Computational Chemistry Benchmarks - Nvidia

More documents

Recommendations

Info

Viz, ―Docking‖ and Related Applications Growing Related Applications Features Supported GPU Perf Release Status Notes Amira 5® 3D visualization of volumetric data and surfaces 70x Released, Version 5.3.3 Single GPU Visualization from Visage Imaging. Next release, 5.4, will use GPU for general purpose processing in some functions http://www.visageimaging.com/overview.html BINDSURF Allows fast processing of large ligand databases 100X Available upon request to authors; single GPU High-Throughput parallel blind Virtual Screening, http://www.biomedcentral.com/1471-2105/13/S14/S13 BUDE Empirical Free Energy Forcefield 6.5-13.4X Released Single GPU University of Bristol http://www.bris.ac.uk/biochemistry/cpfg/bude/bude.htm Core Hopping GPU accelerated application 3.75-5000X Released, Suite 2011 Single and multi-GPUs. Schrodinger, Inc. http://www.schrodinger.com/products/14/32/ FastROCS Real-time shape similarity searching/comparison 800-3000X Released Single and multi-GPUs. Open Eyes Scientific Software http://www.eyesopen.com/fastrocs PyMol Lines: 460% increase Cartoons: 1246% increase Surface: 1746% increase Spheres: 753% increase Ribbon: 426% increase 1700x Released, Version 1.5 Single GPUs http://pymol.org/ VMD High quality rendering, large structures (100 million atoms), analysis and visualization tasks, multiple GPU support for display of molecular orbitals 100-125X or greater on kernels Released, Version 1.9 Visualization from University of Illinois at Urbana-Champaign http://www.ks.uiuc.edu/Research/vmd/ GPU Perf compared against Multi-core x86 CPU socket. GPU Perf benchmarked on GPU supported features and may be a kernel to kernel perf comparison
Bioinformatics Applications Application Features Supported GPU Speedup Release Status Website BarraCUDA Alignment of short sequencing reads 6-10x Version 0.6.2 – 3/2012 Multi-GPU, multi-node http://seqbarracuda.sourceforge.net/ CUDASW++ Parallel search of Smith- Waterman database 10-50x Version 2.0.8 – Q1/2012 Multi-GPU, multi-node http://sourceforge.net/projects/cudasw/ CUSHAW Parallel, accurate long read aligner for large genomes 10x Version 1.0.40 – 6/2012 Multiple-GPU http://cushaw.sourceforge.net/ GPU-BLAST Protein alignment according to BLASTP 3-4x Version 2.2.26 – 3/2012 Single GPU http://eudoxus.cheme.cmu.edu/gpublast/gpu blast.html GPU-HMMER Parallel local and global search of Hidden Markov Models 60-100x Version 2.3.2 – Q1/2012 Multi-GPU, multi-node http://www.mpihmmer.org/installguideGPUH MMER.htm mCUDA-MEME Scalable motif discovery algorithm based on MEME 4-10x Version 3.0.12 Multi-GPU, multi-node https://sites.google.com/site/yongchaosoftwa re/mcuda-meme SeqNFind Hardware and software for reference assembly, blast, SW, HMM, de novo assembly 400x Released. Multi-GPU, multi-node http://www.seqnfind.com/ UGENE Fast short read alignment 6-8x WideLM Parallel linear regression on multiple similarly-shaped models 150x Version 1.11 – 5/2012 Multi-GPU, multi-node Version 0.1-1 – 3/2012 Multi-GPU, multi-node http://ugene.unipro.ru/ http://insilicos.com/products/widelm GPU Perf compared against same or similar code running on single CPU machine Performance measured internally or independently
Page 1 and 2: update Updated: February 4, 2013
Page 3 and 4: New/Additional MD Applications Ramp
Page 5 and 6: Quantum Chemistry Applications Appl
Page 7: Quantum Chemistry Applications Appl
Page 12 and 13: Molecular Dynamics (MD) Application
Page 14 and 15: Built from Ground Up for GPUs Compu
Page 16 and 17: Nanoseconds / Day Kepler - Our Fast
Page 18 and 19: Speedup Compared to CPU Only K20 Ac
Page 20 and 21: Nanoseconds / Day K10 Strong Scalin
Page 22 and 23: Nanoseconds / Day K10 Extreme Perfo
Page 24 and 25: Replace 8 Nodes with 1 K20 GPU 90.0
Page 26 and 27: Nanoseconds / Day 1 CPU 2 GPUs 2 CP
Page 28 and 29: Recommended GPU Node Configuration
Page 30 and 31: NAMD 2.9
Page 32 and 33: Speedup Compared to CPU Only Accele
Page 34 and 35: Nanoseconds / Day Outstanding Stron
Page 36 and 37: Energy Expended (kJ) K20 - Greener:
Page 38 and 39: Recommended GPU Node Configuration
Page 40 and 41: LAMMPS, Jan. 2013 or later
Page 42 and 43: Speedup Relative to CPU Alone K20X,
Page 44 and 45: Loop Time (seconds) Excellent Stron
Page 46 and 47: Energy Expended (kJ) 140 Faster, Gr
Page 48 and 49: Time (s) Time (s) Time (s) Time (s)
Page 50 and 51: Early Titan XK6/XK7 Benchmarks 18 1
Page 52 and 53: GROMACS 4.6 Final, Pre-Beta and 4.6
Page 54 and 55: Nanoseconds / Day Great Scaling in
Page 56 and 57: Replace 3 Nodes with 2 GPUs 9 8 7 6
Page 58 and 59:
The Power of Kepler 140 RNase Solva
Page 60 and 61:
K20X, the Fastest Yet Nanoseconds /
Page 62 and 63:
CHARMM Release C37b1
Page 64 and 65:
More Bang for your Buck Scaled Perf
Page 66 and 67:
www.acellera.com 470 ns/day on 1 GP
Page 69 and 70:
Quantum Chemistry Applications Appl
Page 71 and 72:
Quantum Chemistry Applications Appl
Page 73:
BigDFT
Page 78 and 79:
Performance Relative to CPU Only Ke
Page 80 and 81:
Energy Expended (kJ) Kepler, keepin
Page 82 and 83:
Gaussian Key quantum chemistry code
Page 84 and 85:
GAMESS Partnership Overview Mark Go
Page 86 and 87:
Upcoming GAMESS Q4 2012 Release Mul
Page 88 and 89:
GPAW
Page 98 and 99:
NWChem
Page 100 and 101:
Quantum Espresso/PWscf
Page 102 and 103:
Scaled to CPU Only Extreme Performa
Page 104 and 105:
Elapsed Time (minutes) Replace 72 C
Page 106 and 107:
Power Consumption [kW/h] NVIDIA GPU
Page 108 and 109:
Time (s) QE/PWscf - More Powerful S
Page 110 and 111:
Time (Seconds) TeraChem Supercomput
Page 112 and 113:
Seconds Seconds Kepler’s Even Bet
Page 114 and 115:
FastROCS OpenEye Japan Hideyuki Sat
Page 116 and 117:
Shape Overlays per Second Riding Mo
Page 118 and 119:
Benefits of GPU Accelerated Computi
show all

Computational Chemistry Benchmarks - Nvidia

You also want an ePaper? Increase the reach of your titles

Delete template?

Save as template?