4	779
102	HP 9000/J5000-440	492	551	493	48.0	778
103	Pentium 4/2000	655	328	279	50.5	738
104	Pentium III/800-CM(pgi)	328	170	307	51.5	719
105	SUN Blade 1000/M1750	214	224	-	52.2	716
106	AMD Athlon K7/650 (pgi)	328	174	299	53.0	694
107	HP 9000/C3000-400	492	492	453	53.0	705
108	Pentium 4/1500 (pgi)	492	250	207	53.5	696
109	Pentium III/1000-CM	393	212	288	53.7	694
110	DEC Alpha 8400/5-625	469	264	245	54.9	675
111	SGI O2000/R10k-250	154	115	-	55.1	679
112	SGI Onyx2 IR2/R10k-250	154	116	-	55.1	678
113	Pentium III/750-CM(pgi)	328	156	282	55.1	669
114	Pentium III/733-CM(pgi)	328	153	282	56.1	658
115	SGI Octane/R10k-250	153	96	-	56.1	665
116	Pentium III/800 (pgi)	393	168	257	56.3	650
117	Pentium 4/1400 (pgi)	492	233	192	57.5	648
118	HP 9000/785-367	492	444	403	57.6	647
119	Pentium III/866-CM	393	170	295	57.9	640
120	DEC Alpha PW/600AU	458	257	266	58.1	637
121	AMD Athlon K7/850	218	320	339	58.4	632
122	HP 9000/N4000-440	246	115	682	59.3	628
123	DEC Alpha 1200/5-533	387	210	218	59.5	614
124	SGI Origin200/225	138	104	-	61.2	611
125	Pentium III/800	328	158	241	62.9	582
126	Pentium III/800-CM	328	158	271	63.4	584
127	Pentium III/650-CM(pgi)	281	136	244	63.6	580
128	SGI O2 R5k/300	95	76	-	66.7	558
129	Pentium III/750-CM	328	148	225	67.9	540
130	DEC Alpha 500/5-500	380	213	247	68.3	542
131	IBM RS/6000-43P	94	127	316	68.3	548
132	AMD Athlon K7/600 (pgi)	393	168	157	68.7	543
133	Pentium III/733-CM	328	142	220	69.8	524
134	AMD Athlon K7/500 (pgi)	281	94	226	76.6	483
135	Pentium 4/1500	492	246	208	76.9	486
136	Pentium III/650-CM	281	129	196	78.4	467
137	AMD Athlon K7/650	197	246	252	78.8	464
138	SGI Octane/R10k-195	120	89	147	79.5	452
139	SGI PChall-R10k/195	120	88	147	79.6	450
140	Pentium 4/1400	492	233	196	79.7	467
141	AMD Athlon K7/600	328	219	256	80.3	460
142	DEC Alpha PW/433AU	330	185	191	80.8	457
143	SUN Ultra80/450	406	256	578	81.1	459
144	SGI O2000/R10k-195	120	76	150	83.4	448
145	Pentium III/550 (pgi)	197	94	126	85.1	417
146	SGI Origin200/180	111	82	136	86.1	417
147	HP 9000/C240	328	287	261	87.4	428
148	SGI O2 R10k/175	111	77	191	87.5	424
149	SGI Octane/R10k-175	107	78	131	88.9	403
150	DEC Alpha 500/5-400	305	171	188	90.8	408
151	SUN HPC4500/400	361	228	514	91.5	407
152	Pentium III/550	197	90	109	96.6	362
153	AMD Athlon K7/500	281	110	217	97.2	379
154	HP 9000/V2250-240	123	63	371	107.1	375
155	DEC Alpha 8400/5-300	229	129	175	111.4	343
156	HP 9000/C200	109	113	708	112.0	337
157	SGI O2 R5k/180	50	60	-	112.6	331
158	IBM RS/6000-397	98	138	115	113.9	326
159	SUN HPC4500/336	40	40	360	115.8	321
160	ProLiant PII/450	164	73	95	116.3	303
161	DEC Alpha 600/5-333	198	53	149	117.4	313
162	Pentium II/400 (pgi)	123	65	74	125.1	276
163	HP 9000/C160	219	186	178	127.0	291
164	SUN Ultra30/300	35	35	80	130.9	286
165	SUN Ultra-2/300	35	35	317	131.7	282
166	Pentium II/400 (abs)	86	51	70	132.7	271
167	Pentium III/500	86	49	33	158.9	226
168	Pentium II/300 (abs)	61	39	49	174.8	201
169	Pentium II/450	82	49	27	175.6	201
170	Pentium II/266 (pgi)	86	42	51	179.4	189
171	Pentium II/266 (abs)	56	34	46	199.7	180
172	Pentium II/400	63	38	24	201.8	175
173	Pentium II/300	53	34	19	282.4	127
174	Pentium II/266	47	30	22	290.9	126
175	Dell Optiplex/266	44	27	19	328.7	111
176	HP 9000/755-99	36	44	31	356.3	105
177	SGI Indigo2 R4400/250	55	33	23	380.2	95
178	Pentium 233 MMX	33	24	12	397.8	84
179	Pentium Pro/200	25	18	16	416.8	90
	MPP nodes
1	IBM RS/6000-SP/375	197	202	449	38.5	970
2	IBM RS/6000-SP/222	109	123	266	63.8	585
3	Cray T3E/1200	193	173	475	77.1	485
4	IBM SP2/160Thin	131	156	341	101.1	370
5	Cray T3E/900	145	130	357	102.6	364
6	HP 9000/V2200	109	52	154	136.2	273
7	IBM SP2/120Thin	140	148	120	152.1	328
8	IBM SP2/77Wide	55	61	49	277.6	135
	Vector Supercomputers
1	NEC SX-6i	3932	5505	6904	4.3	13983
2	NEC SX-5	3932	4587	6905	5.7	12997
3	FUJITSU VPP-300	1175	1023	1158	18.9	3311

(+) Prototype, pre-release System




Table 5. The Matrix-97 Benchmark - Sparse MMO.

The Matrix-97 Benchmark - Sparse MMO. Total CPU times(seconds) for a Series of Sparse MMOs (R = A X B, see text)Implemented in Fortran.

Rank	Machine	Sparsity in B-Matrix
		0% (secs.)	50% (secs.)
1	HP RX1620 Itanium2/1600-H	0.6	0.3
2	IBM eServer p5 570/1.9 GHz	0.7	0.4
3	IBM pSeries 690Turbo/1.7 GHz	1.0	0.5
4	HP RX5670 Itanium2/1500-H	1.0	0.6
5	HP Superdome Itanium2/1500	1.1	0.6
6	HP RX4640 Itanium2/1600-L	1.1	0.6
7	HP RX1620 Itanium2/1600-L	1.1	0.6
8	HP RX4640 Itanium2/1300-H	1.2	0.7
9	SGI Altix3700/Itanium2-1500	1.2	0.6
10	IBM Regatta-HPC/1300	1.3	0.6
11	HP RX2600 Itanium2/1500-L	1.3	0.7
12	IBM pSeries 690Turbo/1.3 GHz	1.4	0.6
13	Intel Tiger Itanium2/1500	1.5	0.8
14	HP DL380 Pentium4/3600 (EM64T)	1.8	0.8
15	IBM pSeries 630/1000	1.8	0.9
16	HP RX2600 Itanium2/1000-H	1.9	0.9
17	Intel Tiger Itanium2/1200 (+)	1.9	1.0
18	SGI Altix3700/Itanium2-1300	2.0	0.8
19	SunFire V40z 2.2GHz (pgi5.2)	2.0	0.9
20	Pentium4/3400 (EM64T,32-bit)	2.1	0.9
21	AMD Opteron850/2400 (pgi5.2)	2.1	0.9
22	SunFire V20 2.2GHz (pgi5.2)	2.1	1.0
23	HP DL145 Opteron248/2200 (pgi5.2)	2.1	1.0
24	Pentium4/3400 (EM64T)	2.2	0.8
25	Pentium4/3200 (EM64T)	2.3	1.0
26	HP RX5670 Itanium2/1000-L	2.3	1.2
27	AMD Opteron846/2000 (pgi5.0)	2.3	0.9
28	Intel Tiger Itanium2/1000	2.3	1.2
29	AMD Opteron848/2200 (pgi5.1)	2.4	0.8
30	SunFire V20 2.2GHz (PathScale)	2.4	0.9
31	AMD Opteron848/2200 (pgi5.2)	2.5	0.8
32	AMD Opteron846/2000 (pgi5.1)	2.5	0.8
33	AMD Opteron848/2200 (pgi5.0)	2.6	1.0
34	AMD Opteron244/1800 (pgi5.1)	2.6	1.0
35	AMD Opteron244/1800 (pgi5.0)	2.7	1.2
36	Compaq Alpha ES45/1250	2.9	1.5
37	HP 9000/RP3440-1000	2.9	1.1
38	AMD Opteron842/1600 (pgi5.0)	3.1	1.3
39	HP DL140 Pentium 4/3200	3.3	1.0
40	HP ZX6000 Itanium2/900	3.4	1.4
41	HP BL20p G2 Pentium 4/2800	3.4	1.3
42	Pentium 4/3066 (ifc)	3.5	1.6
43	IBM RS/6000 44P-270	3.5	1.8
44	HP 9000/J6700-750	3.5	1.2
45	AMD Opteron840/1400 (pgi5.0)	3.5	1.6
46	Pentium 4/2666 (ifc)	3.6	1.5
47	Compaq Alpha ES45/1000	3.6	1.8
48	SGI Onyx 300/R14k-600	3.7	1.9
49	Compaq Marvel EV7/1000	3.8	1.6
50	HP 9000/RP7410-875	3.9	1.1
51	Pentium 4/2666 (pgi)	3.9	1.6
52	Compaq Alpha ES40/833	4.0	2.1
53	SGI O300/R14k-500	4.1	2.5
54	Pentium 4/3066	4.2	1.6
55	Pentium 4/2666	4.5	1.8
56	IBM Itanium/800-4M L3	4.6	2.0
57	Compaq Alpha ES40/667	4.9	2.6
58	Compaq Alpha DS20E/667	4.9	2.6
59	SGI O3800/R14k-500	5.0	2.7
60	Pentium 4/2533 (pgi)	5.0	2.2
61	SGI O2000/R12k-400	5.1	2.7
62	SUN Blade 2000/1056-Cu	5.2	2.6
63	Pentium 4/2000 (ifc)	5.2	2.2
64	Pentium 4/2533	5.4	2.3
65	HP Itanium/733-2M L3	5.4	2.2
66	Compaq PW XP1000/667	5.5	3.0
67	SGI Octane2/R12k-400	5.8	2.9
68	Pentium 4/1500 (pgi)	5.8	2.3
69	SGI O3800/R12k-400	5.9	3.1
70	SUN Fire V880/900-Cu	5.9	3.0
71	Pentium 4/1400 (pgi)	6.0	2.3
72	IBM RS/6000-43P	6.1	3.2
73	Pentium 4/2000 (pgi)	6.1	2.4
74	SUN Blade 1000/M1750	6.3	3.2
75	HP 9000/J6000-552	6.6	2.0
76	Pentium 4/2000	6.7	2.5
77	Pentium 4/1500	6.9	2.8
78	Compaq Alpha DS20/500	7.0	3.8
79	Compaq Alpha GS140	7.1	3.9
80	DEC Alpha 8400/6-575	7.1	3.8
81	Compaq PW XP1000/500	7.2	3.9
82	SGI O2000/R12k-300	7.3	3.9
83	API UP2000 6/833	7.5	3.2
84	AMD MP2400+/2000	7.7	3.8
85	AMD MP1800+/1533 (pgi)	7.7	3.9
86	Pentium 4/1400	7.7	3.0
87	Compaq Alpha ES40/500	7.7	4.1
88	AMD MP1800+/1533	7.9	3.9
89	HP 9000/J5000-440	7.9	3.3
90	AMD MP2400+/2000 (pgi)	8.3	3.9
91	Compaq XP1000 6/450	8.4	4.4
92	HP 9000/C3000-400	8.4	2.9
93	SGI Octane/R12k-270	8.6	4.6
94	AMD MP2000+/1667 (pgi)	8.6	4.1
95	AMD MP2000+/1667	8.7	4.1
96	HP 9000/C240	9.1	5.0
97	HP 9000/N4000-440	9.3	3.4
98	SGI Onyx2 IR2/R10k-250	9.4	5.0
99	API UP2000 6/667	9.5	4.2
100	SUN Fire 6800/900-Cu	9.7	4.9
101	SGI O2000/R10k-250	10.1	5.2
102	SGI Origin200/225	10.8	5.7
103	HP 9000/V2250-240	11.8	6.8
104	IBM RS/6000-397	11.9	6.3
105	AlphaPC 264DP-500	12.5	5.6
106	Compaq Alpha DS10/466	12.8	5.8
107	SGI Troon IA64/666	13.5	4.9
108	SGI O2000/R10k-195	13.7	7.4
109	AMD Athlon K7/1200(pgi)	14.4	7.2
110	DEC Alpha 8400/5-625	15.1	8.0
111	AMD Athlon K7/1400(pgi)	15.1	6.8
112	DEC Alpha 1200/5-533	15.2	7.9
113	DEC Alpha PW/600AU	15.7	8.3
114	AMD Athlon K7/1200	16.4	7.2
115	AMD Athlon K7/1400	16.5	6.4
116	HP 9000/785-367	16.9	5.7
117	HP 9000/C200	17.4	8.5
118	AMD Athlon K7/850 (pgi)	17.5	7.7
119	SUN Ultra80/450	17.6	9.0
120	DEC Alpha 500/5-500	18.5	9.9
121	SGI Octane/R10k-250	18.5	6.8
122	AMD Athlon K7/850	18.9	8.3
123	Pentium III/733-CM(pgi)	19.5	8.1
124	SGI Octane/R10k-195	19.5	7.6
125	SUN HPC4500/400	19.6	10.0
126	Pentium III/1000-CM(pgi)	19.7	9.0
127	SUN HPC4500/336	20.4	10.4
128	SGI Origin200/180	21.2	8.2
129	AMD Athlon K7/650 (pgi)	21.2	8.8
130	Pentium III/733-CM	21.6	9.5
131	SGI Octane/R10k-175	21.8	8.6
132	HP 9000/C160	21.9	10.2
133	Pentium III/800-CM(pgi)	22.2	10.3
134	Pentium III/1000-CM	22.3	10.1
135	AMD Athlon K7/600	22.6	9.8
136	AMD Athlon K7/600 (pgi)	22.6	10.0
137	AMD Athlon K7/1000(pgi)	22.6	9.7
138	Pentium III/866-CM(pgi)	22.9	9.8
139	Pentium III/750-CM(pgi)	23.4	10.7
140	AMD Athlon K7/1000	24.2	9.8
141	AMD Athlon K7/500 (pgi)	24.2	10.6
142	AMD Athlon K7/650	24.4	10.0
143	AMD Athlon K7/500	24.8	10.6
144	Pentium III/650-CM(pgi)	25.1	11.5
145	Pentium III/800 (pgi)	25.4	10.9
146	Pentium III/800	25.4	11.3
147	Pentium III/800-CM	25.4	12.4
148	Pentium III/550 (pgi)	25.6	11.1
149	DEC Alpha 8400/5-300	26.9	14.1
150	DEC Alpha PW/433AU	27.0	12.3
151	DEC Alpha 600/5-333	27.1	14.1
152	Pentium III/750-CM	27.4	12.7
153	Pentium III/550	27.5	11.9
154	Pentium III/866-CM	27.5	12.2
155	SUN Ultra-2/300	28.7	12.9
156	Pentium III/500	28.8	12.9
157	SUN Ultra30/300	29.2	13.4
158	Pentium III/650-CM	29.2	13.4
159	ProLiant PII/450	30.3	11.9
160	DEC Alpha 500/5-400	30.4	13.8
161	Pentium II/400 (pgi)	31.2	14.2
162	Pentium II/450	31.7	13.8
163	Pentium II/400	32.5	14.9
164	SGI PChall-R10k/195	35.2	12.7
165	Pentium II/400 (abs)	40.3	18.1
166	Pentium II/300	49.1	21.0
167	Pentium II/300 (abs)	51.8	21.7
168	SGI O2 R12k/270	52.6	13.4
169	Pentium II/266	60.7	25.5
170	Dell Optiplex/266	62.1	26.3
171	SGI Indigo2 R4400/250	65.0	32.4
172	Pentium II/266 (abs)	68.5	29.8
173	SGI O2 R5k/300	72.4	28.1
174	Pentium II/266 (pgi)	74.2	32.6
175	SGI O2 R10k/175	74.6	19.2
176	Pentium Pro/200	86.0	41.6
177	Pentium 233 MMX	93.4	44.5
178	SGI O2 R5k/180	104.7	46.6
179	HP 9000/755-99	122.2	55.6
	MPP nodes
1	IBM RS/6000-SP/375	3.3	1.7
2	IBM RS/6000-SP/222	6.9	3.5
3	HP 9000/V2200	9.7	5.3
4	IBM SP2/160Thin	11.6	6.3
5	Cray T3E/1200	14.6	7.5
6	IBM SP2/120Thin	15.1	7.9
7	Cray T3E/900	16.3	8.3
8	IBM SP2/77Wide	30.0	15.8
	Vector Supercomputers
1	NEC SX-6i	0.5	0.3
2	NEC SX-5	0.6	0.5
3	FUJITSU VPP-300	3.7	2.5

(+) Prototype, pre-release System




Table 6. The Matrix-97 Benchmark - Similarity Transformation.

The Matrix-97 Benchmark - Sparse MMO. Total CPU times(seconds) for a Series of Similarity Transformations (H=Q*HQ, see text) Using Both Scalar and Vector Algorithms.

Rank	Machine	Algorithm
		Scalar (secs.)	Vector (secs.)
1	HP RX1620 Itanium2/1600-L	16.5	4.7
2	HP RX4640 Itanium2/1600-L	16.5	4.7
3	HP RX1620 Itanium2/1600-H	9.0	5.1
4	HP RX2600 Itanium2/1500-L	18.1	5.3
5	IBM eServer p5 570/1.9 GHz	10.3	5.3
6	Intel Tiger Itanium2/1500	19.4	5.4
7	SGI Altix3700/Itanium2-1500	17.8	5.4
8	HP Superdome Itanium2/1500	9.7	5.5
9	HP RX5670 Itanium2/1500-H	9.7	5.6
10	HP RX4640 Itanium2/1300-H	11.1	6.3
11	SGI Altix3700/Itanium2-1300	20.5	6.4
12	HP DL380 Pentium4/3600 (EM64T)	12.1	6.7
13	Intel Tiger Itanium2/1200 (+)	22.2	7.0
14	Pentium4/3400 (EM64T)	13.2	7.2
15	Pentium4/3200 (EM64T)	14.1	7.5
16	IBM pSeries 690Turbo/1.7 GHz	16.9	7.7
17	Pentium4/3400 (EM64T,32-bit)	14.0	7.7
18	HP DL140 Pentium 4/3200	13.6	7.7
19	HP RX2600 Itanium2/1000-H	14.4	8.0
20	Intel Tiger Itanium2/1000	26.7	8.3
21	HP 9000/RP3440-1000	21.0	9.0
22	HP RX5670 Itanium2/1000-L	26.5	9.4
23	AMD Opteron850/2400 (pgi5.2)	13.3	10.0
24	IBM Regatta-HPC/1300	18.1	10.0
25	AMD Opteron848/2200 (pgi5.2)	13.1	10.1
26	IBM pSeries 690Turbo/1.3 GHz	22.3	10.2
27	SunFire V20 2.2GHz (PathScale)	15.0	10.3
28	HP DL145 Opteron248/2200 (pgi5.2)	13.9	10.4
29	SunFire V20 2.2GHz (pgi5.2)	13.3	10.4
30	SunFire V40z 2.2GHz (pgi5.2)	13.0	10.5
31	AMD Opteron848/2200 (pgi5.1)	13.6	10.8
32	HP 9000/RP7410-875	23.9	11.2
33	HP BL20p G2 Pentium 4/2800	22.1	11.2
34	Pentium 4/3066 (ifc)	24.5	11.3
35	AMD Opteron848/2200 (pgi5.0)	13.9	11.4
36	AMD Opteron846/2000 (pgi5.0)	14.3	11.7
37	AMD Opteron846/2000 (pgi5.1)	15.0	11.8
38	HP 9000/J6700-750	27.9	12.2
39	HP ZX6000 Itanium2/900	29.6	13.1
40	AMD Opteron244/1800 (pgi5.1)	16.3	13.2
41	Pentium 4/2666 (ifc)	27.0	13.3
42	IBM pSeries 630/1000	29.4	13.5
43	HP Itanium/733-2M L3	34.6	13.7
44	Compaq Alpha ES45/1250	20.9	14.2
45	Pentium 4/3066	26.9	14.3
46	AMD Opteron842/1600 (pgi5.0)	19.9	15.5
47	AMD Opteron244/1800 (pgi5.0)	17.0	15.9
48	Pentium 4/2666	31.3	16.0
49	IBM Itanium/800-4M L3	46.2	16.0
50	Pentium 4/2666 (pgi)	24.3	16.9
51	HP 9000/J6000-552	37.6	17.2
52	AMD Opteron840/1400 (pgi5.0)	23.0	17.6
53	Compaq Alpha ES45/1000	26.0	17.9
54	Pentium 4/2533 (pgi)	42.3	18.0
55	Pentium 4/2533	30.6	18.3
56	SUN Blade 2000/1056-Cu	29.3	18.9
57	Compaq Marvel EV7/1000	27.1	19.1
58	SUN Fire V880/900-Cu	33.4	20.7
59	Compaq Alpha ES40/833	36.4	20.7
60	SUN Fire 6800/900-Cu	33.8	20.9
61	HP 9000/J5000-440	74.7	22.5
62	HP 9000/N4000-440	86.0	22.8
63	API UP2000 6/833	45.3	23.4
64	SUN Blade 1000/M1750	62.3	25.4
65	Compaq Alpha DS20E/667	45.4	25.7
66	SGI Onyx 300/R14k-600	45.8	26.3
67	Pentium 4/2000 (ifc)	48.5	26.9
68	SGI Troon IA64/666	39.8	28.9
69	IBM RS/6000 44P-270	43.6	26.9
70	AMD MP2400+/2000	98.2	30.9
71	API UP2000 6/667	49.2	31.0
72	AMD MP2400+/2000 (pgi)	89.7	31.7
73	Pentium 4/2000	53.3	32.1
74	Pentium 4/1500	52.7	32.6
75	Pentium 4/1500 (pgi)	71.5	32.8
76	AMD MP1800+/1533	101.2	32.9
77	Pentium 4/2000 (pgi)	64.4	33.0
78	AMD MP1800+/1533 (pgi)	96.7	33.0
79	SGI O3800/R14k-500	59.3	33.0
80	SGI O300/R14k-500	58.1	34.0
81	AMD MP2000+/1667	104.7	34.1
82	Compaq PW XP1000/667	47.9	34.3
83	AMD MP2000+/1667 (pgi)	95.2	34.3
84	Compaq Alpha DS20/500	61.7	36.1
85	Pentium 4/1400 (pgi)	76.2	36.8
86	Pentium 4/1400	63.2	38.5
87	Compaq Alpha ES40/667	48.0	38.6
88	AMD Athlon K7/1400	139.4	38.7
89	SGI O3800/R12k-400	69.7	39.0
90	AMD Athlon K7/1400(pgi)	116.8	39.2
91	SGI Octane2/R12k-400	68.5	39.7
92	HP 9000/C3000-400	52.7	39.9
93	AMD Athlon K7/1200(pgi)	113.6	40.6
94	SUN Ultra80/450	76.7	40.8
95	AMD Athlon K7/1200	124.9	41.3
96	HP 9000/V2250-240	115.7	42.2
97	SGI O2000/R12k-400	66.2	43.7
98	DEC Alpha 8400/6-575	56.6	44.9
99	SUN HPC4500/400	83.3	46.7
100	IBM RS/6000-43P	82.9	48.2
101	SUN HPC4500/336	99.8	51.1
102	HP 9000/785-367	68.4	51.4
103	Compaq XP1000 6/450	72.4	52.8
104	Compaq Alpha GS140	60.9	54.6
105	AlphaPC 264DP-500	67.9	54.7
106	Compaq Alpha ES40/500	62.3	54.8
107	Compaq PW XP1000/500	63.7	57.3
108	SGI O2000/R12k-300	88.0	58.6
109	AMD Athlon K7/1000	160.1	59.1
110	AMD Athlon K7/1000(pgi)	148.1	61.0
111	AMD Athlon K7/850 (pgi)	155.6	62.6
112	SUN Ultra-2/300	129.6	64.2
113	IBM RS/6000-397	89.0	64.7
114	Compaq Alpha DS10/466	74.6	65.6
115	SGI Octane/R12k-270	100.0	66.1
116	AMD Athlon K7/850	134.7	66.6
117	Pentium III/1000-CM(pgi)	167.5	66.7
118	Pentium III/1000-CM	170.0	67.1
119	HP 9000/C240	103.9	67.3
120	AMD Athlon K7/650 (pgi)	201.5	71.8
121	DEC Alpha 1200/5-533	184.4	73.7
122	DEC Alpha 8400/5-625	101.9	73.9
123	Pentium III/733-CM(pgi)	202.9	74.1
124	Pentium III/800-CM(pgi)	173.8	75.2
125	SGI Onyx2 IR2/R10k-250	108.7	75.5
126	DEC Alpha PW/600AU	104.8	75.9
127	AMD Athlon K7/650	161.0	78.3
128	Pentium III/866-CM	193.5	78.6
129	SGI O2000/R10k-250	123.6	78.8
130	AMD Athlon K7/600	163.5	79.7
131	Pentium III/750-CM(pgi)	218.4	81.2
132	Pentium III/866-CM(pgi)	176.5	81.7
133	Pentium III/800-CM	170.6	81.7
134	Pentium III/733-CM	204.1	82.9
135	HP 9000/C200	151.1	82.9
136	SGI Origin200/225	121.3	83.7
137	Pentium III/750-CM	243.4	84.9
138	SGI Octane/R10k-250	118.7	85.6
139	SGI O2000/R10k-195	149.8	87.2
140	Pentium III/650-CM	247.9	91.1
141	SUN Ultra30/300	153.7	91.5
142	SGI O2 R12k/270	140.5	91.9
143	AMD Athlon K7/500	185.2	91.9
144	Pentium III/550	186.3	92.0
145	DEC Alpha PW/433AU	143.9	92.0
146	SGI Octane/R10k-195	156.3	92.7
147	ProLiant PII/450	186.9	92.8
148	Pentium III/550 (pgi)	285.1	93.2
149	AMD Athlon K7/500 (pgi)	248.2	93.3
150	AMD Athlon K7/600 (pgi)	235.1	94.1
151	Pentium III/800	223.8	94.7
152	Pentium III/650-CM(pgi)	232.6	98.4
153	Pentium III/800 (pgi)	308.6	98.9
154	SGI Origin200/180	169.3	100.6
155	SGI PChall-R10k/195	165.1	103.2
156	SGI Octane/R10k-175	176.1	104.2
157	DEC Alpha 500/5-400	151.0	105.7
158	HP 9000/C160	186.0	107.4
159	DEC Alpha 500/5-500	125.2	107.4
160	Pentium III/500	292.5	108.1
161	DEC Alpha 600/5-333	226.2	113.1
162	DEC Alpha 8400/5-300	181.0	113.4
163	Pentium II/450	314.3	116.9
164	Pentium II/400	365.6	130.3
165	SGI O2 R10k/175	201.8	131.5
166	Pentium II/400 (pgi)	392.9	132.0
167	Pentium II/400 (abs)	279.4	132.0
168	Pentium II/300	485.5	184.1
169	Pentium II/300 (abs)	354.3	194.2
170	Pentium II/266	516.4	203.2
171	Pentium II/266 (abs)	414.1	208.0
172	Pentium II/266 (pgi)	561.2	211.6
173	SGI O2 R5k/300	441.9	338.2
174	Dell Optiplex/266	523.6	442.0
175	HP 9000/755-99	892.3	469.5
176	SGI O2 R5k/180	651.0	498.5
177	Pentium Pro/200	1002.7	590.2
178	Pentium 233 MMX	782.7	844.4
	MPP nodes
1	IBM RS/6000-SP/375	43.5	26.1
2	HP 9000/V2200	158.3	44.1
3	IBM RS/6000-SP/222	75.8	49.4
4	IBM SP2/160Thin	103.2	64.9
5	IBM SP2/120Thin	114.2	84.3
6	Cray T3E/1200	178.9	86.4
7	Cray T3E/900	187.9	94.7
8	IBM SP2/77Wide	231.0	164.7
	Vector Supercomputers
1	NEC SX-6i	36.4	3.8
2	NEC SX-5	61.4	4.5
3	FUJITSU VPP-300	115.1	41.6

(+) Prototype, pre-release System




Table 7. The Matrix-97 Benchmark - Diagonalization.

The Matrix-97 Benchmark - Diagonalization. Total CPU times (seconds) for a Series of Matrix Diagonalizations (see text) Using Eight Different Routines.

Rank	Machine	CPU Time	CPU Time	Compiler
		(seconds)	(-JACO)	Options
1	HP DL380 Pentium4/3600 (EM64T)	5.7	2.8	-fast
2	Pentium4/3400 (EM64T)	6.0	3.0	-fast
3	IBM eServer p5 570/1.9 GHz	6.0	3.0	-O5 qarch=pwr5
4	HP RX1620 Itanium2/1600-H	6.3	2.3	+Ofaster +O3 +DSitanium2
5	Pentium4/3200 (EM64T)	6.4	3.2	-fast
6	AMD Opteron850/2400 (pgi5.2)	6.5	3.2	-fastsse -Mipa=safe
7	IBM pSeries 690Turbo/1.7 GHz	6.6	3.5	-O4 -qipa=level=2:noobject
8	AMD Opteron848/2200 (pgi5.1)	6.6	3.3	-fastsse -Minline -Mipa
9	HP DL140 Pentium 4/3200	6.7	3.7	-O3 -w -ipo -tpp7 -fno-alias
10	SunFire V20 2.2GHz (PathScale)	6.8	3.6	-O3
11	AMD Opteron848/2200 (pgi5.2)	6.8	3.3	-fastsse -Mipa=safe
12	SunFire V40z 2.2GHz (pgi5.2)	7.0	3.4	-tp=k8-64 -fastsse -Mipa=safe
13	Pentium4/3400 (EM64T,32-bit)	7.0	3.8	-fast
14	SunFire V20 2.2GHz (pgi5.2)	7.0	3.4	-tp=k8-64 -fastsse -Mipa=safe
15	HP DL145 Opteron248/2200 (pgi5.2)	7.1	3.5	-fastsse -Mipa=safe
16	HP RX5670 Itanium2/1500-H	7.1	2.9	+Ofaster +O3 +DSitanium2
17	HP Superdome Itanium2/1500	7.1	2.9	+Ofaster +O3 +DSitanium2
18	AMD Opteron846/2000 (pgi5.1)	7.3	3.6	-fastsse -Minline -Mipa
19	AMD Opteron848/2200 (pgi5.0)	7.5	4.1	-fastsse -Minline -Mipa
20	HP RX4640 Itanium2/1600-L	7.5	3.5	-O3 -ipo -tpp2 -fno-alias
21	HP RX1620 Itanium2/1600-L	7.6	3.5	-O3 -ipo -tpp2 -fno-alias
22	SGI Altix3700/Itanium2-1500	8.0	3.7	-O3 -ipo -w -tpp2
23	AMD Opteron244/1800 (pgi5.1)	8.1	4.1	-fastsse -Minline -Mipa
24	HP RX4640 Itanium2/1300-H	8.2	3.4	+Ofaster +O3 +DSitanium2
25	AMD Opteron846/2000 (pgi5.0)	8.2	4.5	-fastsse -Minline -Mipa
26	HP RX2600 Itanium2/1500-L	8.5	4.4	-O3 -ipo -w -tpp2
27	Intel Tiger Itanium2/1500	8.5	4.3	-O3 -ipo -tpp2 -fno-alias
28	IBM pSeries 690Turbo/1.3 GHz	8.7	4.6	-O4 -qipa=level=2:noobject
29	HP 9000/RP3440-1000	8.7	3.7	+Ofast +DA2.0
30	IBM Regatta-HPC/1300	8.9	4.7	-O4 -qarch=auto
31	Pentium 4/3066 (ifc)	9.2	4.9	-O3 -ipo -tpp7 -fno-alias
32	AMD Opteron244/1800 (pgi5.0)	9.3	5.1	-fastsse
33	Pentium 4/2666 (ifc)	9.7	5.1	-O3 -ipo -tpp7 -fno-alias
34	Compaq Alpha ES45/1250	9.8	4.5	-fast -arch ev68
35	HP BL20p G2 Pentium 4/2800	9.8	5.3	-O3 -w -ipo -tpp7 -fno-alias
36	SGI Altix3700/Itanium2-1300	9.8	5.0	-O3 -ipo -w -tpp2
37	HP 9000/RP7410-875	10.0	4.2	+O3 +Oaggressive +DA2.0
38	HP RX2600 Itanium2/1000-H	10.2	3.9	+Ofaster +O3 +DSitanium2
39	Compaq Marvel EV7/1000	10.3	5.1	-fast -arch ev68
40	AMD Opteron842/1600 (pgi5.0)	10.4	5.6	-fastsse -Minline -Mipa
41	Intel Tiger Itanium2/1200 (+)	10.7	5.4	-O3 -ipo -w -tpp2
42	Pentium 4/3066	10.9	5.7	-O3 -malign-double
43	Pentium 4/2666 (pgi)	10.9	5.8	-fast -Mdalign
44	IBM pSeries 630/1000	11.5	6.2	-O4 -qipa=level=2:noobject
45	HP 9000/J6700-750	11.8	4.9	+O3 +Oaggressive +DA2.0
46	AMD Opteron840/1400 (pgi5.0)	11.8	6.4	-fastsse -Minline -Mipa
47	Compaq Alpha ES45/1000	11.9	5.5	-fast -arch ev68
48	Pentium 4/2533 (pgi)	12.0	6.3	-fast -Mdalign
49	Pentium 4/2666	12.4	6.2	-O3 -malign-double
50	HP RX5670 Itanium2/1000-L	12.4	6.1	-O3 -ipo -tpp2
51	Intel Tiger Itanium2/1000	12.8	6.5	-O3 -ipo -tpp2
52	Pentium 4/2533	13.0	6.2	-O3 -malign-double
53	Compaq Alpha ES40/833	13.8	6.6	-fast -arch ev67
54	HP ZX6000 Itanium2/900	14.1	6.9	-O3 -ipo -tpp2
55	HP 9000/J6000-552	16.1	6.7	+O3 +Oaggressive +DA2.0
56	Compaq Alpha DS20E/667	17.3	8.3	-fast -arch ev67
57	Compaq Alpha ES40/667	17.4	8.5	-fast -arch EV6
58	API UP2000 6/833	17.8	8.6	-fast -arch ev67 -tune ev67
59	Pentium 4/2000 (ifc)	17.9	8.7	-O3 -ipo ifpp -tpp7
60	SGI Onyx 300/R14k-600	19.0	9.8	-Ofast=ip35 -LNO:fusion=2
61	AMD MP2400+/2000 (pgi)	19.4	10.3	-fast -Mdalign
62	Pentium 4/2000 (pgi)	20.1	9.8	-fast -Mdalign
63	HP 9000/N4000-440	20.3	8.6	+O3 +DA2.0
64	AMD MP2400+/2000	20.4	10.6	-O3 -malign-double
65	Pentium 4/2000	20.6	9.4	-O3 -malign-double
66	HP 9000/J5000-440	20.7	9.1	+O3 +Oaggressive +DA2.0
67	Pentium 4/1500 (pgi)	21.0	10.5	-fast -Mdalign
68	Compaq PW XP1000/667	21.1	10.2	-fast -arch ev6
69	AMD MP2000+/1667 (pgi)	21.6	11.7	-fast -Mdalign
70	SUN Blade 2000/1056-Cu	21.7	9.8	-fast -xarch=v8plusb -xchip=ultra3
71	HP Itanium/733-2M L3	21.9	8.9	+Ofast +DSitanium
72	AMD MP1800+/1533 (pgi)	22.1	11.8	-fast -Mdalign
73	API UP2000 6/667	22.2	10.5	-fast -arch ev67 -tune ev67
74	SGI O300/R14k-500	22.2	11.2	-Ofast=ip35 -LNO:fusion=2
75	AMD MP2000+/1667	22.3	11.6	-O3 -malign-double
76	Pentium 4/1400 (pgi)	22.4	11.1	-fast -Mdalign
77	HP 9000/C3000-400	22.5	9.8	+O3 +Oaggressive +DA2.0
78	SGI O3800/R14k-500	22.7	11.4	-Ofast=ip35 -LNO:fusion=2
79	Pentium 4/1500	23.1	10.7	-O3 -malign-double
80	AMD MP1800+/1533	23.4	12.0	-O3 -malign-double
81	IBM RS/6000 44P-270	23.7	10.7	-O3 -qhot
82	IBM Itanium/800-4M L3	24.5	12.6	-O2 -ipo
83	SUN Fire V880/900-Cu	24.6	11.5	-fast -xarch=v8plusb -xchip=ultra3
84	Pentium 4/1400	24.7	11.2	-O3 -malign-double
85	Compaq Alpha DS20/500	24.9	11.7	-fast -arch ev6
86	SUN Fire 6800/900-Cu	25.7	12.6	-fast -xarch=v8plusb -xchip=ultra3
87	Compaq Alpha GS140	26.1	12.6	-fast -arch ev6
88	SGI Octane2/R12k-400	26.9	13.8	-Ofast=ip30 -LNO:fusion=2
89	DEC Alpha 8400/6-575	27.1	12.4	-fast
90	SGI O2000/R12k-400	27.1	14.3	-Ofast=ip27 -LNO:fusion=2
91	Compaq Alpha ES40/500	27.2	11.8	-fast -arch ev6
92	AMD Athlon K7/1400(pgi)	27.5	15.2	-fast -Mdalign
93	HP 9000/785-367	27.5	11.8	+O3 +Oaggressive +DA2.0
94	SGI O3800/R12k-400	27.6	14.1	-Ofast=ip35 -LNO:fusion=2
95	AlphaPC 264DP-500	27.9	12.3	-arch host -tune ev6 -O5
96	Compaq PW XP1000/500	28.0	13.3	-fast -arch EV6
97	AMD Athlon K7/1400	28.5	15.2	-O3 -malign-double
98	Compaq XP1000 6/450	28.9	13.4	-fast -arch ev6
99	AMD Athlon K7/1200(pgi)	29.4	16.0	-fast -Mdalign
100	SUN Blade 1000/M1750	30.3	14.5	-fast -xarch=v9 -xchip=ultra2 -xO5
101	AMD Athlon K7/1200	30.7	16.1	-O3 -malign-double
102	Compaq Alpha DS10/466	31.2	15.3	-fast -arch ev6
103	SGI O2000/R12k-300	35.1	17.8	-Ofast=ip27 -LNO:fusion=2
104	SUN Ultra80/450	37.9	19.2	-fast -O5 -xsafe=mem -fsimple=2
105	Pentium III/1000-CM(pgi)	38.0	21.3	-fast -Mdalign
106	AMD Athlon K7/1000(pgi)	38.4	20.8	-fast -Mdalign
107	AMD Athlon K7/850 (pgi)	38.6	20.1	-O -Mdalign
108	SGI Octane/R12k-270	39.0	19.8	-Ofast=ip27 -LNO:fusion=2
109	AMD Athlon K7/1000	39.2	21.2	-O3 -malign-double
110	HP 9000/C240	39.7	18.2	+O2 +Oaggressive
111	Pentium III/866-CM(pgi)	40.7	23.6	-fast -Mdalign
112	Pentium III/1000-CM	40.8	22.2	-O3 -malign-double
113	IBM RS/6000-43P	40.9	18.3	-O3 -qnosave -qalias=noaryovrlp -qarch=p
114	HP 9000/V2250-240	40.9	18.9	+O3 +DA2.0
115	SGI O2000/R10k-250	41.1	20.4	-Ofast=ip27 -LNO:fusion=2
116	AMD Athlon K7/850	41.2	20.7	-O3 -malign-double
117	SGI Onyx2 IR2/R10k-250	41.4	20.7	-Ofast=ip27 -LNO:fusion=2
118	SUN HPC4500/400	41.6	21.6	-fast -O5 -xsafe=mem -fsimple=2
119	Pentium III/866-CM	43.0	23.2	-O3 -malign-double
120	Pentium III/733-CM(pgi)	43.8	24.8	-fast -Mdalign
121	AMD Athlon K7/650 (pgi)	45.5	24.2	-O -Mdalign
122	SGI Origin200/225	45.7	22.6	-Ofast=ip27 -LNO:fusion=2
123	DEC Alpha 8400/5-625	45.8	21.8	-fast
124	AMD Athlon K7/600	46.0	22.2	-O3 -malign-double
125	DEC Alpha PW/600AU	46.6	22.3	-fast
126	Pentium III/733-CM	46.9	25.1	-O3 -malign-double
127	SUN HPC4500/336	47.5	24.7	-fast -X05 -fsimple=2
128	HP 9000/C200	47.8	22.0	+Oall +Oaggressive
129	Pentium III/750-CM(pgi)	47.8	27.3	-fast -Mdalign
130	Pentium III/800-CM(pgi)	48.3	27.0	-fast -Mdalign
131	SGI Octane/R10k-250	48.6	24.5	-Ofast=ip27 -LNO:fusion=2
132	DEC Alpha 1200/5-533	50.0	23.1	-fast
133	AMD Athlon K7/650	50.1	24.1	-O3 -malign-double
134	SGI O2 R12k/270	51.0	26.1	-Ofast=ip32_10k -LNO:fusion=2
135	Pentium III/800-CM	51.6	27.6	-O3 -malign-double
136	Pentium III/650-CM(pgi)	51.7	29.7	-fast -Mdalign
137	Pentium III/750-CM	53.0	28.8	-O3 -malign-double
138	AMD Athlon K7/500 (pgi)	54.2	28.5	-O -Mdalign
139	AMD Athlon K7/500	54.5	25.6	-O3 -malign-double
140	Pentium III/800 (pgi)	54.8	31.3	-O -Mdalign
141	Pentium III/800	55.7	30.0	-O3 -malign-double
142	AMD Athlon K7/600 (pgi)	55.7	28.3	-O -Mdalign
143	SUN Ultra-2/300	56.4	29.8	-fast -O4
144	Pentium III/650-CM	57.2	30.7	-O3 -malign-double
145	Pentium III/550 (pgi)	58.2	34.1	-O -Mdalign
146	SGI O2000/R10k-195	59.3	30.2	-O3 -mips4 etc
147	SGI Octane/R10k-195	60.1	30.8	-O3 -mips4 etc
148	Pentium III/550	60.2	32.7	-O3 -malign-double
149	DEC Alpha 500/5-500	60.8	31.1	-fast
150	SGI PChall-R10k/195	61.4	31.5	-O3 -mips4
151	DEC Alpha PW/433AU	61.5	28.5	-fast
152	SGI Troon IA64/666	61.9	41.5	-O3
153	ProLiant PII/450	62.1	33.3	-O3 -malign-double
154	SUN Ultra30/300	62.7	34.4	-fast -X05 -fsimple=2
155	SGI Origin200/180	65.0	33.3	-O3 -mips4 etc
156	HP 9000/C160	65.2	31.6	+Oall +Oaggressive
157	DEC Alpha 500/5-400	66.8	31.6	-fast
158	SGI Octane/R10k-175	67.6	34.9	-O3 -mips4
159	SGI O2 R10k/175	72.4	37.4	-O3 -mips4
160	Pentium III/500	75.6	43.6	-O -malign-double
161	DEC Alpha 600/5-333	80.9	36.1	-fast
162	DEC Alpha 8400/5-300	82.3	37.5	-fast
163	Pentium II/400 (pgi)	83.6	49.2	-O -Mdalign
164	Pentium II/450	85.4	49.4	-O -malign-double
165	Pentium II/400 (abs)	85.6	48.3	-O -s -f -B108 -B100
166	Pentium II/400	90.9	51.6	-O -malign-double
167	IBM RS/6000-397	91.3	35.1	-O3 -qhot
168	Pentium II/300 (abs)	116.4	63.8	-O
169	Pentium II/266 (pgi)	125.1	73.2	-O -Mdalign
170	Pentium II/266 (abs)	126.0	69.9	-O -s
171	SGI O2 R5k/300	135.6	79.3	-Ofast=ip32_5k -LNO:fusion=2
172	Pentium II/266	159.3	97.4	-O -malign-double
173	Dell Optiplex/266	169.5	104.4	-O -malign-double
174	Pentium II/300	175.5	100.9	-O -malign-double
175	SGI Indigo2 R4400/250	178.8	100.9	-O2 -mips2 -sopt
176	SGI O2 R5k/180	203.6	117.9	-O3 -mips4
177	Pentium Pro/200	218.5	129.5	-G6 -O2
178	HP 9000/755-99	260.2	137.5	+O3
179	Pentium 233 MMX	338.0	209.6	-O -malign-double
	MPP nodes
1	IBM RS/6000-SP/375	23.8	10.8	-O4 -qhot
2	IBM RS/6000-SP/222	39.1	17.8	-O3 -qnosave -qalias=noaryovrlp -qarch=p
3	HP 9000/V2200	45.2	19.4	+O2 +Oaggressive
4	IBM SP2/160Thin	94.1	35.7	-O3 -qhot
5	Cray T3E/1200	107.0	50.4	-O,aggress,scalar3,unroll2,inline3
6	IBM SP2/120Thin	118.1	48.1	-O3 -qhot
7	Cray T3E/900	123.5	58.7	-O,aggress,scalar3,unroll2,inline3
8	IBM SP2/77Wide	171.9	65.9
	Vector Supercomputers
1	NEC SX-6i	43.8	20.0	-Nw -float0 -C hopt -Wf"-Nesc"
2	NEC SX-5	63.1	29.2	-Nw -float0 -C hopt -Wf"-Nesc"
3	FUJITSU VPP-300	310.0	80.6

(+) Prototype, pre-release System




Table 8. The Matrix-97 Benchmark: Performance Relative to the HP RX5670 Itanium2/1500-H.

The Matrix-97 Benchmark: Performance Relative to the HP RX5670 Itanium2/1500-H

Rank	Machine	MMO (FORTRAN)	QHQ transform	Diagonalisation	TOTAL (%)
1	HP RX1620 Itanium2/1600-H	176%	110%	113%	133%
2	IBM eServer p5 570/1.9 GHz	148%	105%	118%	123%
3	HP RX4640 Itanium2/1600-L	93%	117%	94%	102%
4	HP RX1620 Itanium2/1600-L	89%	118%	94%	100%
5	HP RX5670 Itanium2/1500-H	100%	100%	100%	100%
6	HP Superdome Itanium2/1500	95%	100%	100%	99%
7	IBM pSeries 690Turbo/1.7 GHz	102%	73%	108%	94%
8	SGI Altix3700/Itanium2-1500	86%	103%	89%	93%
9	HP RX2600 Itanium2/1500-L	77%	106%	84%	89%
10	HP DL380 Pentium4/3600 (EM64T)	57%	82%	126%	88%
11	HP RX4640 Itanium2/1300-H	86%	89%	87%	87%
12	Intel Tiger Itanium2/1500	67%	103%	83%	85%
13	Pentium4/3400 (EM64T)	47%	78%	118%	81%
14	Pentium4/3200 (EM64T)	45%	74%	111%	77%
15	Pentium4/3400 (EM64T,32-bit)	50%	72%	101%	74%
16	AMD Opteron850/2400 (pgi5.2)	48%	56%	110%	71%
17	IBM Regatta-HPC/1300	77%	56%	79%	71%
18	SGI Altix3700/Itanium2-1300	52%	87%	73%	71%
19	IBM pSeries 690Turbo/1.3 GHz	74%	54%	82%	70%
20	HP DL140 Pentium 4/3200	31%	72%	107%	70%
21	SunFire V40z 2.2GHz (pgi5.2)	52%	53%	102%	69%
22	SunFire V20 2.2GHz (pgi5.2)	47%	53%	101%	67%
23	AMD Opteron848/2200 (pgi5.1)	43%	51%	107%	67%
24	HP DL145 Opteron248/2200 (pgi5.2)	47%	54%	100%	67%
25	SunFire V20 2.2GHz (PathScale)	43%	54%	105%	67%
26	AMD Opteron848/2200 (pgi5.2)	41%	55%	104%	67%
27	Intel Tiger Itanium2/1200 (+)	53%	80%	67%	67%
28	HP RX2600 Itanium2/1000-H	53%	70%	70%	64%
29	AMD Opteron846/2000 (pgi5.1)	40%	47%	98%	62%
30	AMD Opteron848/2200 (pgi5.0)	39%	49%	95%	61%
31	HP 9000/RP3440-1000	35%	62%	82%	59%
32	AMD Opteron846/2000 (pgi5.0)	44%	47%	86%	59%
33	AMD Opteron244/1800 (pgi5.1)	39%	42%	88%	56%
34	Intel Tiger Itanium2/1000	44%	67%	56%	56%
35	HP RX5670 Itanium2/1000-L	45%	59%	57%	54%
36	IBM pSeries 630/1000	57%	41%	62%	53%
37	Pentium 4/3066 (ifc)	29%	49%	77%	52%
38	HP BL20p G2 Pentium 4/2800	30%	50%	73%	51%
39	AMD Opteron244/1800 (pgi5.0)	38%	35%	77%	50%
40	Compaq Alpha ES45/1250	35%	39%	73%	49%
41	HP 9000/RP7410-875	26%	50%	71%	49%
42	Pentium 4/2666 (ifc)	28%	42%	73%	48%
43	AMD Opteron842/1600 (pgi5.0)	33%	36%	68%	46%
44	HP 9000/J6700-750	29%	46%	60%	45%
45	Pentium 4/3066	25%	39%	65%	43%
46	Compaq Marvel EV7/1000	27%	29%	69%	42%
47	Pentium 4/2666 (pgi)	26%	33%	65%	41%
48	HP ZX6000 Itanium2/900	30%	43%	51%	41%
49	AMD Opteron840/1400 (pgi5.0)	29%	32%	60%	40%
50	Compaq Alpha ES45/1000	28%	31%	60%	40%
51	Pentium 4/2666	23%	35%	57%	38%
52	Pentium 4/2533 (pgi)	20%	31%	59%	37%
53	Compaq Alpha ES40/833	26%	27%	52%	35%
54	Pentium 4/2533	19%	30%	55%	35%
55	HP 9000/J6000-552	15%	32%	44%	31%
56	HP Itanium/733-2M L3	19%	40%	32%	31%
57	SGI Onyx 300/R14k-600	28%	21%	37%	29%
58	IBM Itanium/800-4M L3	22%	35%	29%	29%
59	Compaq Alpha DS20E/667	21%	22%	41%	28%
60	SUN Blade 2000/1056-Cu	20%	29%	33%	27%
61	Pentium 4/2000 (ifc)	19%	21%	40%	27%
62	IBM RS/6000 44P-270	29%	19%	30%	26%
63	API UP2000 6/833	14%	24%	40%	26%
64	Compaq Alpha ES40/667	21%	14%	41%	25%
65	SGI O300/R14k-500	25%	16%	32%	24%
66	SUN Fire V880/900-Cu	17%	27%	29%	24%
67	HP 9000/J5000-440	13%	25%	34%	24%
68	HP 9000/N4000-440	11%	24%	35%	23%
69	Pentium 4/2000 (pgi)	17%	17%	35%	23%
70	Compaq PW XP1000/667	19%	16%	34%	23%
71	SGI O3800/R14k-500	20%	17%	31%	23%
72	Pentium 4/1500 (pgi)	17%	17%	34%	23%
73	Pentium 4/2000	15%	17%	35%	22%
74	AMD MP2400+/2000 (pgi)	12%	18%	37%	22%
75	AMD MP2400+/2000	13%	18%	35%	22%
76	SUN Fire 6800/900-Cu	11%	27%	28%	22%
77	Pentium 4/1400 (pgi)	17%	15%	32%	21%
78	Pentium 4/1500	15%	17%	31%	21%
79	AMD MP1800+/1533 (pgi)	13%	17%	32%	21%
80	SUN Blade 1000/M1750	16%	22%	23%	21%
81	AMD MP2000+/1667 (pgi)	12%	16%	33%	20%
82	API UP2000 6/667	11%	18%	32%	20%
83	AMD MP1800+/1533	13%	17%	30%	20%
84	AMD MP2000+/1667	12%	16%	32%	20%
85	SGI O2000/R12k-400	20%	13%	26%	20%
86	Compaq Alpha DS20/500	15%	15%	29%	20%
87	SGI Octane2/R12k-400	17%	14%	26%	19%
88	HP 9000/C3000-400	12%	14%	32%	19%
89	SGI O3800/R12k-400	17%	14%	26%	19%
90	Pentium 4/1400	13%	14%	29%	19%
91	DEC Alpha 8400/6-575	14%	12%	26%	18%
92	Compaq Alpha GS140	14%	10%	27%	17%
93	Compaq Alpha ES40/500	13%	10%	26%	16%
94	Compaq PW XP1000/500	14%	10%	25%	16%
95	Compaq XP1000 6/450	12%	11%	25%	16%
96	AMD Athlon K7/1400(pgi)	7%	14%	26%	16%
97	IBM RS/6000-43P	17%	12%	17%	15%
98	AMD Athlon K7/1400	6%	14%	25%	15%
99	AMD Athlon K7/1200(pgi)	7%	14%	24%	15%
100	AlphaPC 264DP-500	8%	10%	26%	15%
101	SGI O2000/R12k-300	14%	9%	20%	15%
102	AMD Athlon K7/1200	6%	13%	23%	14%
103	HP 9000/785-367	6%	11%	26%	14%
104	HP 9000/V2250-240	9%	13%	17%	13%
105	Compaq Alpha DS10/466	8%	8%	23%	13%
106	SGI Octane/R12k-270	12%	8%	18%	13%
107	SGI Troon IA64/666	8%	19%	11%	13%
108	SUN Ultra80/450	6%	14%	19%	13%
109	HP 9000/C240	11%	8%	18%	12%
110	SGI Onyx2 IR2/R10k-250	11%	7%	17%	12%
111	SGI O2000/R10k-250	10%	7%	17%	12%
112	SUN HPC4500/400	5%	12%	17%	11%
113	AMD Athlon K7/850 (pgi)	6%	9%	18%	11%
114	Pentium III/1000-CM(pgi)	5%	8%	19%	11%
115	AMD Athlon K7/1000(pgi)	5%	9%	18%	11%
116	AMD Athlon K7/1000	4%	9%	18%	11%
117	SGI Origin200/225	9%	7%	16%	11%
118	AMD Athlon K7/850	5%	8%	17%	10%
119	SUN HPC4500/336	5%	11%	15%	10%
120	Pentium III/1000-CM	5%	8%	17%	10%
121	DEC Alpha 8400/5-625	7%	8%	16%	10%
122	DEC Alpha PW/600AU	6%	7%	15%	10%
123	Pentium III/733-CM(pgi)	5%	8%	16%	10%
124	Pentium III/866-CM(pgi)	4%	7%	17%	10%
125	DEC Alpha 1200/5-533	7%	8%	14%	9%
126	AMD Athlon K7/650 (pgi)	5%	8%	16%	9%
127	HP 9000/C200	6%	7%	15%	9%
128	Pentium III/866-CM	4%	7%	17%	9%
129	AMD Athlon K7/600	5%	7%	15%	9%
130	Pentium III/800-CM(pgi)	5%	7%	15%	9%
131	SGI Octane/R10k-250	6%	6%	15%	9%
132	Pentium III/733-CM	5%	7%	15%	9%
133	Pentium III/750-CM(pgi)	4%	7%	15%	9%
134	SGI O2000/R10k-195	7%	6%	12%	9%
135	AMD Athlon K7/650	4%	7%	14%	8%
136	IBM RS/6000-397	9%	9%	8%	8%
137	SUN Ultra-2/300	4%	9%	13%	8%
138	Pentium III/800-CM	4%	7%	14%	8%
139	Pentium III/750-CM	4%	7%	13%	8%
140	Pentium III/650-CM(pgi)	4%	6%	14%	8%
141	AMD Athlon K7/500 (pgi)	4%	6%	13%	8%
142	AMD Athlon K7/500	4%	6%	13%	8%
143	AMD Athlon K7/600 (pgi)	5%	6%	13%	8%
144	SGI Octane/R10k-195	5%	6%	12%	8%
145	Pentium III/800	4%	6%	13%	8%
146	Pentium III/800 (pgi)	4%	6%	13%	8%
147	DEC Alpha 500/5-500	6%	5%	12%	7%
148	Pentium III/550 (pgi)	4%	6%	12%	7%
149	Pentium III/650-CM	3%	6%	12%	7%
150	SGI O2 R12k/270	2%	6%	14%	7%
151	Pentium III/550	4%	6%	12%	7%
152	DEC Alpha PW/433AU	4%	6%	12%	7%
153	SGI Origin200/180	5%	6%	11%	7%
154	SUN Ultra30/300	3%	6%	11%	7%
155	ProLiant PII/450	3%	6%	11%	7%
156	HP 9000/C160	5%	5%	11%	7%
157	SGI Octane/R10k-175	5%	5%	11%	7%
158	SGI PChall-R10k/195	3%	5%	12%	7%
159	DEC Alpha 500/5-400	3%	5%	11%	6%
160	Pentium III/500	4%	5%	9%	6%
161	DEC Alpha 600/5-333	4%	5%	9%	6%
162	DEC Alpha 8400/5-300	4%	5%	9%	6%
163	Pentium II/450	3%	5%	8%	5%
164	Pentium II/400 (pgi)	3%	4%	9%	5%
165	SGI O2 R10k/175	1%	4%	10%	5%
166	Pentium II/400	3%	4%	8%	5%
167	Pentium II/400 (abs)	3%	4%	8%	5%
168	Pentium II/300 (abs)	2%	3%	6%	4%
169	Pentium II/266 (abs)	1%	3%	6%	3%
170	Pentium II/266 (pgi)	1%	3%	6%	3%
171	Pentium II/300	2%	3%	4%	3%
172	Pentium II/266	2%	3%	4%	3%
173	SGI O2 R5k/300	1%	2%	5%	3%
174	Dell Optiplex/266	2%	1%	4%	2%
175	SGI O2 R5k/180	1%	1%	3%	2%
176	SGI Indigo2 R4400/250	2%	0%	4%	2%
177	Pentium Pro/200	1%	1%	3%	2%
178	HP 9000/755-99	1%	1%	3%	2%
179	Pentium 233 MMX	1%	1%	2%	1%
	MPP nodes
1	IBM RS/6000-SP/375	31%	21%	30%	28%
2	IBM RS/6000-SP/222	15%	11%	18%	15%
3	HP 9000/V2200	11%	13%	16%	13%
4	IBM SP2/160Thin	9%	9%	8%	8%
5	Cray T3E/1200	7%	6%	7%	7%
6	IBM SP2/120Thin	7%	7%	6%	6%
7	Cray T3E/900	6%	6%	6%	6%
8	IBM SP2/77Wide	3%	3%	4%	4%
	Vector Supercomputers
1	NEC SX-6i	204%	145%	16%	122%
2	NEC SX-5	167%	124%	11%	101%
3	FUJITSU VPP-300	27%	13%	2%	14%

(+) Prototype, pre-release System




Table 9. The Matrix-89 Benchmark: Performance Relative to the HP RX5670 Itanium2/1500-H.

The Matrix-89 Benchmark: Performance Relative to the HP RX5670 Itanium2/1500-H.

Rank	Machine	MMO (FORTRAN)	QHQ transform	Diagonalisation	TOTAL (%)
1	HP RX1620 Itanium2/1600-H	200%	106%	114%	140%
2	IBM eServer p5 570/1.9 GHz	200%	92%	123%	138%
3	HP DL140 Pentium 4/3200	133%	87%	124%	115%
4	AMD Opteron848/2200 (pgi5.2)	114%	61%	138%	105%
5	HP DL380 Pentium4/3600 (EM64T)	80%	81%	144%	102%
6	HP RX5670 Itanium2/1500-H	100%	100%	100%	100%
7	HP Superdome Itanium2/1500	89%	103%	101%	98%
8	AMD Opteron850/2400 (pgi5.2)	80%	62%	148%	97%
9	HP RX4640 Itanium2/1600-L	86%	112%	88%	95%
10	HP RX1620 Itanium2/1600-L	86%	111%	88%	95%
11	HP DL145 Opteron248/2200 (pgi5.2)	89%	60%	134%	94%
12	Pentium4/3400 (EM64T)	62%	80%	135%	92%
13	SunFire V20 2.2GHz (pgi5.2)	80%	59%	137%	92%
14	IBM pSeries 690Turbo/1.7 GHz	100%	66%	107%	91%
15	SunFire V40z 2.2GHz (pgi5.2)	73%	62%	137%	91%
16	HP RX2600 Itanium2/1500-L	89%	94%	85%	89%
17	SGI Altix3700/Itanium2-1500	81%	105%	83%	89%
18	AMD Opteron848/2200 (pgi5.1)	73%	53%	140%	89%
19	Pentium4/3400 (EM64T,32-bit)	80%	70%	114%	88%
20	Pentium4/3200 (EM64T)	62%	74%	128%	88%
21	Pentium 4/3066 (ifc)	73%	70%	114%	85%
22	HP BL20p G2 Pentium 4/2800	80%	67%	109%	85%
23	SunFire V20 2.2GHz (PathScale)	75%	53%	123%	84%
24	HP RX4640 Itanium2/1300-H	80%	83%	86%	83%
25	AMD Opteron846/2000 (pgi5.1)	73%	49%	124%	82%
26	AMD Opteron848/2200 (pgi5.0)	73%	49%	123%	81%
27	Intel Tiger Itanium2/1500	65%	94%	80%	79%
28	AMD Opteron244/1800 (pgi5.1)	73%	43%	113%	76%
29	Pentium 4/2666 (ifc)	80%	49%	99%	76%
30	AMD Opteron846/2000 (pgi5.0)	67%	46%	113%	75%
31	SGI Altix3700/Itanium2-1300	63%	85%	74%	74%
32	IBM pSeries 690Turbo/1.3 GHz	89%	50%	82%	74%
33	IBM Regatta-HPC/1300	80%	51%	82%	71%
34	AMD Opteron244/1800 (pgi5.0)	67%	33%	98%	66%
35	HP 9000/RP3440-1000	53%	62%	81%	65%
36	Intel Tiger Itanium2/1200 (+)	51%	74%	65%	63%
37	Pentium 4/3066	51%	45%	86%	61%
38	HP RX2600 Itanium2/1000-H	53%	61%	67%	61%
39	Compaq Alpha ES45/1250	44%	39%	97%	60%
40	Pentium 4/2666 (pgi)	53%	44%	81%	60%
41	AMD Opteron842/1600 (pgi5.0)	50%	35%	89%	58%
42	HP 9000/RP7410-875	50%	55%	68%	58%
43	Pentium 4/2533 (pgi)	53%	39%	79%	57%
44	AMD MP2400+/2000 (pgi)	35%	36%	94%	55%
45	IBM pSeries 630/1000	62%	39%	62%	54%
46	Intel Tiger Itanium2/1000	43%	62%	54%	53%
47	AMD MP2400+/2000	36%	34%	86%	52%
48	AMD Opteron840/1400 (pgi5.0)	47%	31%	78%	52%
49	HP RX5670 Itanium2/1000-L	43%	53%	57%	51%
50	Pentium 4/2666	42%	38%	72%	51%
51	Compaq Marvel EV7/1000	37%	31%	82%	50%
52	Pentium 4/2000 (ifc)	50%	28%	72%	50%
53	Compaq Alpha ES45/1000	35%	32%	79%	49%
54	Pentium 4/2533	38%	32%	75%	49%
55	HP 9000/J6700-750	40%	47%	59%	48%
56	HP ZX6000 Itanium2/900	38%	47%	51%	45%
57	AMD MP2000+/1667 (pgi)	28%	30%	78%	45%
58	AMD MP2000+/1667	32%	28%	70%	43%
59	AMD MP1800+/1533 (pgi)	27%	29%	71%	43%
60	Pentium 4/2000 (pgi)	42%	22%	63%	42%
61	Compaq Alpha ES40/833	31%	27%	66%	41%
62	AMD MP1800+/1533	30%	27%	65%	40%
63	API UP2000 6/833	31%	24%	64%	40%
64	Pentium 4/2000	36%	22%	60%	39%
65	AMD Athlon K7/1400(pgi)	24%	24%	65%	38%
66	HP 9000/J6000-552	31%	35%	43%	36%
67	AMD Athlon K7/1400	24%	24%	59%	36%
68	Pentium 4/1500 (pgi)	33%	23%	47%	34%
69	HP Itanium/733-2M L3	32%	36%	34%	34%
70	AMD Athlon K7/1200(pgi)	21%	21%	59%	34%
71	Compaq Alpha DS20E/667	25%	22%	52%	33%
72	SGI Onyx 300/R14k-600	28%	20%	49%	32%
73	Pentium 4/1400 (pgi)	29%	21%	45%	32%
74	AMD Athlon K7/1200	22%	21%	52%	31%
75	API UP2000 6/667	24%	19%	51%	31%
76	Compaq PW XP1000/667	24%	17%	52%	31%
77	Compaq Alpha ES40/667	24%	15%	52%	30%
78	IBM RS/6000 44P-270	33%	19%	38%	30%
79	Pentium 4/1500	23%	21%	43%	29%
80	IBM Itanium/800-4M L3	29%	24%	33%	29%
81	SUN Blade 2000/1056-Cu	19%	24%	43%	29%
82	Pentium 4/1400	21%	21%	43%	28%
83	HP 9000/N4000-440	25%	25%	34%	28%
84	HP 9000/J5000-440	25%	25%	34%	28%
85	AMD Athlon K7/1000(pgi)	17%	17%	49%	28%
86	SGI O300/R14k-500	23%	16%	42%	27%
87	AMD Athlon K7/1000	15%	17%	46%	26%
88	SGI O3800/R14k-500	20%	16%	41%	26%
89	Pentium III/1000-CM(pgi)	19%	14%	44%	25%
90	SUN Fire 6800/900-Cu	17%	20%	37%	25%
91	SUN Fire V880/900-Cu	16%	21%	36%	25%
92	Compaq Alpha DS20/500	19%	16%	39%	24%
93	DEC Alpha 8400/6-575	18%	13%	40%	24%
94	Compaq Alpha GS140	17%	11%	41%	23%
95	AlphaPC 264DP-500	18%	12%	38%	23%
96	HP 9000/C3000-400	23%	14%	31%	23%
97	Compaq PW XP1000/500	18%	10%	39%	22%
98	SGI Octane2/R12k-400	19%	13%	33%	22%
99	Pentium III/1000-CM	14%	14%	38%	22%
100	Pentium III/866-CM(pgi)	16%	12%	37%	22%
101	SGI O2000/R12k-400	20%	12%	33%	22%
102	AMD Athlon K7/850 (pgi)	13%	13%	38%	22%
103	SGI O3800/R12k-400	17%	14%	33%	21%
104	SUN Blade 1000/M1750	15%	17%	31%	21%
105	AMD Athlon K7/850	13%	13%	37%	21%
106	Compaq Alpha ES40/500	16%	11%	36%	21%
107	Compaq Alpha DS10/466	16%	10%	36%	21%
108	HP 9000/785-367	20%	13%	28%	20%
109	Pentium III/800-CM(pgi)	15%	11%	33%	20%
110	Compaq XP1000 6/450	14%	11%	32%	19%
111	Pentium III/800 (pgi)	14%	10%	31%	19%
112	Pentium III/866-CM	12%	11%	32%	19%
113	Pentium III/750-CM(pgi)	14%	11%	31%	19%
114	Pentium III/733-CM(pgi)	13%	11%	31%	18%
115	Pentium III/800-CM	11%	10%	30%	17%
116	AMD Athlon K7/650 (pgi)	11%	11%	29%	17%
117	AMD Athlon K7/650	10%	11%	29%	17%
118	Pentium III/800	11%	9%	30%	17%
119	Pentium III/650-CM(pgi)	12%	9%	27%	16%
120	Pentium III/750-CM	11%	10%	28%	16%
121	SGI O2000/R12k-300	14%	9%	25%	16%
122	Pentium III/733-CM	10%	10%	27%	16%
123	AMD Athlon K7/600	11%	9%	26%	15%
124	HP 9000/V2250-240	13%	13%	19%	15%
125	SGI Troon IA64/666	14%	17%	13%	15%
126	IBM RS/6000-43P	11%	11%	21%	14%
127	SGI Octane/R12k-270	13%	8%	23%	14%
128	AMD Athlon K7/600 (pgi)	9%	8%	26%	14%
129	Pentium III/650-CM	9%	9%	24%	14%
130	SGI O2 R12k/270	13%	7%	21%	14%
131	HP 9000/C240	12%	8%	18%	13%
132	DEC Alpha 8400/5-625	8%	9%	22%	13%
133	SUN Ultra80/450	6%	12%	21%	13%
134	AMD Athlon K7/500	9%	8%	22%	13%
135	DEC Alpha PW/600AU	8%	9%	21%	13%
136	AMD Athlon K7/500 (pgi)	8%	8%	22%	13%
137	SGI Onyx2 IR2/R10k-250	11%	7%	20%	13%
138	SGI O2000/R10k-250	10%	7%	20%	13%
139	DEC Alpha 1200/5-533	8%	9%	19%	12%
140	SGI Octane/R10k-250	10%	7%	19%	12%
141	SUN HPC4500/400	5%	11%	19%	12%
142	Pentium III/550 (pgi)	9%	7%	19%	12%
143	HP 9000/C200	11%	7%	16%	11%
144	SGI Origin200/225	10%	6%	18%	11%
145	IBM RS/6000-397	11%	10%	13%	11%
146	Pentium III/550	7%	7%	19%	11%
147	SUN HPC4500/336	5%	10%	17%	11%
148	DEC Alpha 500/5-500	7%	6%	16%	10%
149	SGI O2000/R10k-195	7%	6%	15%	9%
150	SGI Octane/R10k-195	7%	6%	15%	9%
151	DEC Alpha PW/433AU	6%	7%	15%	9%
152	SUN Ultra-2/300	5%	9%	15%	9%
153	IBM RS/6000-595	9%	8%	11%	9%
154	ProLiant PII/450	6%	6%	15%	9%
155	SUN Ultra30/300	4%	8%	14%	9%
156	SGI PChall-R10k/195	7%	6%	14%	9%
157	SGI Origin200/180	7%	6%	14%	9%
158	DEC Alpha 500/5-400	5%	6%	14%	9%
159	Pentium III/500	5%	6%	15%	9%
160	HP 9000/C160	8%	4%	13%	8%
161	SGI Octane/R10k-175	6%	5%	14%	8%
162	SGI O2 R10k/175	6%	5%	14%	8%
163	SGI Indigo2-R10k/175	7%	5%	12%	8%
164	Pentium II/400 (pgi)	6%	5%	13%	8%
165	Pentium II/450	5%	5%	13%	8%
166	Pentium II/400	5%	5%	12%	7%
167	DEC Alpha 600/5-333	4%	6%	11%	7%
168	Pentium II/400 (abs)	3%	5%	13%	7%
169	DEC Alpha 8400/5-300	4%	5%	11%	7%
170	SUN Ultra-2/200	3%	5%	10%	6%
171	SGI O2 R5k/300	6%	2%	10%	6%
172	DEC Alpha 2100/5-250	3%	4%	9%	5%
173	Pentium II/300 (abs)	3%	3%	10%	5%
174	SUN Ultra-1/170	3%	4%	8%	5%
175	DEC Alpha 600/5-266	3%	3%	8%	5%
176	Pentium II/266 (abs)	2%	3%	9%	5%
177	Pentium II/266 (pgi)	3%	3%	8%	5%
178	SGI PChall-R8k/75	4%	4%	6%	5%
179	SGI R8k Indigo2	4%	4%	6%	5%
180	Pentium II/300	4%	3%	7%	5%
181	Pentium Pro/200	3%	2%	8%	5%
182	IBM RS/6000-3CT	4%	4%	5%	4%
183	SUN Ultra-1/140	2%	4%	7%	4%
184	Pentium II/266	3%	3%	6%	4%
185	IBM RS/6000-590	4%	3%	5%	4%
186	DEC Alpha 250/4-266	3%	3%	6%	4%
187	SGI O2 R5k/180	3%	1%	6%	4%
188	HP 9000/735-125	3%	2%	5%	4%
189	IBM RS/6000-3BT	4%	2%	5%	4%
190	HP 9000/J200-100	4%	2%	5%	4%
191	Dell Optiplex/266	3%	2%	6%	3%
192	DEC AXP/3000-700	2%	2%	5%	3%
193	SGI Indy-R5k	2%	1%	6%	3%
194	HP 9000/735-99	3%	2%	4%	3%
195	HP 9000/755-99	3%	2%	4%	3%
196	SGI Indigo2 R4400/250	2%	1%	5%	3%
197	HP 9000/715-100	3%	1%	3%	2%
198	DEC AXP/3000-600	2%	2%	4%	2%
199	SUN SPARC 20/HS21	2%	1%	4%	2%
200	IBM PowerPC-43P	2%	1%	4%	2%
201	IBM RS/6000-370	2%	2%	3%	2%
202	HP 9000/715-80	2%	1%	3%	2%
203	DEC AXP/3000-500	1%	1%	3%	2%
204	DEC AXP/3000-300	1%	1%	3%	2%
205	Pentium 233 MMX	1%	1%	3%	2%
206	IBM RS/6000-360	1%	1%	3%	2%
207	SGI Indigo2 R4400/150	1%	1%	3%	2%
208	SGI Challenge L/150	1%	1%	3%	2%
209	IBM RS/6000-550	1%	1%	2%	2%
210	HP 9000/750-66	2%	1%	2%	2%
211	IBM RS/6000-350	1%	1%	2%	1%
212	SGI Challenge L/100	1%	1%	2%	1%
213	IBM RS/6000-530H	1%	1%	2%	1%
214	IBM RS/6000-340	1%	1%	2%	1%
215	IBM PowerPC-25T	1%	1%	2%	1%
216	IBM PowerPC-250	1%	1%	2%	1%
217	SUN SPARCserver 1000	1%	0%	2%	1%
218	HP 9000/720-50	1%	1%	1%	1%
219	SGI R4000 Indigo	1%	0%	2%	1%
220	SUN SPARC/5-85	1%	0%	1%	1%
221	SUN SPARC 10/41	1%	0%	1%	1%
222	IBM RS/6000-320	1%	0%	1%	1%
223	SUN SPARC 10/30	0%	0%	1%	1%
224	Stardent VISTRA-800	0%	0%	1%	0%
225	SGI 4D/420	0%	0%	1%	0%
226	SGI 4D/320	0%	0%	1%	0%
227	SGI R3000 indigo	0%	0%	1%	0%
228	DEC S5000/200	0%	0%	0%	0%
229	SUN SPARC 2/GS	0%	0%	0%	0%
230	HP/Apollo DN10020	0%	0%	0%	0%
231	SGI 4D/220	0%	0%	0%	0%
232	Stardent 1520	0%	0%	0%	0%
233	DEC S5000/120	0%	0%	0%	0%
234	SOLBOURNE S4000	0%	0%	0%	0%
235	SUN 4/370	0%	0%	0%	0%
	MPP nodes
1	IBM RS/6000-SP/375	30%	20%	38%	29%
2	IBM RS/6000-SP/222	17%	11%	22%	17%
3	HP 9000/V2200	11%	12%	15%	13%
4	Cray T3E/1200	14%	8%	16%	13%
5	IBM SP2/160Thin	10%	10%	12%	11%
6	Cray T3E/900	11%	7%	12%	10%
7	IBM SP2/120Thin	8%	7%	9%	8%
8	IBM SP2/66Thin	4%	4%	5%	4%
9	Hitachi SR2201	3%	3%	5%	4%
10	Cray T3D/AXP-150	1%	1%	2%	1%
11	KSR-2	1%	1%	2%	1%
	Vector Supercomputers
1	NEC SX-6i	73%	113%	30%	72%
2	NEC SX-5	43%	68%	17%	43%
3	NEC SX-4	21%	33%	8%	21%
4	Cray YMP-C98/4256	9%	13%	4%	9%
5	FUJITSU VPP-300	7%	7%	3%	6%
6	CRAY YMP/J90	3%	3%	2%	3%

(+) Prototype, pre-release System




Table 10. The Computational Chemistry Kernels.

The Computational Chemistry Kernels Benchmark.

Rank	Machine	Total CPU Times (seconds)			Mflop	Relative
		SCF	MD	MC	JACOBI	Performance (%)
1	HP RX1620 Itanium2/1600-H	2.5	11.5	2.5	1714.3	112
2	HP RX1620 Itanium2/1600-L	4.4	7.9	3.0	1847.8	109
3	HP RX4640 Itanium2/1600-L	4.5	7.9	3.0	1627.6	105
4	SGI Altix3700/Itanium2-1500	4.7	8.5	3.2	1718.6	102
5	IBM eServer p5 570/1.9 GHz	4.6	12.3	2.6	2000.0	100
6	HP RX5670 Itanium2/1500-H	2.8	12.9	2.7	1500.0	100
7	HP DL380 Pentium4/3600 (EM64T)	6.4	14.3	1.5	1043.5	94
8	AMD Opteron850/2400 (pgi5.2)	5.4	13.3	1.5	648.7	93
9	HP Superdome Itanium2/1500	2.8	12.9	2.7	1043.5	92
10	SunFire V20 2.2GHz (PathScale)	3.9	15.1	1.5	489.8	92
11	HP RX2600 Itanium2/1500-L	6.4	9.7	3.3	1480.5	89
12	Pentium4/3400 (EM64T)	6.9	15.3	1.6	888.9	87
13	HP DL145 Opteron248/2200 (pgi5.2)	6.2	14.5	1.6	727.3	86
14	SunFire V20 2.2GHz (pgi5.2)	5.9	14.7	1.6	685.7	86
15	SunFire V40z 2.2GHz (pgi5.2)	6.0	14.5	1.6	615.4	86
16	AMD Opteron848/2200 (pgi5.2)	6.0	14.4	1.6	558.1	85
17	Pentium4/3200 (EM64T)	7.3	16.2	1.7	1000.0	84
18	Intel Tiger Itanium2/1500	6.9	10.4	3.4	1388.5	84
19	AMD Opteron848/2200 (pgi5.1)	6.1	14.5	1.6	470.6	83
20	HP RX4640 Itanium2/1300-H	5.4	14.9	3.1	1500.0	81
21	SGI Altix3700/Itanium2-1300	7.4	11.2	3.9	1526.5	81
22	AMD Opteron846/2000 (pgi5.1)	6.6	16.0	1.8	600.0	78
23	AMD Opteron848/2200 (pgi5.0)	7.0	14.6	1.9	571.4	77
24	Pentium4/3400 (EM64T,32-bit)	7.0	16.4	2.1	857.1	76
25	HP DL140 Pentium 4/3200	6.3	14.6	1.9	480.0	76
26	Pentium 4/3066 (ifc)	6.2	16.8	2.2	827.6	75
27	AMD Opteron846/2000 (pgi5.0)	7.7	16.0	2.1	615.4	72
28	AMD Opteron244/1800 (pgi5.1)	7.0	17.7	1.9	545.5	72
29	IBM pSeries 690Turbo/1.7 GHz	5.9	16.3	3.4	1142.9	70
30	Intel Tiger Itanium2/1200 (+)	9.9	13.1	4.4	1357.8	70
31	HP BL20p G2 Pentium 4/2800	8.3	16.6	2.3	666.7	68
32	HP RX2600 Itanium2/1000-H	5.5	19.7	4.0	1333.3	68
33	Pentium 4/2666 (ifc)	7.0	19.4	2.4	615.4	65
34	AMD Opteron244/1800 (pgi5.0)	8.6	17.9	2.3	558.1	65
35	HP RX5670 Itanium2/1000-L	10.7	15.6	5.2	1286.7	61
36	Intel Tiger Itanium2/1000	12.0	15.7	5.2	1307.2	61
37	HP 9000/RP3440-1000	7.2	21.7	4.1	960.0	57
38	AMD Opteron842/1600 (pgi5.0)	9.6	20.1	2.6	452.8	57
39	Compaq Alpha ES45/1250	5.8	18.8	3.4	416.1	56
40	Pentium 4/2666 (pgi)	10.8	21.6	2.6	489.8	55
41	IBM Regatta-HPC/1300	7.2	22.4	4.3	923.1	55
42	IBM pSeries 690Turbo/1.3 GHz	7.8	21.4	4.4	923.1	55
43	HP ZX6000 Itanium2/900	12.3	17.4	5.8	948.9	52
44	AMD MP2400+/2000 (pgi)	7.4	23.2	2.7	193.6	51
45	Compaq Marvel EV7/1000	7.2	25.1	4.2	677.5	50
46	AMD Opteron840/1400 (pgi5.0)	11.0	23.0	3.0	393.4	49
47	Pentium 4/2000 (ifc)	11.0	26.0	2.9	444.4	49
48	Pentium 4/3066	11.2	18.9	4.3	469.0	47
49	Compaq Alpha ES45/1000	7.2	22.9	4.6	355.4	44
50	AMD MP2400+/2000	9.6	23.5	3.5	187.5	43
51	AMD MP2000+/1667 (pgi)	8.8	27.9	3.3	189.0	43
52	Pentium 4/2533 (pgi)	11.4	29.1	4.0	436.4	41
53	AMD MP1800+/1533 (pgi)	9.6	30.5	3.5	201.7	40
54	IBM pSeries 630/1000	10.2	28.1	5.8	571.4	39
55	Pentium 4/2666	13.9	22.9	5.1	406.8	39
56	HP 9000/RP7410-875	8.4	26.1	4.9	214.3	38
57	Pentium 4/2533	14.4	23.9	5.4	406.8	38
58	Compaq Alpha ES40/833	8.4	27.8	5.4	231.3	36
59	AMD Athlon K7/1400(pgi)	10.6	33.6	3.9	160.0	36
60	API UP2000 6/833	9.0	28.4	5.6	254.4	35
61	Pentium 4/2000 (pgi)	14.6	38.5	5.2	490.0	34
62	AMD MP2000+/1667	12.7	31.6	4.4	187.5	34
63	HP 9000/J6700-750	10.1	30.8	5.8	300.0	34
64	SUN Blade 2000/1056-Cu	10.8	39.8	5.2	360.8	33
65	Pentium 4/2000	18.4	31.3	6.8	480.0	32
66	AMD Athlon K7/1200(pgi)	12.3	38.2	4.5	141.2	31
67	SGI O3800/R14k-500	8.5	40.4	5.8	192.3	31
68	SGI O300/R14k-500	8.5	40.2	5.8	180.1	31
69	SGI Onyx 300/R14k-600	10.4	33.9	6.0	186.5	30
70	AMD MP1800+/1533	14.4	34.4	5.2	192.0	30
71	Compaq Alpha DS20E/667	10.5	34.7	6.9	264.7	30
72	SUN Fire V880/900-Cu	12.8	49.1	6.1	351.1	29
73	SUN Fire 6800/900-Cu	14.4	44.4	6.0	320.0	29
74	Compaq Alpha ES40/667	11.7	35.2	7.5	269.0	28
75	API UP2000 6/667	10.6	37.9	6.9	226.5	28
76	Pentium 4/1500 (pgi)	19.5	50.9	7.0	500.0	28
77	AMD Athlon K7/1400	16.0	38.4	5.7	161.1	27
78	HP Itanium/733-2M L3	17.9	36.4	9.1	421.1	27
79	Compaq PW XP1000/667	11.7	36.0	7.1	160.8	27
80	HP 9000/J6000-552	13.7	40.9	7.9	279.1	26
81	IBM Itanium/800-4M L3	30.1	33.1	8.7	356.2	26
82	SGI O3800/R12k-400	10.6	47.5	7.2	179.8	26
83	AMD Athlon K7/1000(pgi)	14.8	46.2	5.5	102.1	25
84	SGI Troon IA64/666	27.8	21.2	12.1	118.8	25
85	Pentium 4/1500	28.2	46.7	9.7	500.0	25
86	SGI Octane2/R12k-400	10.5	49.7	7.3	127.3	24
87	Pentium 4/1400 (pgi)	20.9	55.3	7.8	347.8	24
88	IBM RS/6000 44P-270	14.7	48.6	8.5	258.1	24
89	AMD Athlon K7/1200	18.8	45.0	6.9	149.1	23
90	SGI O2000/R12k-400	11.7	49.6	8.0	110.5	23
91	Compaq Alpha DS20/500	14.1	48.6	8.8	202.2	23
92	Pentium 4/1400	26.6	44.2	9.8	342.9	22
93	Compaq Alpha ES40/500	15.4	46.5	9.6	204.1	22
94	Compaq Alpha GS140	14.9	47.3	9.2	170.8	22
95	AMD Athlon K7/850 (pgi)	17.8	58.1	6.8	129.0	21
96	AlphaPC 264DP-500	15.2	48.1	9.5	189.6	21
97	SUN Blade 1000/M1750	14.3	53.5	12.4	298.3	21
98	HP 9000/J5000-440	18.2	52.1	9.9	255.3	21
99	Compaq PW XP1000/500	15.7	50.8	9.3	176.9	21
100	AMD Athlon K7/1000	21.0	56.5	7.5	101.3	20
101	Compaq XP1000 6/450	17.0	50.6	10.6	142.9	19
102	Compaq Alpha DS10/466	17.2	52.4	10.4	139.6	19
103	HP 9000/N4000-440	23.8	52.9	11.2	228.6	19
104	HP 9000/C3000-400	20.0	59.6	10.9	228.6	19
105	Pentium III/1000-CM(pgi)	23.3	68.2	7.7	136.4	19
106	SGI O2000/R12k-300	14.1	63.8	9.9	95.8	18
107	DEC Alpha 8400/6-575	19.7	47.6	10.8	81.9	18
108	AMD Athlon K7/850	24.6	72.7	8.6	129.0	17
109	AMD Athlon K7/650 (pgi)	23.4	78.3	8.8	124.4	17
110	SGI Octane/R12k-270	15.8	70.9	11.0	97.7	17
111	Pentium III/866-CM(pgi)	26.8	74.0	8.8	108.1	16
112	HP 9000/785-367	23.1	65.4	11.9	118.2	16
113	Pentium III/800-CM(pgi)	28.9	81.2	9.5	103.0	15
114	AMD Athlon K7/600 (pgi)	28.8	85.9	10.5	121.0	15
115	SGI O2 R12k/270	17.4	75.1	11.4	17.8	14
116	SGI O2000/R10k-250	18.3	83.4	12.7	85.7	14
117	Pentium III/750-CM(pgi)	31.1	85.5	10.2	110.6	14
118	Pentium III/733-CM(pgi)	31.7	87.1	10.4	123.1	14
119	Pentium III/866-CM	30.3	84.1	10.4	107.1	14
120	Pentium III/800 (pgi)	31.6	83.8	10.7	112.2	14
121	SUN Ultra80/450	33.0	54.9	17.9	131.6	14
122	IBM RS/6000-43P	27.0	91.4	15.4	205.1	14
123	Pentium III/1000-CM	31.8	94.5	10.8	120.0	14
124	DEC Alpha PW/600AU	24.6	75.2	12.3	66.4	14
125	SGI Onyx2 IR2/R10k-250	22.1	87.0	12.6	88.9	14
126	DEC Alpha 8400/5-625	23.9	75.5	12.7	63.6	13
127	SGI Octane/R10k-250	22.4	91.0	12.7	87.3	13
128	Pentium III/800	32.9	91.6	11.3	96.4	13
129	Pentium III/800-CM	33.3	91.6	11.4	98.4	13
130	AMD Athlon K7/600	35.6	89.2	12.3	121.8	13
131	SGI Origin200/225	20.3	92.5	14.0	78.5	13
132	DEC Alpha 1200/5-533	26.9	81.0	13.2	73.7	13
133	Pentium III/733-CM	36.1	100.3	12.4	124.4	13
134	Pentium III/750-CM	35.2	98.4	12.1	108.6	13
135	Pentium III/650-CM(pgi)	36.0	100.7	11.8	103.9	13
136	AMD Athlon K7/500 (pgi)	34.1	98.3	12.5	112.2	13
137	AMD Athlon K7/650	35.6	101.6	12.4	108.6	12
138	SUN HPC4500/400	36.8	61.8	20.1	106.6	12
139	IBM RS/6000-397	59.5	107.6	16.8	237.6	12
140	AMD Athlon K7/500	41.0	100.4	14.8	111.6	11
141	Pentium III/650-CM	40.6	113.3	14.0	106.7	11
142	HP 9000/C240	33.4	101.2	17.7	121.2	11
143	DEC Alpha 500/5-500	32.0	94.0	14.9	50.9	11
144	Pentium III/550 (pgi)	46.5	120.3	15.7	100.9	10
145	DEC Alpha PW/433AU	33.7	102.6	17.0	58.8	10
146	SUN HPC4500/336	47.3	73.2	22.2	71.7	10
147	HP 9000/C200	40.5	110.6	20.0	110.1	10
148	HP 9000/V2250-240	43.3	95.1	20.5	81.1	10
149	IBM RS/6000-595	72.4	137.1	20.0	167.8	9
150	SGI Octane/R10k-195	42.6	112.0	20.1	82.7	9
151	SGI O2000/R10k-195	41.5	112.1	20.4	84.1	9
152	DEC Alpha 500/5-400	38.7	112.8	18.1	50.4	9
153	Pentium III/550	48.7	187.0	16.7	100.8	9
154	SGI Origin200/180	46.1	121.3	21.8	75.0	8
155	SGI PChall-R10k/195	42.7	106.0	22.4	39.1	8
156	Pentium II/400 (abs)	56.4	150.5	19.6	82.5	8
157	SGI Octane/R10k-175	46.6	126.9	23.5	74.3	8
158	SUN Ultra-2/300	55.1	85.2	29.5	41.8	8
159	DEC Alpha 600/5-333	45.8	136.6	20.9	44.9	8
160	ProLiant PII/450	59.0	166.4	20.2	78.4	8
161	SUN Ultra30/300	58.5	96.1	29.6	54.3	8
162	Pentium II/400 (pgi)	64.7	167.0	21.5	82.5	7
163	Pentium III/500	62.2	171.9	22.1	85.4	7
164	SGI O2 R5k/300	68.4	136.7	18.8	16.2	7
165	SGI O2 R10k/175	48.1	126.4	23.7	12.6	7
166	HP 9000/C160	49.1	142.8	25.4	39.6	7
167	Pentium II/450	64.8	205.9	23.6	83.3	7
168	DEC Alpha 8400/5-300	52.6	149.3	25.8	46.6	7
169	SGI Indigo2-R10k/175	62.1	134.7	24.3	30.3	7
170	Pentium II/400	74.9	230.0	27.5	82.5	6
171	SUN Ultra-2/200	81.8	120.9	43.0	56.5	6
172	DEC Alpha 2100/5-250	59.4	186.9	28.0	27.3	6
173	Pentium II/300 (abs)	74.5	216.8	27.2	47.6	6
174	DEC Alpha 600/5-266	57.3	209.8	30.6	41.0	6
175	Pentium II/266 (abs)	85.0	227.1	29.6	48.7	5
176	SUN Ultra-1/170	98.3	145.8	58.4	47.3	5
177	Pentium II/266 (pgi)	96.0	262.1	33.0	49.1	5
178	Pentium II/300	99.6	291.4	37.0	44.9	4
179	SGI O2 R5k/180	115.9	212.6	33.6	14.2	4
180	Pentium II/266	106.3	339.1	35.5	47.5	4
181	DEC Alpha 250/4-266	78.8	272.0	37.0	20.9	4
182	SUN Ultra-1/140	110.1	186.0	60.7	45.2	4
183	IBM RS/6000-590	146.0	453.0	43.0	86.0	4
184	IBM RS/6000-3CT	138.3	461.6	42.2	81.1	4
185	SGI PChall-R8k/75	138.5	217.8	47.6	32.9	4
186	SGI Indy-R5k	123.2	219.4	42.2	12.5	4
187	SGI R8k Indigo2	138.9	218.3	47.9	23.7	4
188	Pentium Pro/200	142.0	321.0	38.0	28.2	4
189	IBM RS/6000-3BT	140.1	389.4	43.1	44.8	4
190	SGI Indigo2 R4400/250	136.3	319.6	46.6	16.9	3
191	Pentium Pro/200-f77m	141.9	424.2	45.6	20.4	3
192	HP 9000/735-125	103.0	236.0	91.0	15.0	3
193	DEC AXP/3000-700	146.3	410.8	54.0	21.7	3
194	Convex C-3860	453.0	342.0	141.0	69.9	3
195	IBM PowerPC-43P	134.8	470.3	55.6	14.8	3
196	HP 9000/735-99	131.0	284.0	102.0	18.6	3
197	Pentium 233 MMX	177.0	480.2	50.3	14.1	3
198	HP 9000/J200-100	121.1	292.3	92.4	12.4	3
199	SUN SPARC 20/HS21	132.3	319.0	84.0	16.3	3
200	HP 9000/715-100	176.5	354.0	106.7	17.3	2
201	IBM RS/6000-370	370.0	844.0	93.0	44.2	2
202	DEC AXP/3000-600	313.9	540.5	79.0	21.6	2
203	SGI Indigo2 R4400/150	234.0	522.9	77.5	13.7	2
204	SGI Challenge L/150	233.0	517.0	78.0	13.3	2
205	HP 9000/715-80	220.0	426.6	133.2	17.1	2
206	IBM RS/6000-360	433.0	1071.0	117.0	35.1	2
207	DEC AXP/3000-300	426.6	715.5	98.1	12.0	1
208	IBM RS/6000-550	563.0	1266.0	149.0	39.9	1
209	HP 9000/750-66	237.0	554.0	163.0	11.5	1
210	SGI R4000 Indigo	375.0	809.0	112.0	9.5	1
211	IBM RS/6000-350	523.0	1288.0	141.0	27.4	1
212	SGI Challenge L/100	400.0	794.0	130.0	9.5	1
213	DEC AXP/3000-500	194.0	853.0	248.0	12.1	1
214	IBM PowerPC-250	409.0	1258.0	145.0	15.6	1
215	Convex C-220	1015.0	1017.0	329.0	33.1	1
216	SUN SPARCserver 1000	441.0	783.0	148.0	6.7	1
217	IBM RS/6000-340	692.0	1608.0	187.0	21.2	1
218	SUN SPARC 5/85	1036.8	1082.2	195.0	6.7	1
219	SUN SPARC 10/41	665.0	1788.0	266.0	6.6	1
220	SUN SPARC 10/30	758.0	2214.0	299.0	6.9	1
	MPP nodes
1	IBM RS/6000-SP/375	14.6	48.6	8.5	260.0	24
2	IBM RS/6000-SP/222	24.3	83.0	14.0	195.1	15
3	IBM SP2/160Thin	59.3	108.0	16.6	244.9	12
4	Cray T3E/1200	26.3	119.2	18.4	147.8	11
5	Cray T3E/900	36.2	158.2	22.8	139.4	9
6	HP 9000/V2200	50.7	111.1	21.2	106.7	9
7	IBM SP2/120Thin	79.3	141.8	22.4	173.9	9
8	Hitachi SR2201	133.9	309.1	45.4	100.4	5
9	IBM SP2/66Thin	137.7	382.8	42.5	68.6	4
	Vector Supercomputers
1	NEC SX-6i	32.7	27.1	17.8	8000.0	151
2	NEC SX-5	81.9	28.1	26.9	7892.9	146
4	FUJITSU VPP-300	347.1	64.1	120.0	1811.2	36

(+) Prototype, pre-release System




Table 11. The STREAM Benchmark - TRIAD Rates.

The STREAM Benchmark: TRIAD Rates (MBytes/second) (see text). Performance Relative to the HP RX5670 Itanium2/1500-H.

Rank	Machine	TRIAD rate	Relative Values (%)
		MBytes / second
1	NEC SX-5	47104.2 (*)	1251%
2	NEC SX-6i	26813.3	712%
3	NEC SX-4	15898.0	422%
4	IBM pSeries 690Turbo/1.7 GHz	5870.9	156%
5	IBM eServer p5 570/1.9 GHz	5604.5	149%
6	Compaq Marvel EV7/1000	5396.5	143%
7	HP RX1620 Itanium2/1600-H	4876.8	129%
8	HP RX1620 Itanium2/1600-L	4402.9	117%
9	HP RX2600 Itanium2/1500-L	4027.5	107%
10	HP RX4640 Itanium2/1600-L	3991.5	106%
11	HP 9000/RP3440-1000	3942.6	105%
12	SGI Altix3700/Itanium2-1500	3869.3	103%
13	SGI Altix3700/Itanium2-1300	3857.7	102%
14	IBM pSeries 690Turbo/1.3 GHz	3853.4	102%
15	HP RX4640 Itanium2/1300-H	3839.0	102%
16	Intel Tiger Itanium2/1500	3783.3	100%
17	HP RX5670 Itanium2/1500-H	3766.1	100%
18	HP RX2600 Itanium2/1000-H	3591.2	95%
19	HP Superdome Itanium2/1500	3552.2	94%
20	Pentium4/3200 (EM64T)	3446.1	92%
21	SunFire V20 2.2GHz (PathScale)	3439.6	91%
22	HP RX5670 Itanium2/1000-L	3433.1	91%
23	HP DL380 Pentium4/3600 (EM64T)	3428.6	91%
24	Intel Tiger Itanium2/1000	3385.1	90%
25	Intel Tiger Itanium2/1200 (+)	3373.0	90%
26	Pentium4/3400 (EM64T,32-bit)	3250.5	86%
27	Pentium4/3400 (EM64T)	3211.3	85%
28	HP DL145 Opteron248/2200 (pgi5.2)	3000.0	80%
29	Pentium 4/3066 (ifc)	2866.5	76%
30	Pentium 4/2666 (ifc)	2797.3	74%
31	AMD Opteron848/2200 (pgi5.1)	2635.0	70%
32	HP ZX6000 Itanium2/900	2634.0	70%
33	AMD Opteron850/2400 (pgi5.2)	2581.6	69%
34	AMD Opteron846/2000 (pgi5.0)	2566.2	68%
35	AMD Opteron848/2200 (pgi5.2)	2557.3	68%
36	IBM Regatta-HPC/1300	2528.4	67%
37	Pentium 4/2666 (pgi)	2400.0	64%
38	SunFire V40z 2.2GHz (pgi5.2)	2299.7	61%
39	AMD Opteron848/2200 (pgi5.0)	2270.1	60%
40	SunFire V20 2.2GHz (pgi5.2)	2260.4	60%
41	Compaq Alpha ES40/667	2240.0	59%
42	AMD Opteron244/1800 (pgi5.0)	2152.7	57%
43	AMD Opteron846/2000 (pgi5.1)	2097.4	56%
44	AMD Opteron244/1800 (pgi5.1)	2064.1	55%
45	Compaq Alpha ES45/1250	2036.0	54%
46	Pentium 4/3066	2000.0	53%
47	Pentium 4/2666	2000.0	53%
48	Compaq Alpha ES45/1000	1943.9	52%
49	AMD Opteron842/1600 (pgi5.0)	1822.2	48%
50	HP BL20p G2 Pentium 4/2800	1745.6	46%
51	AMD Opteron840/1400 (pgi5.0)	1716.0	46%
52	IBM pSeries 630/1000	1690.3	45%
53	Pentium 4/2000 (ifc)	1600.0	42%
54	Pentium 4/1500	1600.0	42%
55	Pentium 4/1500 (pgi)	1600.0	42%
56	Pentium 4/1400	1600.0	42%
57	Pentium 4/1400 (pgi)	1600.0	42%
58	Pentium 4/2000	1600.0	42%
59	Pentium 4/2000 (pgi)	1600.0	42%
60	HP DL140 Pentium 4/3200	1478.7	39%
61	Compaq Alpha ES40/833	1281.8	34%
62	HP Itanium/733-2M L3	1238.0	33%
63	Pentium 4/2533	1200.0	32%
64	Pentium 4/2533 (pgi)	1200.0	32%
65	Compaq Alpha DS20E/667	1199.5	32%
66	Compaq Alpha DS20/500	1171.0	31%
67	Compaq Alpha ES40/500	1171.0	31%
68	IBM Itanium/800-4M L3	1151.6	31%
69	SUN Blade 1000/M1750	1138.5	30%
70	AlphaPC 264DP-500	1063.8	28%
71	API UP2000 6/667	1045.8	28%
72	API UP2000 6/833	1010.0	27%
73	SUN Fire V880/900-Cu	987.6	26%
74	HP 9000/J6700-750	986.6	26%
75	HP 9000/J6000-552	979.3	26%
76	SUN Fire 6800/900-Cu	977.5	26%
77	SUN Blade 2000/1056-Cu	963.3	26%
78	Compaq PW XP1000/500	960.0	25%
79	Compaq PW XP1000/667	919.3	24%
80	IBM RS/6000-397	882.4	23%
81	AMD MP2000+/1667 (pgi)	800.0	21%
82	AMD MP2400+/2000 (pgi)	800.0	21%
83	AMD Athlon K7/1400(pgi)	800.0	21%
84	AMD MP1800+/1533 (pgi)	800.0	21%
85	IBM RS/6000 44P-270	789.7	21%
86	IBM SP2/120Thin	782.1	21%
87	IBM RS/6000-SP/375	780.7	21%
88	HP 9000/RP7410-875	771.0	20%
89	IBM RS/6000-43P	763.8	20%
90	AMD MP2400+/2000	720.0	19%
91	AMD MP2000+/1667	720.0	19%
92	HP 9000/J5000-440	692.0	18%
93	IBM RS/6000-595	689.5	18%
94	AMD Athlon K7/1400	685.7	18%
95	Pentium III/733-CM(pgi)	685.7	18%
96	AMD Athlon K7/1200(pgi)	685.7	18%
97	SGI O300/R14k-500	668.9	18%
98	AMD MP1800+/1533	654.5	17%
99	SGI Onyx 300/R14k-600	649.0	17%
100	HP 9000/C3000-400	637.0	17%
101	SGI O3800/R14k-500	628.3	17%
102	Cray T3E/1200	604.0	16%
103	Compaq Alpha DS10/466	603.4	16%
104	AMD Athlon K7/1200	600.0	16%
105	SGI O3800/R12k-400	588.2	16%
106	IBM RS/6000-SP/222	563.1	15%
107	AMD Athlon K7/650 (pgi)	533.3	14%
108	AMD Athlon K7/850	533.3	14%
109	AMD Athlon K7/850 (pgi)	533.3	14%
110	AMD Athlon K7/650	533.3	14%
111	Compaq XP1000 6/450	528.8	14%
112	SGI Troon IA64/666	522.9	14%
113	Cray T3E/900	515.0	14%
114	SGI Octane2/R12k-400	514.1	14%
115	Compaq Alpha GS140	491.8	13%
116	AMD Athlon K7/600	480.0	13%
117	Pentium III/866-CM(pgi)	480.0	13%
118	AMD Athlon K7/600 (pgi)	480.0	13%
119	Pentium III/800-CM(pgi)	480.0	13%
120	AMD Athlon K7/500 (pgi)	480.0	13%
121	Pentium III/800 (pgi)	480.0	13%
122	AMD Athlon K7/1000(pgi)	480.0	13%
123	HP 9000/N4000-440	468.7	12%
124	Pentium III/1000-CM(pgi)	436.5	12%
125	Pentium III/650-CM(pgi)	436.4	12%
126	AMD Athlon K7/1000	436.4	12%
127	AMD Athlon K7/500	436.4	12%
128	Pentium III/733-CM	400.0	11%
129	Pentium III/750-CM(pgi)	400.0	11%
130	Pentium III/550 (pgi)	400.0	11%
131	SGI Octane/R12k-270	393.4	10%
132	SGI Octane/R10k-250	389.3	10%
133	SGI O2000/R12k-300	388.0	10%
134	SGI O2000/R12k-400	380.0	10%
135	SUN Ultra80/450	375.2	10%
136	Pentium III/1000-CM	369.2	10%
137	Pentium III/800-CM	369.2	10%
138	SGI O2000/R10k-250	358.0	10%
139	SGI Octane/R10k-195	351.8	9%
140	Pentium II/400 (pgi)	342.9	9%
141	Pentium II/400 (abs)	342.9	9%
142	Pentium III/866-CM	342.9	9%
143	SUN HPC4500/400	326.5	9%
144	Pentium III/650-CM	320.0	8%
145	Pentium III/750-CM	320.0	8%
146	HP 9000/785-367	320.0	8%
147	SGI O2000/R10k-195	317.0	8%
148	SGI Octane/R10k-175	316.6	8%
149	SGI Origin200/225	310.9	8%
150	Pentium III/800	300.0	8%
151	Pentium II/400	300.0	8%
152	SGI Origin200/180	286.0	8%
153	SUN HPC4500/336	285.0	8%
154	Pentium III/550	282.4	7%
155	Pentium III/500	270.4	7%
156	SUN Ultra30/300	268.0	7%
157	ProLiant PII/450	266.7	7%
158	Pentium II/450	266.0	7%
159	DEC Alpha 1200/5-533	259.0	7%
160	SUN Ultra-1/170	253.8	7%
161	DEC Alpha PW/600AU	246.0	7%
162	HP 9000/V2200	235.3	6%
163	DEC Alpha PW/433AU	223.5	6%
164	HP 9000/V2250-240	217.0	6%
165	DEC Alpha 8400/5-625	214.8	6%
166	HP 9000/C240	212.6	6%
167	Pentium II/300 (abs)	209.0	6%
168	Pentium II/300	209.0	6%
169	DEC Alpha 8400/5-300	196.7	5%
170	HP 9000/C160	194.6	5%
171	SGI PChall-R10k/195	193.9	5%
172	Pentium II/266 (pgi)	192.0	5%
173	Pentium II/266 (abs)	192.0	5%
174	SUN Ultra-2/200	189.9	5%
175	Cray T3D/AXP-150	176.8	5%
176	Pentium 233 MMX	160.0	4%
177	SUN Ultra-1/140	154.4	4%
178	SGI PChall-R8k/75	134.8	4%
179	SGI Indigo2-R10k/175	116.7	3%
180	Stardent VISTRA-800	115.2	3%
181	IBM RS/6000-530H	114.3	3%
182	DEC AXP/3000-500	99.7	3%
183	SGI R8k Indigo2	82.1	2%
184	SGI O2 R5k/300	80.2	2%
185	HP 9000/735-99	80.0	2%
186	SGI Indigo2 R4400/250	76.0	2%
187	HP 9000/735-125	73.6	2%
188	SGI O2 R12k/270	72.8	2%
189	IBM PowerPC-43P	70.6	2%
190	SGI Indigo2 R4400/150	70.6	2%
191	Stardent 1520	70.1	2%
192	SGI O2 R5k/180	69.3	2%
193	IBM RS/6000-320	60.0	2%
194	SUN SPARC 10/30	57.0	2%
195	SGI Challenge L/150	54.5	1%
196	SGI Challenge L/100	54.5	1%
197	SUN SPARC 10/41	54.0	1%
198	HP 9000/720-50	53.3	1%
199	DEC AXP/3000-300	38.9	1%
200	DEC S5000/200	23.0	1%
201	SGI 4D/220	19.0	1%

(*) value from STREAM web site = 47779.0
(+) Prototype, pre-release System




Table 12. The GAMESS-UK-89 Benchmark.

The GAMESS-UK-89 Benchmark: Total CPU Time (User and System), Elapsed Time (minutes) and Efficiency (%) for Calculations 1-12 (see text)

Rank	Machine	CPU Time			Elapsed	Efficiency	Relative
		User	System	Total	Time	(%)	Performance (%)
1	AMD Opteron850/2400 (pgi5.2)	1.7	0.2	2.0	2.6	75%	118%
2	IBM eServer p5 570/1.9 GHz	1.8	0.2	2.0	2.2	91%	118%
3	HP DL145 Opteron248/2200 (pgi5.2)	1.9	0.3	2.1	2.3	92%	110%
4	AMD Opteron848/2200 (pgi5.2)	1.9	0.3	2.1	2.5	85%	110%
5	SunFire V20 2.2GHz (pgi5.2)	1.9	0.3	2.2	2.3	94%	109%
6	SunFire V40z 2.2GHz (pgi5.2)	1.9	0.3	2.2	2.3	95%	109%
7	HP RX1620 Itanium2/1600-H	2.0	0.3	2.2	2.4	93%	105%
8	HP RX1620 Itanium2/1600-L	2.0	0.2	2.3	2.5	91%	103%
9	HP DL380 Pentium4/3600 (EM64T)	2.0	0.3	2.3	2.4	97%	102%
10	HP RX4640 Itanium2/1600-L	2.0	0.3	2.3	2.4	97%	101%
11	IBM pSeries 690Turbo/1.7 GHz	2.0	0.4	2.3	3.3	71%	100%
12	HP RX5670 Itanium2/1500-H	2.1	0.3	2.3	2.4	97%	100%
13	AMD Opteron848/2200 (pgi5.1)	2.1	0.3	2.4	2.5	95%	98%
14	Pentium4/3400 (EM64T,32-bit)	2.2	0.2	2.5	3.6	68%	95%
15	SGI Altix3700/Itanium2-1500	2.1	0.3	2.5	2.5	101%	95%
16	AMD Opteron848/2200 (pgi5.0)	2.2	0.3	2.5	2.6	96%	95%
17	HP RX2600 Itanium2/1500-L	2.3	0.2	2.5	2.5	100%	94%
18	AMD Opteron846/2000 (pgi5.1)	2.2	0.3	2.5	2.7	92%	94%
19	Intel Tiger Itanium2/1500	2.3	0.2	2.5	2.8	91%	93%
20	Pentium4/3200 (EM64T)	2.3	0.3	2.6	3.0	87%	91%
21	HP Superdome Itanium2/1500	2.1	0.5	2.6	2.7	96%	90%
22	AMD Opteron846/2000 (pgi5.0)	2.4	0.3	2.7	2.8	93%	88%
23	HP DL140 Pentium 4/3200	2.3	0.4	2.7	2.9	93%	87%
24	HP RX4640 Itanium2/1300-H	2.4	0.3	2.7	2.8	97%	85%
25	AMD Opteron244/1800 (pgi5.1)	2.5	0.4	2.8	3.4	84%	83%
26	SGI Altix3700/Itanium2-1300	2.8	0.2	3.0	3.0	100%	79%
27	Pentium 4/3066 (ifc)	2.6	0.4	3.0	3.1	96%	78%
28	HP BL20p G2 Pentium 4/2800	2.6	0.4	3.0	3.3	92%	78%
29	IBM Regatta-HPC/1300	2.7	0.4	3.1	3.2	95%	76%
30	AMD Opteron244/1800 (pgi5.0)	2.7	0.4	3.1	3.6	85%	76%
31	Intel Tiger Itanium2/1200 (+)	2.9	0.3	3.2	3.3	97%	74%
32	IBM pSeries 690Turbo/1.3 GHz	2.6	0.6	3.2	4.1	78%	73%
33	HP 9000/RP3440-1000	2.9	0.4	3.3	3.6	90%	72%
34	Pentium 4/2666 (ifc)	2.9	0.4	3.3	3.6	93%	70%
35	AMD Opteron842/1600 (pgi5.0)	3.0	0.4	3.4	3.5	96%	70%
36	Intel Tiger Itanium2/1000	3.4	0.4	3.8	4.1	93%	62%
37	HP RX2600 Itanium2/1000-H	3.2	0.6	3.8	3.9	99%	61%
38	Pentium 4/2666 (pgi)	3.5	0.4	3.8	4.2	92%	61%
39	AMD Opteron840/1400 (pgi5.0)	3.4	0.5	3.8	4.0	96%	61%
40	HP RX5670 Itanium2/1000-L	3.5	0.4	3.9	4.1	95%	61%
41	Pentium 4/3066	3.6	0.4	3.9	4.1	96%	60%
42	Compaq Alpha ES45/1250	3.0	1.0	4.0	4.0	101%	58%
43	IBM pSeries 630/1000	3.5	0.6	4.1	4.4	94%	57%
44	HP 9000/RP7410-875	3.5	0.7	4.1	4.2	98%	57%
45	Pentium 4/2533 (pgi)	3.9	0.4	4.3	4.8	88%	55%
46	HP ZX6000 Itanium2/900	4.0	0.3	4.3	4.3	100%	55%
47	Compaq Marvel EV7/1000	3.2	1.2	4.4	4.6	95%	53%
48	AMD MP2400+/2000 (pgi)	3.8	0.6	4.4	4.4	100%	53%
49	Pentium 4/2666	4.0	0.4	4.4	4.6	96%	53%
50	Compaq Alpha ES45/1000	3.7	0.9	4.6	4.7	98%	51%
51	Pentium 4/2000 (ifc)	4.4	0.4	4.8	4.8	99%	49%
52	AMD MP2000+/1667 (pgi)	4.3	0.8	5.0	4.9	103%	47%
53	HP 9000/J6700-750	4.4	0.7	5.1	5.7	89%	46%
54	Pentium 4/2533	4.8	0.4	5.2	5.6	92%	45%
55	Pentium 4/2000 (pgi)	5.0	0.4	5.4	5.5	99%	43%
56	AMD MP1800+/1533 (pgi)	4.8	0.7	5.5	5.5	100%	43%
57	SUN Blade 2000/1056-Cu	5.4	0.5	5.9	6.0	99%	40%
58	AMD MP2000+/1667	5.3	0.8	6.1	6.1	101%	38%
59	Compaq Alpha ES40/833	4.4	1.9	6.3	6.4	99%	37%
60	AMD Athlon K7/1400(pgi)	5.4	1.1	6.5	6.6	99%	36%
61	Pentium 4/1500 (pgi)	6.2	0.3	6.5	6.5	100%	36%
62	Pentium 4/2000	6.3	0.4	6.7	6.7	100%	35%
63	AMD MP1800+/1533	6.0	0.7	6.7	6.7	100%	35%
64	Compaq Alpha ES40/667	5.6	1.2	6.8	7.1	95%	35%
65	HP 9000/J6000-552	6.1	0.7	6.8	7.3	93%	35%
66	SUN Fire V880/900-Cu	6.5	0.6	7.0	7.2	98%	33%
67	AMD Athlon K7/1200(pgi)	6.3	0.9	7.2	10.9	66%	33%
68	Compaq Alpha DS20E/667	5.5	1.9	7.4	7.7	96%	32%
69	Compaq Alpha ES40/667 (+)	5.5	1.9	7.4	7.7	96%	32%
70	Compaq PW XP1000/667	6.3	1.4	7.7	9.0	85%	31%
71	AMD Athlon K7/1400	6.7	1.1	7.7	7.8	99%	30%
72	SUN Fire 6800/900-Cu	6.8	0.9	7.8	8.0	97%	30%
73	SGI Onyx 300/R14k-600	6.8	1.2	8.0	8.0	100%	29%
74	Pentium 4/1500	7.8	0.3	8.1	8.1	100%	29%
75	IBM RS/6000-SP/375	7.2	1.0	8.2	9.2	89%	29%
76	IBM RS/6000 44P-270	7.4	1.0	8.4	9.6	87%	28%
77	AMD Athlon K7/1200	7.7	0.8	8.5	12.4	68%	28%
78	AMD Athlon K7/1000(pgi)	7.8	1.2	9.0	9.8	92%	26%
79	IBM Itanium/800-4M L3	7.6	1.5	9.0	17.0	53%	26%
80	HP 9000/J5000-440	7.9	1.2	9.1	10.1	90%	26%
81	API UP2000 6/833	5.6	3.7	9.2	9.4	98%	25%
82	Pentium III/1000-CM(pgi)	8.2	1.2	9.4	9.4	100%	25%
83	SGI O300/R14k-500	8.1	1.3	9.5	9.5	100%	25%
84	HP 9000/C3000-400	8.4	1.1	9.5	11.1	85%	25%
85	HP Itanium/733-2M L3	7.1	2.4	9.5	16.6	57%	25%
86	Compaq Alpha ES40/500	8.0	1.9	9.9	11.2	89%	24%
87	SGI O3800/R14k-500	8.0	1.9	10.0	12.4	80%	24%
88	API UP2000 6/667	6.9	3.1	10.0	10.4	96%	24%
89	SUN Blade 1000/M1750	9.5	0.8	10.2	10.4	99%	23%
90	Compaq Alpha DS20/500	8.0	2.3	10.3	13.3	77%	23%
91	Compaq Alpha DS20/500 (+)	8.0	2.3	10.3	13.3	77%	23%
92	Pentium 4/1400 (pgi)	8.4	2.0	10.3	12.1	85%	23%
93	AMD Athlon K7/1000	9.3	1.1	10.4	11.1	94%	23%
94	DEC Alpha 8400/6-575-S	8.5	2.0	10.5	16.2	65%	22%
95	DEC Alpha 8400/6-575	8.5	2.0	10.5	11.1	95%	22%
96	Pentium III/866-CM(pgi)	9.3	1.3	10.5	17.2	61%	22%
97	Compaq PW XP1000/500	8.3	2.4	10.7	11.2	95%	22%
98	Compaq XP1000 6/450	9.2	1.6	10.8	17.6	62%	22%
99	AMD Athlon K7/850	9.6	1.3	10.9	11.5	95%	22%
100	AlphaPC 264DP-500	8.9	2.0	11.0	11.1	99%	21%
101	Pentium III/733-CM(pgi)	10.2	0.8	11.0	11.3	97%	21%
102	Pentium III/800-CM(pgi)	10.2	1.0	11.2	13.0	86%	21%
103	Pentium III/1000-CM	10.0	1.3	11.3	11.2	100%	21%
104	SGI Octane2/R12k-400	9.9	1.6	11.5	14.2	81%	20%
105	HP 9000/785-367	10.2	1.6	11.8	14.4	82%	20%
106	Pentium III/866-CM	10.8	1.3	12.1	19.2	63%	19%
107	SGI O3800/R12k-400	9.7	2.5	12.2	15.5	78%	19%
108	SGI O2000/R12k-400	9.9	2.3	12.2	12.3	99%	19%
109	Compaq Alpha DS10/466	9.2	3.1	12.3	52.1	24%	19%
110	Pentium 4/1400	10.7	1.9	12.6	14.3	88%	19%
111	Compaq Alpha GS140	8.2	4.4	12.6	13.1	96%	19%
112	Pentium III/733-CM	12.1	0.9	12.9	13.0	100%	18%
113	Pentium III/800-CM	12.1	1.0	13.1	14.9	88%	18%
114	AMD Athlon K7/650	11.8	1.4	13.2	14.4	92%	18%
115	Pentium III/750-CM(pgi)	11.0	2.4	13.4	13.6	98%	18%
116	Pentium III/800	12.6	1.5	14.0	15.6	90%	17%
117	IBM RS/6000-SP/222	12.7	1.8	14.5	15.6	93%	16%
118	Pentium III/650-CM(pgi)	12.1	2.4	14.6	14.6	100%	16%
119	IBM RS/6000-43P	13.2	1.5	14.8	16.3	90%	16%
120	SGI O2000/R12k-300	12.8	2.3	15.1	15.1	100%	16%
121	Pentium III/750-CM	12.9	2.3	15.2	15.6	98%	15%
122	AMD Athlon K7/600	13.7	1.6	15.2	17.2	89%	15%
123	SUN Ultra80/450	14.4	1.5	15.9	16.6	96%	15%
124	HP 9000/C240	14.4	1.7	16.1	20.2	80%	15%
125	Pentium III/650-CM	14.4	2.4	16.8	17.4	97%	14%
126	SGI Octane/R12k-270	14.6	2.3	16.9	29.2	58%	14%
127	DEC Alpha PW/600AU	14.0	3.1	17.1	24.3	70%	14%
128	AMD Athlon K7/500	16.1	2.1	18.1	26.9	67%	13%
129	SUN HPC3500/400	16.4	1.8	18.2	19.5	93%	13%
130	SUN HPC4500/400	16.9	1.8	18.6	21.1	88%	13%
131	SGI O2000/R10k-250	16.6	2.6	19.2	19.7	97%	12%
132	Pentium III/550	17.6	1.8	19.4	19.7	99%	12%
133	HP 9000/V2200	17.0	2.7	19.7	21.3	92%	12%
134	SGI Origin200/225	17.5	2.5	20.0	20.3	98%	12%
135	DEC Alpha 8400/5-625	14.4	6.0	20.4	20.4	100%	12%
136	IBM RS/6000-397 (+)	18.9	1.5	20.4	24.6	83%	11%
137	IBM RS/6000-397	18.9	1.5	20.4	24.6	83%	11%
138	SUN HPC4500/336	18.8	1.9	20.8	29.9	69%	11%
139	DEC Alpha PW/433AU	18.8	3.6	22.4	25.1	89%	10%
140	DEC Alpha 1200/5-533	14.3	8.3	22.6	28.5	79%	10%
141	Pentium II/450	21.4	1.8	23.2	28.9	80%	10%
142	SGI Octane/R10k-195	20.6	2.6	23.2	24.2	96%	10%
143	SGI O2000/R10k-195 (+)	20.2	3.1	23.2	27.3	85%	10%
144	SGI O2000/R10k-195	20.0	3.2	23.2	24.5	95%	10%
145	HP 9000/C200 (+)	22.1	1.9	24.1	34.8	69%	10%
146	SUN Ultra-2/300 (+)	22.1	2.4	24.4	26.3	93%	10%
147	SUN Ultra30/300	22.7	2.4	25.1	82.6	30%	9%
148	IBM RS/6000-595 (+)	23.2	1.9	25.1	32.9	76%	9%
149	SGI Origin200/180	22.4	3.2	25.6	27.9	92%	9%
150	DEC Alpha 500/5-500	17.7	8.0	25.7	29.5	87%	9%
151	SGI Octane/R10k-175	23.0	3.0	26.0	27.8	93%	9%
152	DEC Alpha 500/5-400	21.5	4.7	26.2	37.2	70%	9%
153	SGI PChall-R10k/195 (+)	23.1	3.6	26.7	36.6	73%	9%
154	Pentium II/400	24.4	2.5	26.9	36.9	73%	9%
155	IBM SP2/120Thin	25.0	2.0	26.9	32.2	84%	9%
156	Pentium III/500	25.5	2.8	28.2	29.9	94%	8%
157	SGI O2 R12k/270	21.4	7.0	28.5	34.1	83%	8%
158	SGI PChall-R10k/175 (+)	26.0	4.8	30.8	70.7	44%	8%
159	DEC Alpha 600/5-333 (+)	23.4	7.6	31.0	34.2	91%	8%
160	HP 9000/C160 (+)	27.7	3.7	31.4	45.9	68%	7%
161	SUN Ultra-2/200 (+)	31.5	3.0	34.5	38.1	91%	7%
162	DEC Alpha 8400/5-300 (+)	25.9	9.5	35.4	37.3	95%	7%
163	Pentium II/300	33.1	2.5	35.6	37.2	96%	7%
164	DEC Alpha 600/5-266 (+)	30.6	6.4	37.0	62.2	59%	6%
165	DEC Alpha 2100/5-250 (+)	30.2	8.2	38.3	43.6	88%	6%
166	SGI O2 R10k/175	29.9	8.7	38.6	48.5	80%	6%
167	Cray YMP-C98/4256 (+)	37.9	2.0	40.0	41.2	97%	6%
168	Pentium II/266	40.7	3.7	44.4	58.7	76%	5%
169	IBM RS/6000-3CT (+)	40.1	4.4	44.5	58.3	76%	5%
170	SUN Ultra-1/170 (+)	40.5	4.9	45.4	56.3	81%	5%
171	SGI O2 R5k/300	38.6	9.4	48.0	52.3	92%	5%
172	IBM RS/6000-590H	43.9	4.8	48.7	74.5	65%	5%
173	SUN Ultra-1/140 (+)	45.3	5.7	51.0	139.8	36%	5%
174	IBM RS/6000-590 (+)	46.6	4.7	51.2	74.8	68%	5%
175	SGI PChall-R8k/75 (+)	48.1	5.5	53.7	55.2	97%	4%
176	DEC Alpha 250/4-266 (+)	46.0	9.7	55.7	85.7	65%	4%
177	SGI R8k Indigo2 (+)	50.8	7.2	58.0	60.6	96%	4%
178	Pentium Pro/200	55.0	4.5	59.5	91.4	65%	4%
179	HP 9000/735-125 (+)	51.6	8.4	59.9	93.3	64%	4%
180	Cray Y-MP (+)	56.8	6.0	62.8	188.4	33%	4%
181	DEC AXP/3000-700 (+)	53.4	11.5	64.9	105.4	62%	4%
182	SGI Indigo2 R4400/250 (+)	59.1	8.8	67.9	88.2	77%	3%
183	SGI O2 R5k/180	57.4	12.6	70.0	92.0	76%	3%
184	HP 9000/735-99 (+)	64.9	7.3	72.2	93.1	78%	3%
185	HP 9000/755-99 (+)	64.9	8.8	73.7	98.2	75%	3%
186	HP 9000/735-99	62.9	11.1	74.0	112.4	66%	3%
187	PowerPC	67.4	12.4	79.8	98.3	81%	3%
188	Hitachi SR2201	82.7	0.0	82.7	245.3	34%	3%
189	SGI Indy-R5k (+)	72.9	11.3	84.2	101.0	83%	3%
190	DEC AXP/3000-600 (+)	72.1	14.1	86.2	195.8	44%	3%
191	IBM RS/6000-370 (+)	78.7	8.4	87.1	108.8	80%	3%
192	Pentium 233 MMX	73.6	18.7	92.3	127.4	72%	3%
193	HP 9000/715-100 (+)	81.3	14.3	95.6	124.7	77%	2%
194	SUN SPARC 20/HS21 (+)	82.5	16.2	98.7	202.7	49%	2%
195	SGI Challenge L/150 (+)	93.3	10.3	103.5	110.6	94%	2%
196	SGI Indigo2 R4400/150 (+)	94.3	10.0	104.3	120.6	86%	2%
197	HP 9000/715-80 (+)	93.7	15.9	109.6	140.1	78%	2%
198	Pentium Pro/200-f2c	108.8	5.3	114.1	143.0	80%	2%
199	Convex C3840 (+)	124.3	3.5	127.9	210.9	61%	2%
200	HP 9000/750-66 (+)	111.1	17.4	128.5	140.2	92%	2%
201	IBM RS/6000-550 (+)	117.6	11.2	128.8	146.4	88%	2%
202	SUN SPARC 20	115.8	15.4	131.2	152.0	86%	2%
203	DEC AXP/3000-500 (+)	93.0	40.0	133.1	148.6	90%	2%
204	SGI Challenge L/100 (+)	140.8	17.1	158.0	178.0	89%	1%
205	DEC AXP/3000-300 (+)	129.8	39.6	169.4	184.0	92%	1%
206	SGI R4000 Indigo (+)	153.2	17.7	170.9	206.0	83%	1%
207	SUN SPARC 10/41 (+)	201.4	21.0	222.5	260.4	85%	1%
208	SUN SPARCserver 1000 (+)	175.1	50.8	225.9	445.6	51%	1%
209	SUN SPARC/5-85 (+)	240.5	25.8	266.3	299.0	89%	1%
210	SUN SPARC 10/30 (+)	251.4	31.7	283.1	306.3	92%	1%

(+) Version 5.0 of the GAMESS-UK Code
(*) Prototype, pre-release System




Table 13. The GAMESS-UK-97 Benchmark.

The GAMESS-UK Benchmark-97: Total CPU Time (User and System), Elapsed Time (minutes) and Efficiency (%) for Calculations 1-10 (see text)

Rank	Machine	CPU Time			Elapsed	Efficiency	Relative
		User	System	Total	Time	(%)	Performance (%)
1	IBM eServer p5 570/1.9 GHz	15.3	1.3	16.6	16.8	98%	115%
2	HP RX1620 Itanium2/1600-H	17.5	1.3	18.8	19.7	95%	101%
3	HP RX5670 Itanium2/1500-H	17.7	1.3	19.0	19.1	100%	100%
4	HP RX1620 Itanium2/1600-L	18.3	1.2	19.5	20.5	95%	97%
5	HP RX4640 Itanium2/1600-L	17.1	2.4	19.5	20.0	97%	97%
6	HP DL380 Pentium4/3600 (EM64T)	18.5	1.5	20.1	21.1	95%	95%
7	IBM pSeries 690Turbo/1.7 GHz	18.3	2.0	20.3	23.1	88%	93%
8	HP RX2600 Itanium2/1500-L	19.6	1.0	20.5	20.5	100%	92%
9	SGI Altix3700/Itanium2-1500	18.8	1.8	20.6	20.6	100%	92%
10	SunFire V40z 2.2GHz (pgi5.2)	18.9	1.7	20.6	21.0	98%	92%
11	AMD Opteron848/2200 (pgi5.2)	19.0	1.7	20.7	22.1	93%	92%
12	HP Superdome Itanium2/1500	18.3	2.4	20.7	20.8	100%	92%
13	HP DL145 Opteron248/2200 (pgi5.2)	19.3	1.6	20.8	21.9	95%	91%
14	Pentium4/3400 (EM64T,32-bit)	19.7	1.5	21.2	27.2	78%	90%
15	SunFire V20 2.2GHz (pgi5.2)	19.5	1.8	21.3	21.7	98%	89%
16	Intel Tiger Itanium2/1500	20.7	1.4	22.1	23.2	95%	86%
17	HP RX4640 Itanium2/1300-H	20.6	1.7	22.3	23.4	95%	85%
18	Pentium4/3200 (EM64T)	21.1	1.7	22.8	24.4	93%	83%
19	HP DL140 Pentium 4/3200	22.0	2.7	24.7	26.3	94%	77%
20	SGI Altix3700/Itanium2-1300	24.1	1.4	25.4	25.5	100%	75%
21	AMD Opteron848/2200 (pgi5.0)	24.2	1.9	26.0	26.3	99%	73%
22	AMD Opteron848/2200 (pgi5.1)	24.3	2.0	26.3	26.6	99%	72%
23	AMD Opteron846/2000 (pgi5.0)	24.7	2.0	26.8	27.3	98%	71%
24	HP 9000/RP3440-1000	25.1	2.0	27.1	32.7	83%	70%
25	IBM pSeries 690Turbo/1.3 GHz	23.7	3.4	27.1	37.7	72%	70%
26	Pentium 4/3066 (ifc)	25.5	2.2	27.7	28.5	97%	69%
27	HP BL20p G2 Pentium 4/2800	25.6	2.4	28.0	29.0	97%	68%
28	HP RX2600 Itanium2/1000-H	24.9	3.5	28.4	29.8	95%	67%
29	AMD Opteron244/1800 (pgi5.1)	26.2	2.3	28.5	38.5	74%	67%
30	Intel Tiger Itanium2/1200 (+)	27.4	1.8	29.2	29.9	98%	65%
31	IBM Regatta-HPC/1300	27.4	2.2	29.6	29.8	99%	64%
32	Pentium 4/2666 (ifc)	28.2	2.8	31.0	33.2	93%	61%
33	AMD Opteron244/1800 (pgi5.0)	28.7	2.3	31.0	40.1	77%	61%
34	Compaq Alpha ES45/1250	26.0	6.0	31.9	31.9	100%	59%
35	Intel Tiger Itanium2/1000	31.5	2.2	33.7	34.5	98%	56%
36	HP RX5670 Itanium2/1000-L	31.3	2.4	33.8	34.8	97%	56%
37	AMD Opteron842/1600 (pgi5.0)	32.7	2.5	35.2	35.5	99%	54%
38	Pentium 4/2666 (pgi)	33.4	2.5	35.9	38.0	95%	53%
39	IBM pSeries 630/1000	32.4	3.6	36.0	37.5	96%	53%
40	Pentium 4/3066	34.1	2.4	36.4	37.4	97%	52%
41	HP 9000/RP7410-875	32.9	3.8	36.7	37.3	98%	52%
42	Compaq Marvel EV7/1000	30.2	7.1	37.3	37.5	100%	51%
43	Compaq Alpha ES45/1000	32.6	5.1	37.7	38.7	97%	50%
44	HP ZX6000 Itanium2/900	36.7	2.0	38.8	40.5	96%	49%
45	AMD Opteron840/1400 (pgi5.0)	36.0	2.9	38.9	39.1	100%	49%
46	HP 9000/J6700-750	36.5	3.4	39.9	49.7	80%	48%
47	Pentium 4/2666	38.1	2.8	40.8	43.4	94%	46%
48	Pentium 4/2533 (pgi)	39.0	2.3	41.3	46.9	88%	46%
49	AMD MP2400+/2000 (pgi)	44.6	3.9	48.5	48.7	100%	39%
50	Pentium 4/2000 (ifc)	46.2	2.6	48.8	49.7	98%	39%
51	Pentium 4/2533	46.6	2.3	48.9	54.8	89%	39%
52	Compaq Alpha ES40/833	38.5	11.4	49.9	54.3	92%	38%
53	Pentium 4/2000 (pgi)	50.2	2.6	52.7	54.0	98%	36%
54	HP 9000/J6000-552	49.5	3.6	53.1	63.2	84%	36%
55	AMD MP2000+/1667 (pgi)	48.5	5.0	53.5	66.0	81%	35%
56	AMD MP1800+/1533 (pgi)	50.9	4.4	55.3	59.6	93%	34%
57	SUN Blade 2000/1056-Cu	54.5	3.7	58.3	58.4	100%	33%
58	Compaq Alpha DS20E/667	47.7	11.3	59.0	77.0	77%	32%
59	AMD MP2000+/1667	54.2	5.0	59.2	71.5	83%	32%
60	Pentium 4/1500 (pgi)	57.3	2.1	59.5	76.5	78%	32%
61	Compaq PW XP1000/667	57.3	8.4	65.7	103.8	63%	29%
62	AMD Athlon K7/1400(pgi)	59.9	6.7	66.6	82.6	81%	28%
63	API UP2000 6/833	50.0	16.9	66.9	78.6	85%	28%
64	IBM RS/6000 44P-270	62.0	5.2	67.2	72.3	93%	28%
65	SGI Onyx 300/R14k-600	60.3	7.0	67.3	68.5	98%	28%
66	Pentium 4/1500	65.9	2.1	68.0	84.8	80%	28%
67	SUN Fire V880/900-Cu	67.3	3.8	71.1	71.3	100%	27%
68	HP 9000/J5000-440	65.9	7.8	73.7	81.9	90%	26%
69	HP 9000/N4000-440	67.8	6.6	74.4	75.9	98%	25%
70	AMD Athlon K7/1200(pgi)	69.2	5.3	74.5	124.8	60%	25%
71	AMD Athlon K7/1400	67.9	6.7	74.6	93.1	80%	25%
72	HP 9000/C3000-400	68.9	6.8	75.7	83.6	91%	25%
73	SGI O3800/R14k-500	69.1	11.0	80.1	95.3	84%	24%
74	AMD Athlon K7/1200	77.2	5.1	82.3	119.3	69%	23%
75	API UP2000 6/667	71.6	12.2	83.8	107.1	78%	23%
76	Compaq Alpha DS20/500	66.7	18.3	85.0	99.8	85%	22%
77	Compaq PW XP1000/500	75.8	13.9	89.7	109.1	82%	21%
78	SGI O3800/R12k-400	80.4	13.7	94.0	121.8	77%	20%
79	AMD Athlon K7/1000	87.0	7.3	94.3	121.0	78%	20%
80	SGI Octane2/R12k-400	90.6	10.0	100.5	143.6	70%	19%
81	Pentium III/1000-CM(pgi)	94.1	7.9	102.1	102.5	100%	19%
82	AMD Athlon K7/850	94.6	10.3	104.9	136.7	77%	18%
83	Compaq Alpha DS10/466	87.8	17.8	105.6	224.9	47%	18%
84	Pentium III/1000-CM	104.4	7.9	112.3	112.6	100%	17%
85	Pentium III/733-CM(pgi)	108.6	6.2	114.8	148.3	77%	17%
86	AMD Athlon K7/650	112.3	8.9	121.1	188.3	64%	16%
87	SGI O2000/R12k-300	112.7	15.7	128.4	213.4	60%	15%
88	Pentium III/733-CM	128.1	5.7	133.8	162.4	82%	14%
89	Pentium III/750-CM(pgi)	116.7	17.8	134.5	161.8	83%	14%
90	AMD Athlon K7/600	126.3	11.6	137.9	187.0	74%	14%
91	SUN Ultra80/450	130.4	10.1	140.4	142.5	99%	14%
92	AMD Athlon K7/500	146.1	12.9	159.0	232.4	68%	12%
93	Pentium III/550	153.4	10.7	164.1	197.2	83%	12%
94	DEC Alpha 1200/5-533	128.7	45.0	173.8	231.0	75%	11%
95	DEC Alpha PW/433AU	161.2	20.0	181.2	243.8	74%	10%
96	SGI O2000/R10k-195	176.8	19.4	196.2	220.0	89%	10%
97	IBM SP2/120Thin	207.0	20.2	227.2	329.0	69%	8%
98	SGI O2 R12k/270	230.1	46.3	276.4	368.4	75%	7%

(+) Prototype, pre-release System




Table 14. The DL_POLY Benchmark.

The DLPOLY Benchmark: Total CPU Time (User and System) and Elapsed Time (minutes) for Calculations 1-6 (see text) and Performance Relative to the HP RX5670 Itanium2/1500-H.

Rank	Machine	CPU Time			Elapsed	Relative
		User	System	Total	Time	Performance (%)
1	HP RX1620 Itanium2/1600-H	2.8	0.0	2.8	2.8	124%
2	HP RX4640 Itanium2/1600-L	3.1	0.0	3.1	3.1	113%
3	HP RX1620 Itanium2/1600-L	3.1	0.0	3.1	3.1	112%
4	Intel Tiger Itanium2/1500	3.1	0.0	3.1	3.1	111%
5	HP RX2600 Itanium2/1500-L	3.2	0.0	3.2	3.3	108%
6	SGI Altix3700/Itanium2-1500	3.3	0.0	3.3	3.3	106%
7	HP RX5670 Itanium2/1500-H	3.5	0.0	3.5	3.5	100%
8	HP Superdome Itanium2/1500	3.5	0.0	3.5	3.5	100%
9	SGI Altix3700/Itanium2-1300	3.8	0.0	3.8	3.8	91%
10	SunFire V20 2.2GHz (PathScale)	3.9	0.0	3.9	3.9	89%
11	Intel Tiger Itanium2/1200 (+)	4.0	0.0	4.0	4.0	88%
12	HP RX4640 Itanium2/1300-H	4.0	0.0	4.0	4.0	86%
13	AMD Opteron850/2400 (pgi5.2)	4.1	0.0	4.1	4.1	85%
14	HP DL380 Pentium4/3600 (EM64T)	4.3	0.0	4.3	4.3	81%
15	AMD Opteron848/2200 (pgi5.2)	4.4	0.0	4.4	4.3	80%
16	SunFire V40z 2.2GHz (pgi5.2)	4.4	0.0	4.4	4.4	80%
17	SunFire V20 2.2GHz (pgi5.2)	4.4	0.0	4.4	4.4	79%
18	HP DL145 Opteron248/2200 (pgi5.2)	4.4	0.0	4.4	4.4	78%
19	AMD Opteron848/2200 (pgi5.1)	4.6	0.0	4.6	4.6	76%
20	Pentium4/3400 (EM64T)	4.6	0.0	4.6	4.7	76%
21	AMD Opteron848/2200 (pgi5.0)	4.6	0.0	4.6	4.6	75%
22	Intel Tiger Itanium2/1000	4.7	0.0	4.7	4.7	75%
23	HP RX5670 Itanium2/1000-L	4.7	0.0	4.7	4.7	74%
24	Pentium4/3400 (EM64T,32-bit)	4.7	0.0	4.7	4.7	73%
25	IBM eServer p5 570/1.9 GHz	4.8	0.0	4.8	4.8	72%
26	HP DL140 Pentium 4/3200	4.8	0.0	4.8	4.8	72%
27	Pentium4/3200 (EM64T)	4.8	0.0	4.9	4.8	72%
28	AMD Opteron846/2000 (pgi5.1)	5.0	0.0	5.0	5.0	70%
29	AMD Opteron846/2000 (pgi5.0)	5.0	0.0	5.0	5.0	69%
30	HP ZX6000 Itanium2/900	5.3	0.0	5.3	5.3	66%
31	AMD Opteron244/1800 (pgi5.0)	5.5	0.0	5.5	5.5	64%
32	HP RX2600 Itanium2/1000-H	5.5	0.0	5.5	5.6	64%
33	AMD Opteron244/1800 (pgi5.1)	5.5	0.0	5.5	5.5	63%
34	IBM pSeries 690Turbo/1.7 GHz	5.5	0.0	5.5	5.4	63%
35	Compaq Alpha ES45/1250	5.7	0.0	5.7	6.3	61%
36	AMD Opteron842/1600 (pgi5.0)	6.3	0.0	6.3	6.3	55%
37	Pentium 4/3066 (ifc)	6.3	0.0	6.3	6.3	55%
38	Compaq Marvel EV7/1000	6.4	0.0	6.4	6.9	54%
39	HP 9000/RP3440-1000	6.7	0.0	6.7	6.8	52%
40	Compaq Alpha ES45/1000	7.0	0.0	7.0	7.0	49%
41	Pentium 4/2666 (ifc)	7.1	0.0	7.1	7.1	49%
42	AMD Opteron840/1400 (pgi5.0)	7.1	0.0	7.1	7.1	49%
43	IBM pSeries 690Turbo/1.3 GHz	7.3	0.0	7.3	7.2	48%
44	HP 9000/RP7410-875	7.6	0.0	7.6	7.6	46%
45	IBM Regatta-HPC/1300	8.4	0.0	8.4	8.4	41%
46	Compaq Alpha ES40/833	8.4	0.0	8.5	8.5	41%
47	HP 9000/J6700-750	9.0	0.0	9.0	9.0	39%
48	SGI Onyx 300/R14k-600	9.2	0.0	9.2	9.3	38%
49	AMD MP2400+/2000 (pgi)	9.5	0.0	9.6	9.5	36%
50	Pentium 4/2666 (pgi)	9.6	0.0	9.6	9.6	36%
51	Pentium 4/2533 (pgi)	10.4	0.0	10.4	10.4	34%
52	AMD MP2000+/1667 (pgi)	10.4	0.0	10.4	10.4	33%
53	Compaq Alpha DS20E/667	10.5	0.0	10.6	10.5	33%
54	API UP2000 6/833	10.6	0.0	10.6	10.7	33%
55	Compaq Alpha ES40/667	10.8	0.0	10.8	10.7	32%
56	SGI O3800/R14k-500	10.9	0.0	10.9	10.9	32%
57	IBM pSeries 630/1000	10.9	0.0	10.9	10.9	32%
58	SGI O300/R14k-500	11.1	0.0	11.1	11.1	31%
59	Compaq PW XP1000/667	11.2	0.0	11.2	11.2	31%
60	Pentium 4/2000 (ifc)	11.3	0.0	11.3	11.3	31%
61	SUN Blade 2000/1056-Cu	11.4	0.0	11.4	11.4	30%
62	AMD MP1800+/1533 (pgi)	11.6	0.0	11.6	11.6	30%
63	HP 9000/J6000-552	12.2	0.0	12.2	12.3	28%
64	API UP2000 6/667	12.5	0.0	12.5	12.7	28%
65	SGI O2000/R12k-400	13.0	0.0	13.0	13.3	27%
66	AMD Athlon K7/1400(pgi)	13.1	0.0	13.1	13.1	27%
67	SGI Octane2/R12k-400	13.1	0.1	13.2	13.6	26%
68	SGI O3800/R12k-400	13.3	0.0	13.3	13.3	26%
69	SUN Fire V880/900-Cu	13.3	0.0	13.3	13.4	26%
70	Compaq Alpha GS140	13.9	0.0	13.9	13.9	25%
71	SUN Fire 6800/900-Cu	14.1	0.0	14.1	14.1	25%
72	Pentium 4/2000 (pgi)	14.1	0.0	14.1	14.1	25%
73	Compaq Alpha DS20/500	14.4	0.0	14.4	14.4	24%
74	Compaq Alpha ES40/500	14.5	0.0	14.6	14.5	24%
75	AMD Athlon K7/1200(pgi)	14.7	0.0	14.7	14.7	24%
76	Compaq PW XP1000/500	14.8	0.0	14.8	14.8	23%
77	HP Itanium/733-2M L3	10.3	4.9	15.2	15.2	23%
78	IBM Itanium/800-4M L3	15.1	0.1	15.2	15.3	23%
79	Compaq XP1000 6/450	15.6	0.0	15.6	15.6	22%
80	HP 9000/N4000-440	15.9	0.1	16.0	16.1	22%
81	Compaq Alpha DS10/466	16.3	0.2	16.5	33.1	21%
82	SUN Blade 1000/M1750	16.7	0.0	16.7	16.7	21%
83	SGI Troon IA64/666	17.6	0.0	17.6	17.6	20%
84	Pentium 4/1500 (pgi)	17.7	0.0	17.8	17.8	20%
85	AMD Athlon K7/1000(pgi)	17.9	0.0	17.9	17.9	19%
86	AlphaPC 264DP-500	18.4	0.0	18.4	18.8	19%
87	Pentium III/1000-CM(pgi)	18.6	0.0	18.6	18.6	19%
88	SGI O2000/R12k-300	18.7	0.0	18.7	18.8	19%
89	HP 9000/J5000-440	19.6	0.0	19.6	19.7	18%
90	DEC Alpha 8400/5-625	19.7	0.1	19.8	19.9	18%
91	Pentium 4/1400 (pgi)	19.9	0.0	20.0	20.5	17%
92	HP 9000/C3000-400	20.1	0.0	20.1	20.2	17%
93	DEC Alpha PW/600AU	20.2	0.0	20.3	20.2	17%
94	SGI Octane/R12k-270	20.9	0.0	20.9	20.9	17%
95	Pentium III/866-CM(pgi)	21.2	0.0	21.2	21.2	16%
96	AMD Athlon K7/850 (pgi)	21.3	0.0	21.3	21.3	16%
97	SGI O2000/R10k-250	21.6	0.0	21.6	21.7	16%
98	IBM RS/6000 44P-270	21.7	0.0	21.7	21.6	16%
99	IBM RS/6000-SP/375	21.7	0.0	21.7	21.7	16%
100	Pentium III/800-CM(pgi)	23.0	0.0	23.0	23.3	15%
101	DEC Alpha 1200/5-533	23.2	0.1	23.3	24.6	15%
102	Pentium III/800 (pgi)	24.0	0.0	24.1	24.0	14%
103	SGI Origin200/225	24.0	0.0	24.1	24.1	14%
104	SGI Octane/R10k-250	24.5	0.0	24.5	24.5	14%
105	Pentium III/750-CM(pgi)	24.7	0.0	24.7	24.7	14%
106	Pentium III/800-CM(pgi)	24.9	0.0	24.9	24.9	14%
107	AMD Athlon K7/650 (pgi)	26.5	0.0	26.5	26.9	13%
108	DEC Alpha PW/433AU	28.1	0.0	28.2	28.8	12%
109	AMD Athlon K7/600 (pgi)	28.3	0.0	28.3	28.3	12%
110	Pentium III/650-CM(pgi)	28.4	0.0	28.4	28.4	12%
111	SGI O2 R12k/270	29.5	0.1	29.6	30.3	12%
112	SGI O2000/R10k-195	29.9	0.0	29.9	30.8	12%
113	HP 9000/V2250-240	30.7	0.1	30.7	30.8	11%
114	SGI PChall-R10k/195	33.0	0.1	33.1	33.2	11%
115	AMD Athlon K7/500 (pgi)	33.3	0.0	33.3	33.3	10%
116	HP 9000/V2200	33.5	0.1	33.5	33.6	10%
117	Pentium III/550 (pgi)	34.3	0.0	34.3	34.3	10%
118	IBM RS/6000-43P	35.8	0.0	35.8	35.7	10%
119	HP 9000/C240	36.3	0.0	36.3	36.5	10%
120	DEC Alpha 8400/5-300	37.2	0.1	37.2	37.2	9%
121	IBM RS/6000-SP/222	38.0	0.0	38.0	38.0	9%
122	SGI Octane/R10k-175	40.4	0.1	40.4	40.5	9%
123	ProLiant PII/450 (pgi)	41.1	0.0	41.1	41.2	8%
124	Cray T3E/1200	41.2	0.6	41.8	42.6	8%
125	SUN Ultra80/450	44.0	0.0	44.0	44.1	8%
126	SUN HPC4500/400	49.1	0.0	49.1	49.2	7%
127	Pentium II/400 (pgi)	50.4	0.0	50.5	50.5	7%
128	Cray T3E/900	51.0	0.7	51.7	52.3	7%
129	SGI O2 R5k/300	51.7	0.1	51.8	53.0	7%
130	SUN HPC4500/336	62.6	0.0	62.6	62.7	6%
131	Pentium II/300 (pgi)	65.6	0.0	65.6	65.7	5%
132	IBM SP2/120Thin	67.5	0.0	67.5	68.1	5%
133	Pentium II/300 (abs)	72.1	0.0	72.1	72.1	5%
134	Pentium II/266 (pgi)	76.4	0.0	76.5	76.5	5%
135	SGI O2 R5k/180	80.5	0.2	80.7	83.5	4%
136	Pentium II/266 (abs)	83.8	0.0	83.8	83.8	4%
137	IBM RS/6000-59H	107.7	0.0	107.7	108.3	3%

(+) Prototype, pre-release System




Table 15. Summary of the Serial Chemistry Benchmarks.

The Chemistry Benchmark: Performance (%, see text) Relative to the HP RX5670 Itanium2/1500-H.

Rank	Machine	SPEC	SPEC	Matrix-97	Chem.	GAMESS	DLPOLY	Final
		fp2000	int2000	(Mat-89)	Kernels	CPU	CPU	Index(%)
1	HP RX1620 Itanium2/1600-H	128	111	133	112	105	124	119
2	HP RX1620 Itanium2/1600-L	128	121	100	109	103	112	106
3	HP RX4640 Itanium2/1600-L	129	121	102	105	101	113	105
4	IBM eServer p5 570/1.9 GHz	128	111	123	100	118	72	103
5	HP RX5670 Itanium2/1500-H	100	100	100	100	100	100	100
6	SGI Altix3700/Itanium2-1500	100	95	93	102	95	106	99
7	HP Superdome Itanium2/1500	100	100	99	92	90	100	95
8	HP RX2600 Itanium2/1500-L	101	102	89	89	94	108	95
9	Intel Tiger Itanium2/1500	97	93	85	84	93	111	93
10	AMD Opteron850/2400 (pgi5.2)	78	113	71	93	118	85	92
11	HP DL380 Pentium4/3600 (EM64T)	72	118	88	94	102	81	91
12	SunFire V40z 2.2GHz (pgi5.2)	80	111	69	86	109	80	86
13	AMD Opteron848/2200 (pgi5.2)	72	107	67	85	110	80	86
14	SunFire V20 2.2GHz (pgi5.2)	80	111	67	86	109	79	85
15	HP DL145 Opteron248/2200 (pgi5.2)	80	111	67	86	110	78	85
16	HP RX4640 Itanium2/1300-H	86	81	87	81	85	86	85
17	SunFire V20 2.2GHz (PathScale)	80	111	67	92	-	89	83
18	IBM pSeries 690Turbo/1.7 GHz	81	85	94	70	100	63	82
19	Pentium4/3400 (EM64T)	69	113	81	87	-	76	81
20	AMD Opteron848/2200 (pgi5.1)	72	107	67	83	98	76	81
21	SGI Altix3700/Itanium2-1300	85	78	71	81	79	91	81
22	Pentium4/3400 (EM64T,32-bit)	69	113	74	76	95	73	80
23	AMD Opteron848/2200 (pgi5.0)	72	107	61	77	95	75	77
24	HP DL140 Pentium 4/3200	47	96	70	76	87	72	76
25	AMD Opteron846/2000 (pgi5.1)	62	96	62	78	94	70	76
26	Intel Tiger Itanium2/1200 (+)	77	74	67	70	74	88	75
27	AMD Opteron846/2000 (pgi5.0)	62	96	59	72	88	69	72
28	AMD Opteron244/1800 (pgi5.1)	55	89	56	72	83	63	69
29	HP BL20p G2 Pentium 4/2800	47	75	51	68	78	-	66
30	Pentium 4/3066 (ifc)	52	84	52	75	78	55	65
31	HP RX2600 Itanium2/1000-H	62	62	64	68	61	64	64
32	AMD Opteron244/1800 (pgi5.0)	55	89	50	65	76	64	64
33	Intel Tiger Itanium2/1000	64	62	56	61	62	75	64
34	HP RX5670 Itanium2/1000-L	68	62	54	61	61	74	63
35	IBM pSeries 690Turbo/1.3 GHz	60	64	70	55	73	48	62
36	IBM Regatta-HPC/1300	60	64	71	55	76	41	61
37	HP 9000/RP3440-1000	47	65	59	57	72	52	60
38	Pentium 4/2666 (ifc)	49	77	48	65	70	49	58
39	AMD Opteron842/1600 (pgi5.0)	53	80	46	57	70	55	57
40	Compaq Alpha ES45/1250	65	71	49	56	58	61	56
41	HP ZX6000 Itanium2/900	54	51	41	52	55	66	54
42	Pentium 4/3066	52	84	43	47	60	-	50
43	AMD Opteron840/1400 (pgi5.0)	48	71	40	49	61	49	50
44	Compaq Marvel EV7/1000	61	58	42	50	53	54	50
45	Pentium 4/2666 (pgi)	49	77	41	55	61	36	48
46	HP 9000/RP7410-875	32	52	49	38	57	46	48
47	Compaq Alpha ES45/1000	46	52	40	44	51	49	46
48	IBM pSeries 630/1000	42	49	53	39	57	32	45
49	Pentium 4/2666	49	77	38	39	53	-	43
50	Pentium 4/2533 (pgi)	47	73	37	41	55	34	42
51	HP 9000/J6700-750	28	43	45	34	46	39	41
52	AMD MP2400+/2000 (pgi)	31	58	22	51	53	36	41
53	Pentium 4/2533	47	73	35	38	45	-	39
54	Pentium 4/2000 (ifc)	34	50	27	49	49	31	39
55	Compaq Alpha ES40/833	37	43	35	36	37	41	37
56	AMD MP2000+/1667 (pgi)	28	50	20	43	47	33	36
57	AMD MP1800+/1533 (pgi)	26	46	21	40	43	30	34
58	AMD MP2400+/2000	31	58	22	43	-	-	33
59	SUN Blade 2000/1056-Cu	39	44	27	33	40	30	33
60	SGI Onyx 300/R14k-600	23	37	29	30	29	38	32
61	Pentium 4/2000 (pgi)	34	50	23	34	43	25	31
62	Compaq Alpha DS20E/667	28	35	28	30	32	33	31
63	AMD MP2000+/1667	28	50	20	34	38	-	31
64	HP 9000/J6000-552	21	34	31	26	35	28	30
65	Compaq Alpha ES40/667	27	33	25	28	35	32	30
66	API UP2000 6/833	31	41	26	35	25	33	30
67	Pentium 4/2000	34	50	22	32	35	-	30
68	AMD Athlon K7/1400(pgi)	22	42	16	36	36	27	29
69	AMD MP1800+/1533	26	46	20	30	35	-	28
70	Compaq PW XP1000/667	21	31	23	27	31	31	28
71	SUN Fire V880/900-Cu	33	39	24	29	33	26	28
72	SGI O300/R14k-500	18	29	24	31	25	31	28
73	SGI O3800/R14k-500	22	33	23	31	24	32	28
74	Pentium 4/1500 (pgi)	29	41	23	28	36	20	27
75	HP Itanium/733-2M L3	30	42	31	27	25	23	27
76	SUN Fire 6800/900-Cu	33	39	22	29	30	25	27
77	IBM Itanium/800-4M L3	33	50	29	26	26	23	26
78	AMD Athlon K7/1200(pgi)	20	38	15	31	33	24	26
79	Pentium 4/1500	29	41	21	25	29	-	25
80	API UP2000 6/667	17	26	20	28	24	28	25
81	IBM RS/6000-SP/375	18	20	28	24	29	16	24
82	AMD Athlon K7/1400	22	42	15	27	30	-	24
83	IBM RS/6000 44P-270	18	21	26	24	28	16	24
84	Compaq Alpha DS20/500	20	24	20	23	23	24	23
85	SGI O3800/R12k-400	19	27	19	26	19	26	23
86	HP 9000/J5000-440	17	26	24	21	26	18	22
87	SGI Octane2/R12k-400	16	26	19	24	20	26	22
88	SGI O2000/R12k-400	16	26	20	23	19	27	22
89	AMD Athlon K7/1200	20	38	14	23	28	-	22
90	SUN Blade 1000/M1750	20	30	21	21	23	21	22
91	Compaq Alpha ES40/500	20	24	16	22	24	24	22
92	HP 9000/N4000-440	14	0	23	19	-	22	21
93	Pentium 4/1400 (pgi)	28	39	21	24	23	17	21
94	Compaq Alpha GS140	15	24	17	22	19	25	21
95	Compaq PW XP1000/500	18	23	16	21	22	23	21
96	AMD Athlon K7/1000(pgi)	15	29	11	25	26	19	20
97	HP 9000/C3000-400	17	0	19	19	25	17	20
98	Pentium 4/1400	28	39	19	22	19	-	20
99	DEC Alpha 8400/6-575	16	21	18	18	22	-	19
100	Pentium III/1000-CM(pgi)	15	33	11	19	25	19	19
101	Compaq Alpha DS10/466	16	21	13	19	19	21	18
102	AMD Athlon K7/1000	15	29	11	20	23	-	18
103	SGI O2000/R12k-300	13	20	15	18	16	19	17
104	HP 9000/785-367	9	21	14	16	20	-	17
105	AMD Athlon K7/850	13	25	10	17	22	-	16
106	AMD Athlon K7/850 (pgi)	13	25	11	21	-	16	16
107	Pentium III/866-CM(pgi)	14	31	10	16	22	16	16
108	SGI Octane/R12k-270	9	17	13	17	14	17	15
109	Pentium III/733-CM(pgi)	12	26	10	14	21	-	15
110	Pentium III/1000-CM	15	33	10	14	21	-	15
111	Pentium III/800-CM(pgi)	13	28	9	15	21	15	15
112	Pentium III/866-CM	14	31	9	14	19	-	14
113	Pentium III/750-CM(pgi)	11	25	9	14	18	14	14
114	IBM RS/6000-43P	11	13	15	14	16	10	14
115	IBM RS/6000-SP/222	10	14	15	15	16	9	14
116	DEC Alpha PW/600AU	8	14	10	14	14	17	14
117	SGI O2000/R10k-250	9	17	12	14	12	16	14
118	Pentium III/733-CM	12	26	9	13	18	-	13
119	DEC Alpha 8400/5-625	8	15	10	13	12	18	13
120	SGI Onyx2 IR2/R10k-250	9	17	12	14	-	-	13
121	AMD Athlon K7/650 (pgi)	11	20	9	17	-	13	13
122	Pentium III/800-CM	13	28	8	13	18	-	13
123	Pentium III/800	13	28	8	13	17	-	13
124	AMD Athlon K7/650	11	20	8	12	18	-	13
125	SUN Ultra80/450	13	16	13	14	15	8	13
126	AMD Athlon K7/600	10	18	9	13	15	-	12
127	Pentium III/650-CM(pgi)	10	23	8	13	16	12	12
128	HP 9000/C240	8	14	12	11	15	10	12
129	Pentium III/800 (pgi)	13	28	8	14	-	14	12
130	Pentium III/750-CM	11	25	8	13	15	-	12
131	SGI Octane/R10k-250	8	15	9	13	-	14	12
132	DEC Alpha 1200/5-533	8	14	9	13	10	15	12
133	AMD Athlon K7/600 (pgi)	10	18	8	15	-	12	12
134	HP 9000/V2250-240	8	13	13	10	-	11	11
135	HP 9000/V2200	7	11	13	9	12	10	11
136	SUN HPC4500/400	12	16	11	12	13	7	11
137	AMD Athlon K7/500	9	15	8	11	13	-	11
138	Pentium III/650-CM	10	15	7	11	14	-	11
139	AMD Athlon K7/500 (pgi)	9	15	8	13	-	10	10
140	IBM RS/6000-397	9	9	8	12	11	-	10
141	SGI O2 R12k/270	4	13	7	14	8	12	10
142	IBM SP2/160Thin	9	9	8	12	-	-	10
143	SGI O2000/R10k-195	7	10	9	9	10	12	10
144	DEC Alpha PW/433AU	6	12	7	10	10	12	10
145	HP 9000/C200	7	11	9	10	10	-	10
146	Pentium III/550	7	16	7	9	12	-	9
147	SUN HPC4500/336	11	14	10	10	11	6	9
148	DEC Alpha 500/5-500	7	12	7	11	9	-	9
149	Pentium III/550 (pgi)	7	16	7	10	-	10	9
150	IBM RS/6000-595	6	6	( 9)	9	9	-	9
151	SGI Octane/R10k-195	6	10	8	9	10	-	9
152	SGI PChall-R10k/195	5	10	7	8	9	11	9
153	SUN Ultra-2/300	6	11	8	8	10	-	9
154	Cray T3E/1200	8	15	7	11	-	8	9
155	SGI Octane/R10k-175	6	9	7	8	9	9	8
156	DEC Alpha 500/5-400	5	10	6	9	9	-	8
157	SUN Ultra30/300	7	11	7	8	9	-	8
158	ProLiant PII/450	6	12	7	8	-	-	8
159	DEC Alpha 600/5-333	5	8	6	8	8	-	7
160	Pentium II/450	6	13	5	7	10	-	7
161	Cray T3E/900	6	11	6	9	-	7	7
162	DEC Alpha 8400/5-300	4	6	6	7	7	9	7
163	IBM SP2/120Thin	6	6	6	9	9	5	7
164	HP 9000/C160	5	8	7	7	7	-	7
165	Pentium III/500	7	15	6	7	8	-	7
166	Pentium II/400	6	12	5	6	9	-	7
167	Pentium II/400 (abs)	6	12	5	8	-	-	7
168	SUN Ultra-2/200	4	7	( 6)	6	7	-	6
169	Pentium II/400 (pgi)	6	12	5	7	-	7	6
170	SGI O2 R10k/175	3	10	5	7	6	-	6
171	DEC Alpha 2100/5-250	3	5	( 5)	6	6	-	6
172	DEC Alpha 600/5-266	4	7	( 5)	6	6	-	6
173	SGI O2 R5k/300	3	9	3	7	5	7	6
174	SUN Ultra-1/170	3	5	( 5)	5	5	-	5
175	Pentium II/300 (abs)	4	9	4	6	-	5	5
176	Pentium II/300	4	9	3	4	7	-	5
177	SUN Ultra-1/140	3	4	( 4)	4	5	-	4
178	IBM RS/6000-590	4	3	( 4)	4	5	-	4
179	IBM RS/6000-3CT	4	4	( 4)	4	5	-	4
180	Pentium II/266 (pgi)	3	8	3	5	-	5	4
181	IBM SP2/66Thin	3	3	( 4)	4	-	-	4
182	IBM RS/6000-3BT	3	3	( 4)	4	-	-	4
183	Pentium II/266 (abs)	3	8	3	5	-	4	4
184	DEC Alpha 250/4-266	2	4	( 4)	4	4	-	4
185	Pentium II/266	3	8	3	4	5	-	4
186	HP 9000/735-125	1	3	( 4)	3	4	-	4
187	HP 9000/J200-100	2	3	( 4)	3	-	-	4
188	Pentium Pro/200	3	6	2	4	4	-	3
189	DEC AXP/3000-700	2	3	( 3)	3	4	-	3
190	SGI Indy-R5k	2	5	( 3)	4	3	-	3
191	SGI O2 R5k/180	2	5	2	4	3	4	3
192	HP 9000/735-99	1	3	( 3)	3	3	-	3
193	IBM PowerPC-43P	1	4	( 2)	3	-	-	3
194	Pentium 233 MMX	2	0	1	3	3	-	2
195	DEC AXP/3000-600	2	2	( 2)	2	3	-	2
196	HP 9000/715-100	1	2	( 2)	2	2	-	2
197	DEC AXP/3000-500	1	2	( 2)	1	2	-	2
198	IBM PowerPC-250	1	2	( 1)	1	-	-	1
199	SUN SPARC 10/41	1	1	( 1)	1	1	-	1
200	Dell Optiplex/266	3	8	2	0	-	-	1

(+) Prototype, pre-release System
(*) Estimated SPECfp2000 and SPECint200 values (see text)




Table 16. The Matrix Multiplication RATE Benchmark 1. FORTRAN MMO.

FORTRAN MMO Benchmark. Total Elapsed times (seconds) and RATE Estimate Relative to the HP RX5670 Itanium2/1500-H.

Rank	Machine	NCPUs	Wall Time	RATE
			(seconds)	(%)
1	HP RX4640 Itanium2/1600-L 9ML3	1	18.0	147
2	IBM eServer p5 570/1.9 GHz	1	19.2	138
3	HP RX5670 Itanium2/1500-H	1	26.4	100
4	HP RX1620 Itanium2/1600-L 3ML3	1	29.5	90
5	IBM pSeries 690Turbo/1700	1	30.6	86
6	HP RX4640 Itanium2/1300-H	1	33.0	80
7	SGI Altix3700/Itanium2-1500	1	33.9	78
8	IBM pSeries 690Turbo/1300	1	34.4	77
9	HP Superdome Itanium2/1500	1	35.4	75
10	HP DL380 Pentium4/3600 (EM64T)	1	37.9	70
11	HP RX2600 Itanium2/1000-H	1	42.2	63
12	SM X6DAG-G Pentium4/3200 (EM64T)	1	42.5	62
13	IBM pSeries 630/1000	1	45.0	59
14	Compaq Alpha ES45/1250	1	50.7	52
15	HP DL145 Opteron248/2200 (pgi5.2)	1	53.0	50
16	SunFire V20 Opteron248/2.2GHz (pgi5.2)	1	53.1	50
17	SunFire V40z Opteron248/2.2GHz (pgi5.2)	1	53.4	50
18	HP RX2600 Itanium2/1500-L	1	53.4	49
19	AMD Opteron848/2200 (pgi5.2)	1	54.8	48
20	AMD Opteron246/2000 (pgi5.0)	1	59.1	45
21	Pentium 4 Xeon/3066 (ifc)	1	59.2	45
22	Pentium 4 Xeon/2666 (ifc)	1	61.9	43
23	AMD Opteron244/1800 (pgi5.0)	1	66.4	40
24	HP BL20p G2 Pentium 4/2800	1	67.4	39
25	AMD Opteron848/2200 (pgi5.1)	1	67.6	39
26	Pentium 4/2000 (ifc)	1	76.4	35
27	AMD Opteron848/2200 (pgi5.0)	1	76.6	34
28	HP DL140 Pentium 4 Xeon/3200	1	80.3	33
29	Intel Tiger Madison/1500	1	82.2	32
30	AMD Opteron842/1600 (pgi5.0)	1	88.6	30
31	HP RX5670 Itanium2/1000-L	1	89.6	29
32	AMD Opteron840/1400 (pgi5.0)	1	92.7	29
33	SUN Blade 2000/1056-Cu	1	97.2	27
34	SGI Altix3700/Itanium2-1300	1	99.2	27
35	HP ZX6000 Itanium 2/900-L	1	104.0	25
36	SUN Fire V880/900-Cu	1	110.8	24
37	SGI Origin 3800/R14k-500	1	116.2	23
38	Intel Tiger Madison/1200	1	124.3	21
39	Intel Tiger Itanium 2/1000	1	126.0	21
40	AMD MP2000+ / 1667	1	171.1	15
41	SGI Origin 3800/R14k-600	1	208.8	13
42	SUN Fire 6800/750	1	256.4	10
43	API UP2000 6/667	1	300.9	9
44	HP PA-9000/N4000-440	1	330.6	8
	Number of CPUs = 2
1	IBM eServer p5 570/1.9 GHz	2	19.6	270
2	HP RX4640 Itanium2/1600-L 9ML3	2	24.8	213
3	IBM pSeries 690Turbo/1700	2	31.3	169
4	HP RX5670 Itanium2/1500-H	2	34.8	152
5	IBM pSeries 690Turbo/1300	2	37.1	142
6	HP Superdome Itanium2/1500	2	37.9	140
7	SGI Altix3700/Itanium2-1500	2	42.9	123
8	HP RX1620 Itanium2/1600-L 3ML3	2	43.0	123
9	SunFire V20 Opteron248/2.2GHz (pgi5.2)	2	45.3	117
10	HP RX4640 Itanium2/1300-H	2	45.5	116
11	HP DL145 Opteron248/2200 (pgi5.2)	2	49.0	108
12	Compaq Alpha ES45/1250	2	50.8	104
13	HP RX2600 Itanium2/1000-H	2	52.3	101
14	HP RX2600 Itanium2/1500-L	2	59.5	89
15	AMD Opteron848/2200 (pgi5.2)	2	59.6	89
16	IBM pSeries 630/1000	2	63.5	83
17	SunFire V40z Opteron248/2.2GHz (pgi5.2)	2	71.1	74
18	AMD Opteron244/1800 (pgi5.0)	2	76.7	69
19	Intel Tiger Madison/1500	2	84.4	63
20	HP DL380 Pentium4/3600 (EM64T)	2	88.3	60
21	AMD Opteron848/2200 (pgi5.1)	2	88.6	60
22	AMD Opteron842/1600 (pgi5.0)	2	91.1	58
23	AMD Opteron246/2000 (pgi5.0)	2	91.2	58
24	AMD Opteron840/1400 (pgi5.0)	2	94.5	56
25	HP RX5670 Itanium2/1000-L	2	99.3	53
26	SUN Blade 2000/1056-Cu	2	99.4	53
27	AMD Opteron848/2200 (pgi5.0)	2	101.1	52
28	SGI Altix3700/Itanium2-1300	2	104.9	50
29	SM X6DAG-G Pentium4/3200 (EM64T)	2	107.1	49
30	HP BL20p G2 Pentium 4/2800	2	120.3	44
31	SGI Origin 3800/R14k-500	2	123.0	43
32	Pentium 4 Xeon/2666 (ifc)	2	125.2	42
33	Pentium 4 Xeon/3066 (ifc)	2	128.1	41
34	Intel Tiger Madison/1200	2	129.4	41
35	Pentium 4/2000 (ifc)	2	129.9	41
36	Intel Tiger Itanium 2/1000	2	133.8	40
37	HP DL140 Pentium 4 Xeon/3200	2	154.9	34
38	SUN Fire 6800/750	2	258.7	20
39	AMD MP2000+ / 1667	2	262.8	20
40	API UP2000 6/667	2	331.6	16
41	HP PA-9000/N4000-440	2	354.3	15
	Number of CPUs = 4
1	IBM eServer p5 570/1.9 GHz	4	25.2	420
2	HP RX4640 Itanium2/1600-L 9ML3	4	28.0	377
3	IBM pSeries 690Turbo/1700	4	31.5	336
4	SGI Altix3700/Itanium2-1500	4	41.0	258
5	IBM pSeries 690Turbo/1300	4	46.4	228
6	HP Superdome Itanium2/1500	4	46.5	227
7	Compaq Alpha ES45/1250	4	52.0	203
8	HP RX5670 Itanium2/1500-H	4	54.1	195
9	SunFire V40z Opteron248/2.2GHz (pgi5.2)	4	55.2	191
10	AMD Opteron848/2200 (pgi5.2)	4	70.3	150
11	IBM pSeries 630/1000	4	78.7	134
12	HP RX4640 Itanium2/1300-H	4	79.5	133
13	Intel Tiger Madison/1500	4	91.6	115
14	AMD Opteron842/1600 (pgi5.0)	4	96.8	109
15	SGI Altix3700/Itanium2-1300	4	105.2	100
16	SUN Fire V880/900-Cu	4	111.7	95
17	AMD Opteron848/2200 (pgi5.0)	4	116.4	91
18	SGI Origin 3800/R14k-500	4	122.0	87
19	HP RX5670 Itanium2/1000-L	4	130.5	81
20	Intel Tiger Itanium 2/1000	4	137.8	77
21	Intel Tiger Madison/1200	4	142.7	74
22	AMD Opteron848/2200 (pgi5.1)	4	145.7	73
23	HP PA-9000/N4000-440	4	383.4	28
	Number of CPUs = 8
1	IBM pSeries 690Turbo/1700	8	33.4	634
2	SGI Altix3700/Itanium2-1500	8	34.8	608
3	HP Superdome Itanium2/1500	8	49.1	430
4	IBM pSeries 690Turbo/1300	8	61.2	345
5	SGI Altix3700/Itanium2-1300	8	113.6	186
6	SUN Fire V880/900-Cu	8	115.5	183
7	HP PA-9000/N4000-440	8	488.7	43

Table 17. The Matrix Multiplication RATE Benchmark 2. MXMB0.

The MXMB0 Benchmark. Total Elapsed times (seconds) and RATE Estimate Relative to the HP RX5670 Itanium2/1500-H.

Rank	Machine	NCPUs	Wall Time	RATE
			(seconds)
1	HP RX5670 Itanium2/1500-H	1	10.4	100
2	HP Superdome Itanium2/1500	1	10.7	98
3	HP RX4640 Itanium2/1600-L 9ML3	1	11.8	89
4	HP RX1620 Itanium2/1600-L 3ML3	1	11.9	88
5	Intel Tiger Madison/1500	1	12.2	86
6	HP RX4640 Itanium2/1300-H	1	12.3	85
7	HP RX2600 Itanium2/1500-L	1	12.3	85
8	SGI Altix3700/Itanium2-1500	1	12.7	82
9	HP RX2600 Itanium2/1000-H	1	13.4	78
10	SGI Altix3700/Itanium2-1300	1	15.5	68
11	IBM pSeries 690Turbo/1700	1	19.1	55
12	HP DL145 Opteron248/2200 (pgi5.2)	1	20.0	52
13	AMD Opteron848/2200 (pgi5.1)	1	20.1	52
14	SunFire V20 Opteron248/2.2GHz (pgi5.2)	1	20.2	52
15	AMD Opteron848/2200 (pgi5.2)	1	20.3	51
16	SunFire V40z Opteron248/2.2GHz (pgi5.2)	1	20.4	51
17	IBM eServer p5 570/1.9 GHz	1	21.0	50
18	AMD Opteron848/2200 (pgi5.0)	1	21.6	48
19	Intel Tiger Madison/1200	1	21.8	48
20	AMD Opteron246/2000 (pgi5.0)	1	22.7	46
21	IBM pSeries 690Turbo/1300	1	24.9	42
22	Intel Tiger Itanium 2/1000	1	25.1	42
23	HP RX5670 Itanium2/1000-L	1	25.8	40
24	AMD Opteron244/1800 (pgi5.0)	1	25.9	40
25	SUN Blade 2000/1056-Cu	1	27.7	38
26	HP ZX6000 Itanium 2/900-L	1	29.0	36
27	AMD Opteron842/1600 (pgi5.0)	1	29.6	35
28	HP DL380 Pentium4/3600 (EM64T)	1	30.8	34
29	IBM pSeries 630/1000	1	32.6	32
30	AMD Opteron840/1400 (pgi5.0)	1	33.7	31
31	SM X6DAG-G Pentium4/3200 (EM64T)	1	34.8	30
32	SUN Fire V880/900-Cu	1	41.7	25
33	HP DL140 Pentium 4 Xeon/3200	1	43.8	24
34	HP PA-9000/N4000-440	1	44.7	23
35	Pentium 4 Xeon/3066 (ifc)	1	45.6	23
36	SGI Origin 3800/R14k-600	1	48.6	22
37	AMD MP2000+ / 1667	1	49.4	21
38	Compaq Alpha ES45/1250	1	50.9	21
39	HP BL20p G2 Pentium 4/2800	1	50.9	21
40	Pentium 4 Xeon/2666 (ifc)	1	51.3	20
41	SUN Fire 6800/750	1	55.0	19
42	SGI Origin 3800/R14k-500	1	58.7	18
43	Pentium 4/2000 (ifc)	1	78.3	13
44	API UP2000 6/667	1	169.4	6
	Number of CPUs = 2
1	HP RX5670 Itanium2/1500-H	2	10.5	199
2	HP Superdome Itanium2/1500	2	10.6	196
3	HP RX4640 Itanium2/1600-L 9ML3	2	11.8	177
4	HP RX1620 Itanium2/1600-L 3ML3	2	12.0	174
5	Intel Tiger Madison/1500	2	12.2	171
6	HP RX4640 Itanium2/1300-H	2	12.3	169
7	HP RX2600 Itanium2/1500-L	2	12.4	169
8	SGI Altix3700/Itanium2-1500	2	12.8	163
9	HP RX2600 Itanium2/1000-H	2	13.4	155
10	SGI Altix3700/Itanium2-1300	2	15.9	132
11	IBM pSeries 690Turbo/1700	2	19.1	109
12	HP DL145 Opteron248/2200 (pgi5.2)	2	19.9	105
13	SunFire V20 Opteron248/2.2GHz (pgi5.2)	2	20.0	105
14	AMD Opteron848/2200 (pgi5.2)	2	20.3	103
15	AMD Opteron848/2200 (pgi5.1)	2	20.7	101
16	SunFire V40z Opteron248/2.2GHz (pgi5.2)	2	20.8	101
17	IBM eServer p5 570/1.9 GHz	2	21.1	99
18	Intel Tiger Madison/1200	2	21.8	96
19	AMD Opteron848/2200 (pgi5.0)	2	22.4	93
20	AMD Opteron246/2000 (pgi5.0)	2	23.8	88
21	IBM pSeries 690Turbo/1300	2	24.9	84
22	Intel Tiger Itanium 2/1000	2	25.2	83
23	HP RX5670 Itanium2/1000-L	2	25.6	82
24	AMD Opteron244/1800 (pgi5.0)	2	26.1	80
25	SUN Blade 2000/1056-Cu	2	28.2	74
26	AMD Opteron842/1600 (pgi5.0)	2	29.6	70
27	HP DL380 Pentium4/3600 (EM64T)	2	30.8	68
28	IBM pSeries 630/1000	2	33.4	63
29	AMD Opteron840/1400 (pgi5.0)	2	33.7	62
30	SM X6DAG-G Pentium4/3200 (EM64T)	2	34.8	60
31	HP DL140 Pentium 4 Xeon/3200	2	43.7	48
32	HP PA-9000/N4000-440	2	45.0	46
33	Pentium 4 Xeon/3066 (ifc)	2	45.8	46
34	Compaq Alpha ES45/1250	2	50.6	41
35	AMD MP2000+ / 1667	2	50.7	41
36	HP BL20p G2 Pentium 4/2800	2	51.7	40
37	Pentium 4 Xeon/2666 (ifc)	2	51.9	40
38	SUN Fire 6800/750	2	55.3	38
39	SGI Origin 3800/R14k-500	2	58.8	36
40	Pentium 4/2000 (ifc)	2	80.8	26
41	API UP2000 6/667	2	169.8	12
	Number of CPUs = 4
1	HP Superdome Itanium2/1500	4	10.6	393
2	HP RX5670 Itanium2/1500-H	4	10.7	392
3	HP RX4640 Itanium2/1600-L 9ML3	4	11.8	354
4	Intel Tiger Madison/1500	4	12.3	340
5	HP RX4640 Itanium2/1300-H	4	12.5	335
6	SGI Altix3700/Itanium2-1500	4	13.0	323
7	SGI Altix3700/Itanium2-1300	4	15.4	272
8	IBM pSeries 690Turbo/1700	4	19.1	219
9	SunFire V40z Opteron248/2.2GHz (pgi5.2)	4	20.4	205
10	AMD Opteron848/2200 (pgi5.2)	4	20.4	205
11	IBM eServer p5 570/1.9 GHz	4	21.4	195
12	Intel Tiger Madison/1200	4	21.8	192
13	AMD Opteron848/2200 (pgi5.1)	4	22.2	189
14	AMD Opteron848/2200 (pgi5.0)	4	22.6	185
15	Intel Tiger Itanium 2/1000	4	25.2	166
16	HP RX5670 Itanium2/1000-L	4	25.6	163
17	IBM pSeries 690Turbo/1300	4	25.7	162
18	AMD Opteron842/1600 (pgi5.0)	4	29.6	141
19	IBM pSeries 630/1000	4	40.4	103
20	SUN Fire V880/900-Cu	4	41.7	100
21	HP PA-9000/N4000-440	4	45.8	91
22	Compaq Alpha ES45/1250	4	51.2	82
23	SGI Origin 3800/R14k-500	4	59.0	71
	Number of CPUs = 8
1	HP Superdome Itanium2/1500	8	10.7	783
2	SGI Altix3700/Itanium2-1500	8	12.6	662
3	SGI Altix3700/Itanium2-1300	8	16.5	507
4	IBM pSeries 690Turbo/1700	8	19.1	438
5	IBM pSeries 690Turbo/1300	8	26.0	322
6	SUN Fire V880/900-Cu	8	43.7	191
7	HP PA-9000/N4000-440	8	47.7	175

Table 18. The Matrix Multiplication RATE Benchmark 3. DGEMM.

The DGEMM Benchmark. Total Elapsed times (seconds) and RATE Estimate Relative to the HP RX5670 Itanium2/1500-H.

Rank	Machine	NCPUs	Wall Time	RATE
			(seconds)
1	IBM eServer p5 570/1.9 GHz	1	5.8	112
2	HP RX4640 Itanium2/1600-L 9ML3	1	6.2	103
3	HP RX1620 Itanium2/1600-L 3ML3	1	6.3	102
4	HP RX5670 Itanium2/1500-H	1	6.4	100
5	HP DL380 Pentium4/3600 (EM64T)	1	6.5	99
6	HP Superdome Itanium2/1500	1	6.5	99
7	HP RX2600 Itanium2/1500-L	1	6.7	97
8	SGI Altix3700/Itanium2-1500	1	6.7	96
9	SM X6DAG-G Pentium4/3200 (EM64T)	1	7.0	92
10	Intel Tiger Madison/1500	1	7.3	88
11	HP DL140 Pentium 4 Xeon/3200	1	7.4	87
12	HP RX4640 Itanium2/1300-H	1	7.4	87
13	HP BL20p G2 Pentium 4/2800	1	7.9	81
14	SGI Altix3700/Itanium2-1300	1	8.0	80
15	IBM pSeries 690Turbo/1700	1	8.2	78
16	Pentium 4 Xeon/3066 (ifc)	1	8.3	78
17	Intel Tiger Madison/1200	1	9.2	70
18	Pentium 4 Xeon/2666 (ifc)	1	9.5	68
19	HP RX2600 Itanium2/1000-H	1	9.6	67
20	HP DL145 Opteron248/2200 (pgi5.2)	1	10.1	64
21	SunFire V20 Opteron248/2.2GHz (pgi5.2)	1	10.1	64
22	AMD Opteron848/2200 (pgi5.2)	1	10.2	63
23	SunFire V40z Opteron248/2.2GHz (pgi5.2)	1	10.2	63
24	AMD Opteron848/2200 (pgi5.1)	1	10.5	61
25	IBM pSeries 690Turbo/1300	1	10.7	60
26	AMD Opteron848/2200 (pgi5.0)	1	10.8	59
27	Intel Tiger Itanium 2/1000	1	11.0	58
28	HP RX5670 Itanium2/1000-L	1	11.2	58
29	AMD Opteron246/2000 (pgi5.0)	1	11.5	56
30	HP ZX6000 Itanium 2/900-L	1	12.6	51
31	AMD Opteron244/1800 (pgi5.0)	1	13.4	48
32	Pentium 4/2000 (ifc)	1	14.0	46
33	IBM pSeries 630/1000	1	14.2	45
34	AMD Opteron842/1600 (pgi5.0)	1	15.3	42
35	Compaq Alpha ES45/1250	1	16.6	39
36	AMD Opteron840/1400 (pgi5.0)	1	17.4	37
37	SUN Blade 2000/1056-Cu	1	21.0	31
38	SUN Fire V880/900-Cu	1	22.8	28
39	AMD MP2000+ / 1667	1	28.3	23
40	HP PA-9000/N4000-440	1	31.8	20
41	API UP2000 6/667	1	32.7	20
42	SGI Origin 3800/R14k-600	1	34.1	19
43	SGI Origin 3800/R14k-500	1	40.3	16
44	SUN Fire 6800/750	1	62.4	10
	Number of CPUs = 2
1	IBM eServer p5 570/1.9 GHz	2	5.8	221
2	HP RX4640 Itanium2/1600-L 9ML3	2	6.3	205
3	HP RX5670 Itanium2/1500-H	2	6.4	200
4	HP RX1620 Itanium2/1600-L 3ML3	2	6.5	199
5	HP Superdome Itanium2/1500	2	6.5	198
6	HP DL380 Pentium4/3600 (EM64T)	2	6.6	194
7	HP RX2600 Itanium2/1500-L	2	6.7	191
8	SGI Altix3700/Itanium2-1500	2	6.8	191
9	SM X6DAG-G Pentium4/3200 (EM64T)	2	7.2	178
10	Intel Tiger Madison/1500	2	7.4	174
11	HP RX4640 Itanium2/1300-H	2	7.5	172
12	SGI Altix3700/Itanium2-1300	2	8.0	161
13	IBM pSeries 690Turbo/1700	2	8.2	156
14	HP DL140 Pentium 4 Xeon/3200	2	8.4	153
15	HP BL20p G2 Pentium 4/2800	2	8.8	145
16	Intel Tiger Madison/1200	2	9.3	139
17	Pentium 4 Xeon/3066 (ifc)	2	9.4	136
18	HP RX2600 Itanium2/1000-H	2	9.6	133
19	HP DL145 Opteron248/2200 (pgi5.2)	2	10.1	128
20	SunFire V20 Opteron248/2.2GHz (pgi5.2)	2	10.1	127
21	AMD Opteron848/2200 (pgi5.2)	2	10.2	126
22	SunFire V40z Opteron248/2.2GHz (pgi5.2)	2	10.3	125
23	Pentium 4 Xeon/2666 (ifc)	2	10.5	123
24	AMD Opteron848/2200 (pgi5.1)	2	10.8	120
25	IBM pSeries 690Turbo/1300	2	10.8	119
26	Intel Tiger Itanium 2/1000	2	10.8	119
27	AMD Opteron848/2200 (pgi5.0)	2	11.1	116
28	HP RX5670 Itanium2/1000-L	2	11.4	112
29	AMD Opteron246/2000 (pgi5.0)	2	11.9	108
30	AMD Opteron244/1800 (pgi5.0)	2	13.5	95
31	IBM pSeries 630/1000	2	14.8	87
32	Pentium 4/2000 (ifc)	2	15.3	84
33	AMD Opteron842/1600 (pgi5.0)	2	15.3	84
34	Compaq Alpha ES45/1250	2	16.6	78
35	AMD Opteron840/1400 (pgi5.0)	2	17.5	74
36	SUN Blade 2000/1056-Cu	2	21.6	60
37	AMD MP2000+ / 1667	2	30.0	43
38	HP PA-9000/N4000-440	2	32.2	40
39	API UP2000 6/667	2	37.9	34
40	SGI Origin 3800/R14k-500	2	40.4	32
41	SUN Fire 6800/750	2	62.8	20
	Number of CPUs = 4
1	IBM eServer p5 570/1.9 GHz	4	6.2	413
2	HP RX4640 Itanium2/1600-L 9ML3	4	6.3	409
3	HP RX5670 Itanium2/1500-H	4	6.5	395
4	HP Superdome Itanium2/1500	4	6.5	393
5	SGI Altix3700/Itanium2-1500	4	6.8	378
6	Intel Tiger Madison/1500	4	7.5	344
7	HP RX4640 Itanium2/1300-H	4	7.6	339
8	SGI Altix3700/Itanium2-1300	4	8.0	323
9	IBM pSeries 690Turbo/1700	4	8.2	313
10	Intel Tiger Madison/1200	4	9.5	272
11	SunFire V40z Opteron248/2.2GHz (pgi5.2)	4	10.2	251
12	AMD Opteron848/2200 (pgi5.2)	4	10.2	251
13	Intel Tiger Itanium 2/1000	4	10.9	235
14	AMD Opteron848/2200 (pgi5.0)	4	11.2	230
15	IBM pSeries 690Turbo/1300	4	11.2	229
16	AMD Opteron848/2200 (pgi5.1)	4	11.4	225
17	HP RX5670 Itanium2/1000-L	4	11.7	219
18	AMD Opteron842/1600 (pgi5.0)	4	15.3	168
19	Compaq Alpha ES45/1250	4	16.6	154
20	IBM pSeries 630/1000	4	20.3	127
21	SUN Fire V880/900-Cu	4	22.8	113
22	HP PA-9000/N4000-440	4	33.1	78
23	SGI Origin 3800/R14k-500	4	40.4	64
	Number of CPUs = 8
1	HP Superdome Itanium2/1500	8	6.6	776
2	SGI Altix3700/Itanium2-1500	8	6.8	755
3	IBM pSeries 690Turbo/1700	8	8.2	624
4	SGI Altix3700/Itanium2-1300	8	8.7	593
5	IBM pSeries 690Turbo/1300	8	11.7	440
6	SUN Fire V880/900-Cu	8	23.7	217
7	HP PA-9000/N4000-440	8	35.4	145

Table 19. The Matrix Multiplication RATE Benchmark 4. QHQ (Vector Algorithm).

The QHQ(vector) RATE Benchmark. Total Elapsed times (seconds) and RATE Estimate Relative to the HP RX5670 Itanium2/1500-H.

Rank	Machine	NCPUs	Wall Time	RATE
			(seconds)
1	HP RX5670 Itanium2/1500-H	1	52.1	100
2	Intel Tiger Madison/1500	1	55.1	95
3	HP Superdome Itanium2/1500	1	58.1	90
4	HP RX4640 Itanium2/1600-L 9ML3	1	58.1	90
5	SGI Altix3700/Itanium2-1500	1	63.4	82
6	IBM eServer p5 570/1.9 GHz	1	66.1	79
7	HP RX1620 Itanium2/1600-L 3ML3	1	66.4	78
8	HP RX4640 Itanium2/1300-H	1	79.1	66
9	HP RX2600 Itanium2/1500-L	1	82.2	63
10	Intel Tiger Madison/1200	1	86.8	60
11	HP RX2600 Itanium2/1000-H	1	92.7	56
12	SGI Altix3700/Itanium2-1300	1	93.7	56
13	Intel Tiger Itanium 2/1000	1	95.8	54
14	IBM pSeries 690Turbo/1700	1	103.5	50
15	HP DL380 Pentium4/3600 (EM64T)	1	116.0	45
16	IBM pSeries 690Turbo/1300	1	124.3	42
17	SM X6DAG-G Pentium4/3200 (EM64T)	1	130.9	40
18	HP RX5670 Itanium2/1000-L	1	137.6	38
19	Compaq Alpha ES45/1250	1	153.1	34
20	SunFire V20 Opteron248/2.2GHz (pgi5.2)	1	157.5	33
21	HP DL145 Opteron248/2200 (pgi5.2)	1	159.0	33
22	Pentium 4 Xeon/3066 (ifc)	1	164.4	32
23	IBM pSeries 630/1000	1	168.3	31
24	AMD Opteron246/2000 (pgi5.0)	1	172.1	30
25	SunFire V40z Opteron248/2.2GHz (pgi5.2)	1	172.2	30
26	AMD Opteron848/2200 (pgi5.2)	1	176.7	29
27	SUN Blade 2000/1056-Cu	1	180.1	29
28	Pentium 4 Xeon/2666 (ifc)	1	182.3	29
29	AMD Opteron848/2200 (pgi5.1)	1	182.6	29
30	SUN Fire V880/900-Cu	1	184.4	28
31	HP BL20p G2 Pentium 4/2800	1	199.1	26
32	AMD Opteron244/1800 (pgi5.0)	1	201.3	26
33	AMD Opteron848/2200 (pgi5.0)	1	202.2	26
34	HP DL140 Pentium 4 Xeon/3200	1	214.2	24
35	AMD Opteron842/1600 (pgi5.0)	1	224.9	23
36	AMD Opteron840/1400 (pgi5.0)	1	242.9	21
37	HP ZX6000 Itanium 2/900-L	1	290.6	18
38	Pentium 4/2000 (ifc)	1	301.4	17
39	SGI Origin 3800/R14k-600	1	362.2	14
40	SGI Origin 3800/R14k-500	1	386.5	13
41	HP PA-9000/N4000-440	1	453.0	11
42	API UP2000 6/667	1	522.5	10
43	AMD MP2000+ / 1667	1	533.3	10
44	SUN Fire 6800/750	1	642.7	8
	Number of CPUs = 2
1	HP RX5670 Itanium2/1500-H	2	53.6	194
2	Intel Tiger Madison/1500	2	57.3	182
3	HP Superdome Itanium2/1500	2	57.8	180
4	HP RX4640 Itanium2/1600-L 9ML3	2	60.7	172
5	SGI Altix3700/Itanium2-1500	2	62.5	167
6	IBM eServer p5 570/1.9 GHz	2	66.1	158
7	HP RX1620 Itanium2/1600-L 3ML3	2	77.8	134
8	HP RX4640 Itanium2/1300-H	2	93.4	112
9	Intel Tiger Madison/1200	2	99.2	105
10	HP RX2600 Itanium2/1000-H	2	101.2	103
11	Intel Tiger Itanium 2/1000	2	103.1	101
12	IBM pSeries 690Turbo/1700	2	103.9	100
13	HP RX2600 Itanium2/1500-L	2	109.1	95
14	SGI Altix3700/Itanium2-1300	2	123.3	84
15	IBM pSeries 690Turbo/1300	2	129.3	81
16	HP RX5670 Itanium2/1000-L	2	152.9	68
17	Compaq Alpha ES45/1250	2	153.4	68
18	HP DL145 Opteron248/2200 (pgi5.2)	2	158.6	66
19	SunFire V40z Opteron248/2.2GHz (pgi5.2)	2	171.6	61
20	SUN Blade 2000/1056-Cu	2	182.7	57
21	AMD Opteron848/2200 (pgi5.2)	2	183.9	57
22	SunFire V20 Opteron248/2.2GHz (pgi5.2)	2	184.7	56
23	IBM pSeries 630/1000	2	208.4	50
24	AMD Opteron244/1800 (pgi5.0)	2	218.3	48
25	AMD Opteron848/2200 (pgi5.0)	2	227.9	46
26	HP DL380 Pentium4/3600 (EM64T)	2	228.0	46
27	AMD Opteron848/2200 (pgi5.1)	2	230.4	45
28	AMD Opteron246/2000 (pgi5.0)	2	240.5	43
29	AMD Opteron842/1600 (pgi5.0)	2	253.8	41
30	SM X6DAG-G Pentium4/3200 (EM64T)	2	277.7	38
31	AMD Opteron840/1400 (pgi5.0)	2	280.3	37
32	HP BL20p G2 Pentium 4/2800	2	325.2	32
33	Pentium 4 Xeon/2666 (ifc)	2	340.4	31
34	Pentium 4 Xeon/3066 (ifc)	2	347.7	30
35	HP DL140 Pentium 4 Xeon/3200	2	373.3	28
36	SGI Origin 3800/R14k-500	2	399.2	26
37	Pentium 4/2000 (ifc)	2	466.6	22
38	HP PA-9000/N4000-440	2	498.6	21
39	API UP2000 6/667	2	603.0	17
40	SUN Fire 6800/750	2	644.5	16
41	AMD MP2000+ / 1667	2	723.3	14
	Number of CPUs = 4
1	HP Superdome Itanium2/1500	4	58.1	359
2	SGI Altix3700/Itanium2-1500	4	61.4	339
3	HP RX5670 Itanium2/1500-H	4	61.7	338
4	Intel Tiger Madison/1500	4	63.9	326
5	HP RX4640 Itanium2/1600-L 9ML3	4	79.8	261
6	IBM eServer p5 570/1.9 GHz	4	85.6	243
7	IBM pSeries 690Turbo/1700	4	105.1	198
8	SGI Altix3700/Itanium2-1300	4	124.5	167
9	Intel Tiger Madison/1200	4	132.8	157
10	Intel Tiger Itanium 2/1000	4	138.4	150
11	HP RX4640 Itanium2/1300-H	4	146.3	142
12	IBM pSeries 690Turbo/1300	4	155.2	134
13	Compaq Alpha ES45/1250	4	155.3	134
14	SunFire V40z Opteron248/2.2GHz (pgi5.2)	4	175.2	119
15	HP RX5670 Itanium2/1000-L	4	185.0	113
16	SUN Fire V880/900-Cu	4	186.3	112
17	AMD Opteron848/2200 (pgi5.2)	4	204.3	102
18	AMD Opteron842/1600 (pgi5.0)	4	272.1	77
19	IBM pSeries 630/1000	4	283.6	73
20	AMD Opteron848/2200 (pgi5.0)	4	296.5	70
21	AMD Opteron848/2200 (pgi5.1)	4	321.6	65
22	SGI Origin 3800/R14k-500	4	387.3	54
23	HP PA-9000/N4000-440	4	616.3	34
	Number of CPUs = 8
1	HP Superdome Itanium2/1500	8	72.5	575
2	SGI Altix3700/Itanium2-1500	8	75.7	550
3	IBM pSeries 690Turbo/1700	8	107.3	388
4	SGI Altix3700/Itanium2-1300	8	131.1	318
5	IBM pSeries 690Turbo/1300	8	187.9	222
6	SUN Fire V880/900-Cu	8	189.6	220
7	HP PA-9000/N4000-440	8	1001.5	42

Table 20. The Matrix Multiplication RATE Benchmark 5. QHQ (Scalar Alogorithm).

The QHQ (Scalar) RATE Benchmark. Total Elapsed times (seconds) and RATE Estimate Relative to the HP RX5670 Itanium2/1500-H.

Rank	Machine	NCPUs	Wall Time	RATE
			(seconds)
1	HP RX5670 Itanium2/1500-H	1	90.6	100
2	HP Superdome Itanium2/1500	1	112.9	80
3	HP RX4640 Itanium2/1300-H	1	118.6	76
4	IBM eServer p5 570/1.9 GHz	1	131.9	69
5	HP RX4640 Itanium2/1600-L 9ML3	1	144.2	63
6	HP RX2600 Itanium2/1500-L	1	153.7	59
7	Intel Tiger Madison/1500	1	166.1	55
8	HP RX2600 Itanium2/1000-H	1	168.8	54
9	SGI Altix3700/Itanium2-1500	1	185.4	49
10	HP RX1620 Itanium2/1600-L 3ML3	1	190.5	48
11	SGI Altix3700/Itanium2-1300	1	203.5	44
12	Compaq Alpha ES45/1250	1	212.2	43
13	IBM pSeries 690Turbo/1700	1	214.0	42
14	Intel Tiger Madison/1200	1	225.7	40
15	HP DL380 Pentium4/3600 (EM64T)	1	227.9	40
16	IBM pSeries 690Turbo/1300	1	246.9	37
17	SM X6DAG-G Pentium4/3200 (EM64T)	1	251.6	36
18	Intel Tiger Itanium 2/1000	1	255.7	35
19	HP DL145 Opteron248/2200 (pgi5.2)	1	276.8	33
20	SunFire V20 Opteron248/2.2GHz (pgi5.2)	1	281.7	32
21	HP BL20p G2 Pentium 4/2800	1	307.3	29
22	SUN Blade 2000/1056-Cu	1	318.5	28
23	Pentium 4 Xeon/3066 (ifc)	1	332.4	27
24	IBM pSeries 630/1000	1	335.4	27
25	SUN Fire V880/900-Cu	1	335.8	27
26	AMD Opteron246/2000 (pgi5.0)	1	337.1	27
27	SunFire V40z Opteron248/2.2GHz (pgi5.2)	1	346.4	26
28	Pentium 4 Xeon/2666 (ifc)	1	347.8	26
29	HP DL140 Pentium 4 Xeon/3200	1	349.6	26
30	HP RX5670 Itanium2/1000-L	1	365.3	25
31	AMD Opteron848/2200 (pgi5.2)	1	368.6	25
32	AMD Opteron848/2200 (pgi5.1)	1	380.7	24
33	AMD Opteron848/2200 (pgi5.0)	1	427.1	21
34	HP ZX6000 Itanium 2/900-L	1	430.9	21
35	AMD Opteron244/1800 (pgi5.0)	1	500.3	18
36	Pentium 4/2000 (ifc)	1	518.0	17
37	AMD Opteron842/1600 (pgi5.0)	1	657.7	14
38	AMD Opteron840/1400 (pgi5.0)	1	665.7	14
39	API UP2000 6/667	1	683.8	13
40	SGI Origin 3800/R14k-500	1	724.9	12
41	HP PA-9000/N4000-440	1	846.2	11
42	SGI Origin 3800/R14k-600	1	1046.2	9
43	AMD MP2000+ / 1667	1	1339.9	7
44	SUN Fire 6800/750	1	1664.2	5
	Number of CPUs = 2
1	HP RX5670 Itanium2/1500-H	2	104.7	173
2	HP Superdome Itanium2/1500	2	113.0	160
3	IBM eServer p5 570/1.9 GHz	2	133.2	136
4	HP RX4640 Itanium2/1600-L 9ML3	2	148.2	122
5	HP RX2600 Itanium2/1500-L	2	163.2	111
6	HP RX4640 Itanium2/1300-H	2	165.2	110
7	Intel Tiger Madison/1500	2	175.5	103
8	SGI Altix3700/Itanium2-1500	2	187.9	96
9	HP RX2600 Itanium2/1000-H	2	195.7	93
10	IBM pSeries 690Turbo/1700	2	214.1	85
11	Compaq Alpha ES45/1250	2	223.6	81
12	HP RX1620 Itanium2/1600-L 3ML3	2	225.6	80
13	SGI Altix3700/Itanium2-1300	2	227.4	80
14	Intel Tiger Madison/1200	2	250.9	72
15	IBM pSeries 690Turbo/1300	2	255.6	71
16	Intel Tiger Itanium 2/1000	2	275.0	66
17	HP DL145 Opteron248/2200 (pgi5.2)	2	290.5	62
18	HP DL380 Pentium4/3600 (EM64T)	2	316.5	57
19	SUN Blade 2000/1056-Cu	2	325.2	56
20	SunFire V20 Opteron248/2.2GHz (pgi5.2)	2	328.6	55
21	SM X6DAG-G Pentium4/3200 (EM64T)	2	332.6	54
22	SunFire V40z Opteron248/2.2GHz (pgi5.2)	2	346.9	52
23	AMD Opteron848/2200 (pgi5.2)	2	382.2	47
24	IBM pSeries 630/1000	2	387.0	47
25	HP RX5670 Itanium2/1000-L	2	387.9	47
26	Pentium 4 Xeon/3066 (ifc)	2	412.2	44
27	HP BL20p G2 Pentium 4/2800	2	416.5	43
28	Pentium 4 Xeon/2666 (ifc)	2	425.6	43
29	AMD Opteron848/2200 (pgi5.1)	2	448.8	40
30	AMD Opteron848/2200 (pgi5.0)	2	487.0	37
31	AMD Opteron246/2000 (pgi5.0)	2	496.1	37
32	AMD Opteron244/1800 (pgi5.0)	2	538.7	34
33	Pentium 4/2000 (ifc)	2	573.6	32
34	HP DL140 Pentium 4 Xeon/3200	2	642.0	28
35	AMD Opteron842/1600 (pgi5.0)	2	719.8	25
36	AMD Opteron840/1400 (pgi5.0)	2	804.1	23
37	SGI Origin 3800/R14k-500	2	878.0	21
38	API UP2000 6/667	2	880.0	21
39	HP PA-9000/N4000-440	2	932.1	19
40	AMD MP2000+ / 1667	2	1401.7	13
41	SUN Fire 6800/750	2	1694.1	11
	Number of CPUs = 4
1	HP Superdome Itanium2/1500	4	113.8	318
2	IBM eServer p5 570/1.9 GHz	4	146.4	247
3	HP RX4640 Itanium2/1600-L 9ML3	4	155.1	234
4	HP RX5670 Itanium2/1500-H	4	156.8	231
5	SGI Altix3700/Itanium2-1500	4	189.9	191
6	Intel Tiger Madison/1500	4	209.7	173
7	IBM pSeries 690Turbo/1700	4	215.8	168
8	Compaq Alpha ES45/1250	4	225.7	161
9	SGI Altix3700/Itanium2-1300	4	226.8	160
10	IBM pSeries 690Turbo/1300	4	278.6	130
11	HP RX4640 Itanium2/1300-H	4	298.3	121
12	SUN Fire V880/900-Cu	4	341.6	106
13	Intel Tiger Madison/1200	4	356.0	102
14	Intel Tiger Itanium 2/1000	4	361.2	100
15	SunFire V40z Opteron248/2.2GHz (pgi5.2)	4	361.3	100
16	AMD Opteron848/2200 (pgi5.2)	4	424.1	85
17	HP RX5670 Itanium2/1000-L	4	452.2	80
18	IBM pSeries 630/1000	4	473.8	76
19	AMD Opteron848/2200 (pgi5.0)	4	626.5	58
20	AMD Opteron848/2200 (pgi5.1)	4	639.4	57
21	SGI Origin 3800/R14k-500	4	726.9	50
22	AMD Opteron842/1600 (pgi5.0)	4	751.7	48
23	HP PA-9000/N4000-440	4	1160.5	31
	Number of CPUs = 8
1	HP Superdome Itanium2/1500	8	159.8	453
2	SGI Altix3700/Itanium2-1500	8	207.7	349
3	IBM pSeries 690Turbo/1700	8	221.1	328
4	SGI Altix3700/Itanium2-1300	8	225.2	322
5	IBM pSeries 690Turbo/1300	8	312.9	232
6	SUN Fire V880/900-Cu	8	352.7	205
7	HP PA-9000/N4000-440	8	1919.8	38

Table 21. The Matrix Diagonalisation RATE Benchmark.

The Matrix Diagonalisation Benchmark. Total Elapsed times (seconds) and RATE Estimate Relative to the HP RX5670 Itanium2/1500-H.

Rank	Machine	NCPUs	Wall Time	RATE
			(seconds)
1	HP RX5670 Itanium2/1500-H