Application Performance on High-End and Commodity-class Computers

8/18/2003


Click here to start


Table of Contents

Application Performance on High-End and Commodity-class Computers

Outline

Capability Computing - Cost Effectiveness & Dependencies

High-end Commodity-based Systems

PNNLís HPCS2 - 1400 Processor HP Supercluster

High-End Systems Evaluated

Technical Progress in 2003

Commodity Systems (CSx) Prototype / Evaluation Hardware

Commodity Systems (CSx) II. Evaluation Hardware

Applications Performance Overview

SPEC CPU 2000 - SPECfp2000 Values relative to IBM p-series 690/pwr4 1.3 GHz

The GAMESS-UK Serial Benchmark

STREAM: Measured Sustainable Memory Bandwidth in HPC (TRIAD)

Communications Benchmark PMB and EFF_BW

Collective Operations (Time - usec) - as function of no.of CPUs

Communications Benchmark PMB: Pallas MPI Benchmark Suite (V2.2)

PingPong Performance

MPI_allreduce Performance

Interconnect Benchmark - EFF_BW

Interconnect Benchmark - Latency

Application Codes

Performance Metrics: 1999-2001

Beowulf Comparisons with the T3E & O3800/R14k-500

Commodity Comparisons with High-end Systems: Compaq AlphaServer SC ES45/1000 and the SGI O3800/R14k-500

Performance Metrics: 2003

Molecular Simulation

DL_POLY: A Parallel MD Package

DL_POLY Parallel Benchmarks (Cray T3E/1200)

DL_POLY V2: High-end and Commodity-based Systems

DL_POLY V2: High-end and Commodity-based Systems I.

DL_POLY V2: High-end and Commodity-based Systems II.

DL_POLY V2: High-end and Commodity-based Systems III.

DL_POLY V2: Replicated Data

DLMULTI

DLMULTI: High-end and Commodity-based Systems

CHARMM

Parallel CHARMM Benchmark

Parallel CHARMM Benchmark: LAM MPI vs. MPICH

Migration from Replicated to Distributed data DL_POLY-3: Coulomb Energy Evaluation

Migration from Replicated to Distributed data DL_POLY-3: Coulomb Energy Performance

DL_POLY3 Macromolecular Simulations

DL_POLY3 Commodity Performance

Molecular Electronic Structure

Distributed Data SCF

High-End Computational Chemistry: The NWChem Software

Global Arrays

PeIGS 3.0 Parallel Performance

Scalability of Numerical Algorithms I.

Scalability of Numerical Algorithms II.

Parallel Eigensolvers

Case Studies - Zeolite Fragments

DFT Coulomb Fit - NWChem

DFT Coulomb Fit - NWChem

Memory-driven Approaches: NWChem - DFT (LDA): Performance on the IBM SP/p690

2.2 GAMESS-UK

GAMESS-UK features 1.

GAMESS-UK features 2.

Parallel Implementations of GAMESS-UK

GAMESS-UK ?SCF Performance IBM SP/p690, High-end and Commodity-based Systems

GAMESS-UK. DFT B3LYP Performance IBM SP/p690, High-end and Commodity-based Systems

GAMESS-UK. DFT B3LYP Performance The IBM SP/p690 and High-end Systems

GAMESS-UK. DFT HCTH Performance The IBM SP/p690 and High-end Systems

Auxilliary Basis Coulomb Fit (I)

Auxilliary Basis Coulomb Fit (ii)

GAMESS-UK: DFT HCTH on Valinomycin. Impact of Coulomb Fitting: IBM SP/p690, Compaq AlphaServer SC/1000 and SGI Origin R14k/500

GAMESS-UK: DFT HCTH on Valinomycin. Impact of Coulomb Fitting

GAMESS-UK: DFT HCTH on Valinomycin. Impact of Coulomb Fitting

GAMESS-UK. DFT HCTH Performance: Impact of Coulomb Fitting

Memory-driven Approaches - SCF and DFT

DFT BLYP Gradient: High-end and Commodity-based Systems

MP2 Gradient Algorithms

Performance of MP2 Gradient Module Cray T3E/1200, High-end and Commodity- based Systems

SCF and DFT Analytic 2nd Derivatives

SCF Analytic 2nd Derivatives Performance IBM SP/p690, High-end and Commodity-based Systems

DFT Analytic 2nd Derivatives Performance IBM SP/p690, HP/Compaq SC ES45/1000 and SGI Origin 3800

Modelling Complex Systems: The QM/MM Approach

QM/MM Modelling - Challenges

GAMESS-UK Version 6.3 QM/MM Interface with CHARMM

Quantum Simulation in Industry (QUASI)

QUASI Partners and QM/MM Developments

QM/MM Applications

Performance Analysis of GA-based Applications using Vampir

GAMESS-UK / Si8O25H18 : 8 CPUs: One DFT Cycle

GAMESS-UK / Si8O25H18 : 8 CPUs

Materials Simulation Codes

Materials Simulation. Plane Wave Methods

CPMD 3.5 - Car-Parrinello Molecular Dynamics

CPMD 3.7 - C120 Benchmark

CPMD 3.7 - C120 Benchmark

Computational Engineering

ANGUS: Combustion modelling (regular grid) Cray T3E/1200, High-end and Commodity-based Systems

ANGUS: Combustion modelling (regular grid) Memory Bandwidth Effects: Pentium Xeon and Alpha Linux Systems

ANGUS: Combustion modelling (regular grid) High-end and Commodity-based Systems

Computation Engineering: UK Turbulence Consortium

Parallel Implementation and DNS Benchmark

DNS: 3603 benchmark. High End Systems

DNS: 1203 benchmark. High-end and Commodity-based Systems

DNS: 1203 benchmark. High-end and Commodity-based Systems

Commodity Comparisons with High-end Systems: Compaq AlphaServer SC ES45/1000 and the IBM SP/p690

Commodity Comparisons with High-end Systems: The Compaq AlphaServer SC ES45/1000

Summary

Author: Martyn F. Guest

Email: m.f.guest@dl.ac.uk