Simon McIntosh-Smith, University of Bristol, 'Accelerating molecular docking on Graphics Processors'
Upcoming SlideShare
Loading in...5
×
 

Simon McIntosh-Smith, University of Bristol, 'Accelerating molecular docking on Graphics Processors'

on

  • 1,705 views

 

Statistics

Views

Total Views
1,705
Views on SlideShare
1,528
Embed Views
177

Actions

Likes
0
Downloads
8
Comments
0

2 Embeds 177

http://www.cresset-group.com 173
http://a0.twimg.com 4

Accessibility

Categories

Upload Details

Uploaded via as Adobe PDF

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment

    Simon McIntosh-Smith, University of Bristol, 'Accelerating molecular docking on Graphics Processors' Simon McIntosh-Smith, University of Bristol, 'Accelerating molecular docking on Graphics Processors' Presentation Transcript

    • Accelerating moleculardocking on GraphicsProcessorsSimon McIntosh-SmithRichard B. SessionsUniversity of Bristol, UKsimonm@cs.bris.ac.uk, R.Sessions@bristol.ac.uk 1
    • !  Drug docking examples: Elastase inhibitors 2
    • !  Prion disease Prion protein behind Creutzfeld- Jacob disease in humans and shown here binding with a (pink) porphyrin-based ligand The porphyrins bound iron ion is just showing in yellow 1,719 atoms in the protein 53 atoms in the ligand 3
    • !  BUDE: Bristol University Docking Engine Accuracy Speed Typical docking Empirical Free Free Energy scoring Energy Forcefield calculations functions BUDE MM1,2 QM/MM3Entropy: solvation No Yes Yes configurational Approx Approx YesElectrostatics ? Approx YesAll atom No Yes YesExplicit solvent No No Yes 1. MD Tyka, AR Clarke, RB Sessions, J. Phys. Chem. B 110 17212-20 (2006) 2. MD Tyka, RB Sessions, AR Clarke, J. Phys. Chem. B 111 9571-80 (2007) 3. CJ Woods, FR Manby, AJ Mulholland, J. Chem. Phys. 128 014109 (2008) 4
    • !  Empirical Free Energy Function (atom-atom)ΔGligand binding = Nprotein Nligandi=1 j=1 f(xi,xj) Parameterised using experimental data N. Gibbs, A.R. Clarke & R.B. Sessions, "Ab-initio Protein Folding using Physicochemical Potentials and a Simplified Off-Lattice Model", Proteins 43:186-202,2001 5
    • !  Redocking into Xray Structure 1CIL (Human carbonic anhydrase II) 6
    • !  Another example 1EZQ (Human Factor XA) 7
    • 14 4 0 2 6 8 10 12 16 18 20 RMSD 0 5 10 15 20 3i5z 3b65 1kv2 3g0w 3d7z 1hfs 3e92 3e92 3b65 2fzz 3g0w 1z95 2cji 1mq6 1z95 3i5z 1hfs 2hvc 2nw4 1mq5 2p16 2nw4 2fzz 3d83 1w82 3d7z 1mq6 1kv2 2hvc 1ezq 3d83 2p16 2yxj 2cji 3e5a 2i0d 1mq5 3e93 2i0d 1hfc 1fjs 2yxj 1nfx 3ibu 1hfc 1w82 2oaz 2qo8 3ibu 2is7 3e93 1nfx 1nfw 1fjs 1nfw Best pose (by binding energy) 1ezq 1cil 3e5a 2qo8 2adu 2is7 1gcz 2pvh 3ibn 2rbe 1cil 2adu 3byz !  Self docking results 3ibn 2hoc 3byz 1i91 1gcz 2oaz 2gv7 1s2q 3i6o 2rbe 2hb3 3bzu 1mnc Best pose in top 100 poses (by binding energy) 1s2q 3bzu 1usn 3g9l 2pvh 1jg0 2gv7 2hoc 1jg0 1i91 1lqd 1mnc 1tsl 1usn 3g9l 1tsl 2zb0 1lqd 3i6o 2bq6 2hb3 2zb0 2bq6 (Ligand Binding Database)8
    • !   What are GPUs and OpenCL all about? CPU CPUA modern computer includes: – One or more CPUs – One or more GPUs GPU GPU OpenCL (Open Compute Language) lets programmers write a single portable program that uses ALL resources in the heterogeneous platform 9
    • !  BUDE Acceleration with OpenCL START Copy protein and ligand (input) coordinates (once) Geometry GA – like, ( ) (transform ligand) Rx Rxy Rxz Tx energy Ryx Ry Ryz Ty minimisation Rzx Rzy Rz Tz Energy Nprot Nlig i=1 j=1 Energy of pose f(xi,xj) (output) END Host Processor PCI Express Bus GPU accelerator"Benchmarking energy efficiency, power costs and carbon emissions on heterogeneous systems", Simon McIntosh-Smith, Terry Wilson, 10Amaurys Avila Ibarra, Jon Crisp and Richard B.Sessions, The Computer Journal, September 12th 2011. DOI: 10.1093/comjnl/bxr091
    • !  Systems benchmarkedHigh-end: Medium-end:•  Supermicro 1U dual •  Workstation with 1 GPU server CPU & 1 GPU•  Two Intel 5500 series •  Intel E8500 3.16 GHz 2.4 GHz Xeon dual core CPU ‘Nehalem’ quad-core processors •  Previous generation•  Two Nvidia C2050 Nvidia consumer-level ‘Fermi’ GPUs or GPU, the GTX280•  Two AMD ‘Cypress’ FirePro V7800s 11
    • !  Supermicro GPU server 12
    • !  Systems benchmarkedMiddle-end: Low-end:•  Workstation based on •  Laptop based on an a 3-core AMD 2.8 Intel Core2Duo GHz Phenom II X3 SU9400 ‘Penryn’ 1.4 720 GHz CPU•  4 GBytes of DRAM •  4 GBytes of DRAM•  No GPU! •  No GPU! 13
    • !  Benchmarking methodology•  Use the same power measurement equipment for all the systems under test•  Watts Up? Pro meter•  Measures complete system power at the wall•  Run as fast as possible on all available resources (i.e. all cores or all GPUs simultaneously) 14
    • !  Relative performance 6.9X speedup 1,120 seconds per simulation 162 seconds per simulation Less than 2 days to screen a library of 1 million drug candidates on 1000 GPUs 15
    • !  Relative energy efficiency 68% energy reduction 0.034 kWh for the same work per simulation 0.011 kWh per simulation 0.011 kWh = 0.16 pence per simulation 1 million simulations  GPUs save £3,680 on energy costs for one experiment 16
    • !  Future work•  Modify the code to use the CPUs and GPUs at the same time (done)•  Further optimisations (done – code already several times faster)•  Refine forcefield (in progress … forever)•  Improvements to the Genetic Algorithm and further trimming of the search space (done) 17
    • !  For an introduction to GPUsThe GPU Computing Revolution – aKnowledge Transfer Report from the LondonMathematical Society and the KTN forIndustrial Mathematics•  https://ktn.innovateuk.org/web/mathsktn/ articles/-/blogs/the-gpu-computing- revolution 18
    • !  Conclusions•  GPUs can yield big performance increases if your application has enough of the right sort of parallelism•  GPUs can can also result in significant savings in energy costs (and equipment costs)•  OpenCL can work just as well for multi-core CPUs as it does for GPUsIt’s possible to screen libraries of millions of moleculesagainst complex targets using highly accurate,computationally-expensive methods in one weekendusing equipment costing << £1M 19
    • !  References•  S. McIntosh-Smith, T. Wilson, A.A. Ibarra, J. Crisp and R.B. Sessions, "Benchmarking energy efficiency, power costs and carbon emissions on heterogeneous systems", The Computer Journal, September 12th 2011. DOI: 10.1093/comjnl/bxr091•  N. Gibbs, A.R. Clarke & R.B. Sessions, "Ab- initio Protein Folding using Physicochemical Potentials and a Simplified Off-Lattice Model", Proteins 43:186-202,200 20