vasp 5 4 4
play

VASP 5.4.4 October 2017 Silica IFPEN on V100s PCIe 0.00700 - PowerPoint PPT Presentation

Feb 18, 2024 •260 likes •402 views

VASP 5.4.4 October 2017 Silica IFPEN on V100s PCIe 0.00700 0.00628 0.00600 (Untuned on Volta) 3.0X 0.00537 Running VASP version 5.4.4 0.00500 The blue node contains Dual Intel Xeon 2.6X E5-2690 v4@2.6GHz [3.5GHz Turbo] 0.00418

  1. VASP 5.4.4 October 2017

  2. Silica IFPEN on V100s PCIe 0.00700 0.00628 0.00600 (Untuned on Volta) 3.0X 0.00537 Running VASP version 5.4.4 0.00500 The blue node contains Dual Intel Xeon 2.6X E5-2690 v4@2.6GHz [3.5GHz Turbo] 0.00418 (Broadwell) CPUs 0.00400 2.0X 1/seconds The green nodes contain Dual Intel Xeon E5-2690 v4@2.6GHz [3.5GHz 0.00300 Turbo] (Broadwell) CPUs + 0.00210 Tesla V100 PCIe (16GB) GPUs 0.00200 240 ions, cristobalite (high) bulk 0.00100 720 bands ? plane waves ALGO = Very Fast (RMM-DIIS) 0.00000 1 Broadwell node 1 node + 1 node + 1 node + 2x V100 PCIe 4x V100 PCIe 8x V100 PCIe per node (16GB) per node (16GB) per node (16GB) 65

  3. Silica IFPEN on V100s SXM2 0.00700 0.00580 0.00600 (Untuned on Volta) 0.00541 2.8X Running VASP version 5.4.4 0.00500 The blue node contains Dual Intel Xeon 2.6X 0.00423 E5-2698 v4@2.2GHz [3.6GHz Turbo] (Broadwell) CPUs 0.00400 2.0X 1/seconds The green nodes contain Dual Intel 0.00300 Xeon E5-2698 v4@2.2GHz [3.6GHz Turbo] (Broadwell) CPUs + 0.00210 Tesla V100 SXM2 (16GB) GPUs 0.00200 240 ions, cristobalite (high) bulk 0.00100 720 bands ? plane waves ALGO = Very Fast (RMM-DIIS) 0.00000 1 Broadwell node 1 node + 1 node + 1 node + 2x V100 SXM2 4x V100 SXM2 8x V100 SXM2 per node (16GB) per node (16GB) per node (16GB) 66

  4. Si-Huge on V100s PCIe 0.00070 0.00065 0.00060 0.00057 (Untuned on Volta) 3.8X Running VASP version 5.4.4 0.00050 The blue node contains Dual Intel Xeon 3.4X 0.00045 E5-2690 v4@2.6GHz [3.5GHz Turbo] (Broadwell) CPUs 0.00040 2.6X 1/seconds The green nodes contain Dual Intel 0.00030 Xeon E5-2690 v4@2.6GHz [3.5GHz Turbo] (Broadwell) CPUs + Tesla V100 PCIe (16GB) GPUs 0.00020 0.00017 512 Si atoms 0.00010 1282 bands 864000 Plane Waves Algo = Normal (blocked Davidson) 0.00000 1 Broadwell node 1 node + 1 node + 1 node + 2x V100 PCIe 4x V100 PCIe 8x V100 PCIe per node (16GB) per node (16GB) per node (16GB) 67

  5. Si-Huge on V100s SXM2 0.00080 0.00070 0.00067 (Untuned on Volta) Running VASP version 5.4.4 0.00060 0.00056 4.0X The blue node contains Dual Intel Xeon E5-2698 v4@2.2GHz [3.6GHz Turbo] 0.00050 3.3X 0.00044 (Broadwell) CPUs 1/seconds 0.00040 The green nodes contain Dual Intel 2.6X Xeon E5-2698 v4@2.2GHz [3.6GHz Turbo] (Broadwell) CPUs + 0.00030 Tesla V100 SXM2 (16GB) GPUs 0.00020 0.00017 512 Si atoms 0.00010 1282 bands 864000 Plane Waves Algo = Normal (blocked Davidson) 0.00000 1 Broadwell node 1 node + 1 node + 1 node + 2x V100 SXM2 4x V100 SXM2 8x V100 SXM2 per node (16GB) per node (16GB) per node (16GB) 68

  6. SupportedSystems on V100s PCIe 0.0100 0.0087 0.0090 (Untuned on Volta) 0.0080 Running VASP version 5.4.4 2.4X 0.0068 0.0070 The blue node contains Dual Intel Xeon E5-2690 v4@2.6GHz [3.5GHz Turbo] 1.8X 0.0060 (Broadwell) CPUs 1/seconds 0.0050 The green nodes contain Dual Intel Xeon E5-2690 v4@2.6GHz [3.5GHz 0.0037 0.0040 Turbo] (Broadwell) CPUs + Tesla V100 PCIe (16GB) GPUs 0.0030 0.0020 267 ions 788 bands 0.0010 762048 plane waves ALGO = Fast (Davidson + RMM-DIIS) 0.0000 1 Broadwell node 1 node + 1 node + 2x V100 PCIe 4x V100 PCIe per node (16GB) per node (16GB) 69

  7. SupportedSystems on V100s SXM2 0.0120 0.0100 (Untuned on Volta) 0.0100 Running VASP version 5.4.4 2.7X 0.0087 The blue node contains Dual Intel Xeon 0.0080 E5-2698 v4@2.2GHz [3.6GHz Turbo] 2.4X 0.0068 (Broadwell) CPUs 1/seconds 0.0060 The green nodes contain Dual Intel 1.8X Xeon E5-2698 v4@2.2GHz [3.6GHz Turbo] (Broadwell) CPUs + 0.0037 0.0040 Tesla V100 SXM2 (16GB) GPUs 0.0020 267 ions 788 bands 762048 plane waves ALGO = Fast (Davidson + RMM-DIIS) 0.0000 1 Broadwell node 1 node + 1 node + 1 node + 2x V100 SXM2 4x V100 SXM2 8x V100 SXM2 per node (16GB) per node (16GB) per node (16GB) 70

  8. NiAl-MD on V100s PCIe 0.0080 0.0068 0.0070 (Untuned on Volta) 0.0063 Running VASP version 5.4.4 0.0060 2.2X The blue node contains Dual Intel Xeon E5-2690 v4@2.6GHz [3.5GHz Turbo] 0.0050 2.0X (Broadwell) CPUs 1/seconds 0.0040 The green nodes contain Dual Intel Xeon E5-2690 v4@2.6GHz [3.5GHz 0.0031 Turbo] (Broadwell) CPUs + 0.0030 Tesla V100 PCIe (16GB) GPUs 0.0020 500 ions 0.0010 3200 bands 729000 plane waves ALGO = Fast (Davidson + RMM-DIIS) 0.0000 1 Broadwell node 1 node + 1 node + 2x V100 PCIe 4x V100 PCIe per node (16GB) per node (16GB) 71

  9. NiAl-MD on V100s SXM2 0.0080 0.0074 0.0070 0.0070 2.4X (Untuned on Volta) 0.0064 Running VASP version 5.4.4 0.0060 2.3X The blue node contains Dual Intel Xeon E5-2698 v4@2.2GHz [3.6GHz Turbo] 0.0050 2.1X (Broadwell) CPUs 1/seconds 0.0040 The green nodes contain Dual Intel Xeon E5-2698 v4@2.2GHz [3.6GHz 0.0031 Turbo] (Broadwell) CPUs + 0.0030 Tesla V100 SXM2 (16GB) GPUs 0.0020 500 ions 0.0010 3200 bands 729000 plane waves ALGO = Fast (Davidson + RMM-DIIS) 0.0000 1 Broadwell node 1 node + 1 node + 1 node + 2x V100 SXM2 4x V100 SXM2 8x V100 SXM2 per node (16GB) per node (16GB) per node (16GB) 72

  10. B.hR105 on V100s PCIe 0.0140 (Untuned on Volta) 0.0119 0.0120 Running VASP version 5.4.4 0.0112 14.9X The blue node contains Dual Intel Xeon 0.0100 14.0X E5-2690 v4@2.6GHz [3.5GHz Turbo] (Broadwell) CPUs 0.0077 0.0080 The green nodes contain Dual Intel 1/seconds 9.6X Xeon E5-2690 v4@2.6GHz [3.5GHz 0.0060 Turbo] (Broadwell) CPUs + Tesla V100 PCIe (16GB) GPUs 0.0040 105 Boron atoms ( β -rhombohedral structure) 216 bands 0.0020 110592 plane waves 0.0008 Hybrid Functional with blocked Davicson (ALGO=Normal) 0.0000 LHFCALC=.True. (Exact Exchange) 1 Broadwell node 1 node + 1 node + 1 node + 2x V100 PCIe 4x V100 PCIe 8x V100 PCIe per node (16GB) per node (16GB) per node (16GB) 73

  11. B.hR105 on V100s SXM2 0.0140 0.0128 (Untuned on Volta) 0.0116 0.0120 Running VASP version 5.4.4 16.0X The blue node contains Dual Intel Xeon 0.0100 14.5X E5-2698 v4@2.2GHz [3.6GHz Turbo] (Broadwell) CPUs 0.0079 0.0080 The green nodes contain Dual Intel 1/seconds 9.9X Xeon E5-2698 v4@2.2GHz [3.6GHz 0.0060 Turbo] (Broadwell) CPUs + Tesla V100 SXM2 (16GB) GPUs 0.0040 105 Boron atoms ( β -rhombohedral structure) 216 bands 0.0020 110592 plane waves 0.0008 Hybrid Functional with blocked Davicson (ALGO=Normal) 0.0000 LHFCALC=.True. (Exact Exchange) 1 Broadwell node 1 node + 1 node + 1 node + 2x V100 SXM2 4x V100 SXM2 8x V100 SXM2 per node (16GB) per node (16GB) per node (16GB) 74

  12. B.aP107 on V100s PCIe 0.000600 (Untuned on Volta) Running VASP version 5.4.4 0.000490 0.000500 The blue node contains Dual Intel Xeon 0.000462 E5-2690 v4@2.6GHz [3.5GHz Turbo] 12.9X (Broadwell) CPUs 0.000400 12.2X The green nodes contain Dual Intel 0.000323 Xeon E5-2690 v4@2.6GHz [3.5GHz 1/seconds Turbo] (Broadwell) CPUs + 0.000300 Tesla V100 PCIe (16GB) GPUs 8.5X 0.000200 107 Boron atoms (symmetry broken 107-atom β′ variant) 216 bands 0.000100 110592 plane waves Hybrid functional calculation (exact 0.000038 exchange) with blocked Davidson. No KPoint parallelization. 0.000000 Hybrid Functional with blocked Davidson 1 Broadwell node 1 node + 1 node + 1 node + (ALGO=Normal) 2x V100 PCIe 4x V100 PCIe 8x V100 PCIe LHFCALC=.True. (Exact Exchange) per node (16GB) per node (16GB) per node (16GB) 75

  13. B.aP107 on V100s SXM2 0.000600 (Untuned on Volta) Running VASP version 5.4.4 0.000523 0.000500 The blue node contains Dual Intel Xeon 0.000465 13.8X E5-2698 v4@2.2GHz [3.6GHz Turbo] (Broadwell) CPUs 0.000400 12.2X The green nodes contain Dual Intel 0.000324 Xeon E5-2698 v4@2.2GHz [3.6GHz 1/seconds Turbo] (Broadwell) CPUs + 0.000300 Tesla V100 SXM2 (16GB) GPUs 8.5X 0.000200 107 Boron atoms (symmetry broken 107-atom β′ variant) 216 bands 0.000100 110592 plane waves Hybrid functional calculation (exact 0.000038 exchange) with blocked Davidson. No KPoint parallelization. 0.000000 Hybrid Functional with blocked Davidson 1 Broadwell node 1 node + 1 node + 1 node + (ALGO=Normal) 2x V100 SXM2 4x V100 SXM2 8x V100 SXM2 LHFCALC=.True. (Exact Exchange) per node (16GB) per node (16GB) per node (16GB) 76

Download Presentation
Download Policy: The content available on the website is offered to you 'AS IS' for your personal information and use only. It cannot be commercialized, licensed, or distributed on other websites without prior consent from the author. To download a presentation, simply click this link. If you encounter any difficulties during the download process, it's possible that the publisher has removed the file from their server.

Recommend

NEW GPU FUNCTIONALITY IN VASP WITH OPENACC AND CUDA LIBRARIES Stefan Maintz, 2019/12/18 AGENDA

NEW GPU FUNCTIONALITY IN VASP WITH OPENACC AND CUDA LIBRARIES Stefan Maintz, 2019/12/18 AGENDA

NEW GPU FUNCTIONALITY IN VASP WITH OPENACC AND CUDA LIBRARIES Stefan Maintz, 2019/12/18 AGENDA Introduction to VASP GPU Acceleration in VASP 5 Prioritizing Use Cases for New Porting Efforts OpenACC in VASP 6 and Supported Features

705 views • 42 slides
VASP 5.4.1 February 2017 Interface on P100s PCIe 0.00500 Interface Running VASP version 5.4.1

VASP 5.4.1 February 2017 Interface on P100s PCIe 0.00500 Interface Running VASP version 5.4.1

VASP 5.4.1 February 2017 Interface on P100s PCIe 0.00500 Interface Running VASP version 5.4.1 0.00450 0.00434 The blue node contains Dual Intel Xeon E5-2699 v4@2.2GHz [3.6GHz Turbo] 0.00400 2.5X (Broadwell) CPUs 0.00359 0.00350 2.1X

451 views • 17 slides
PORTING VASP TO GPUS WITH OPENACC Stefan Maintz, Dr. Markus Wetzstein 03/26/2018

PORTING VASP TO GPUS WITH OPENACC Stefan Maintz, Dr. Markus Wetzstein 03/26/2018

PORTING VASP TO GPUS WITH OPENACC Stefan Maintz, Dr. Markus Wetzstein 03/26/2018 smaintz@nvidia.com; mwetzstein@nvidia.com Short introduction to VASP Status of the CUDA port Prioritizing Use Cases for GPU Acceleration AGENDA OpenACC in VASP

776 views • 41 slides
VASP Project Briefing Report Eastern Alliance for Greenhouse Action Northern Alliance for

VASP Project Briefing Report Eastern Alliance for Greenhouse Action Northern Alliance for

Future Assets Forum VASP Project Briefing Report Eastern Alliance for Greenhouse Action Northern Alliance for Greenhouse Action December 2015 Future Assets Forum Workshop 1 Dr Gerard Healey and Aaron Yuen 22 nd October 2015 Activity 1: What

588 views • 26 slides
Type III Secretion System 1 Pedestal Formation Cytoskeletal Proteins Recruited Tip: active in

Type III Secretion System 1 Pedestal Formation Cytoskeletal Proteins Recruited Tip: active in

Type III Secretion System 1 Pedestal Formation Cytoskeletal Proteins Recruited Tip: active in pedestal formation Tir, N-WASP, -actinin, Arp2/3, gelsolin, talin, VASP, Nck, CD44 Length: structural components -actinin, Arp2/3, talin,

512 views • 47 slides
Welcome! Todays Agenda: GPU Execution Model GPGPU Flow GPGPU Low Level Notes

Welcome! Todays Agenda: GPU Execution Model GPGPU Flow GPGPU Low Level Notes

/INFOMOV/ Optimization & Vectorization J. Bikker - Sep-Nov 2019 - Lecture 10: GPGPU (3) Welcome! Todays Agenda: GPU Execution Model GPGPU Flow GPGPU Low Level Notes P3 INFOMOV Lecture 10 GPGPU

437 views • 32 slides
Statistical Learning and Inference Methods for Reasoning in Games March 30, 2005 Brian Mihok

Statistical Learning and Inference Methods for Reasoning in Games March 30, 2005 Brian Mihok

Statistical Learning and Inference Methods for Reasoning in Games March 30, 2005 Brian Mihok Michael Terry Draper Laboratory CSAIL, MIT 1 Outline Intro to Games Texas Hold em Demo Hidden Markov Models Structure Learning

1.46k views • 55 slides
Neutrino diffusion in General Relativity Giovanni Camelio with Stephan Rosswog from Astronomy

Neutrino diffusion in General Relativity Giovanni Camelio with Stephan Rosswog from Astronomy

Neutrino diffusion in General Relativity Giovanni Camelio with Stephan Rosswog from Astronomy Department Stockholm University February 6th, 2018 @ Lund Observatory Motivation There are systems in astrophysics where: matter is opaque to

220 views • 9 slides
CHIRAL DYNAMICS AT KLOE, MAINZ, ELSA AND OTHER LABS F. Ambrosino Universit degli Studi di

CHIRAL DYNAMICS AT KLOE, MAINZ, ELSA AND OTHER LABS F. Ambrosino Universit degli Studi di

CHIRAL DYNAMICS AT KLOE, MAINZ, ELSA AND OTHER LABS F. Ambrosino Universit degli Studi di Napoli Federico II e Sezione INFN, Napoli, Italy Chiral Dynamics Study of (pseudo)Goldstone bosons dynamics: pions, kaons etas The most

840 views • 46 slides
Solving Domain Wall Dirac Equation Using Multisplitting Preconditioned Conjugate Gradient Jiqun

Solving Domain Wall Dirac Equation Using Multisplitting Preconditioned Conjugate Gradient Jiqun

Solving Domain Wall Dirac Equation Using Multisplitting Preconditioned Conjugate Gradient Jiqun Tu 1 1 Department of Physics, Columbia University The 36th International Symposium on Lattice Field Theory, July 23, 2018 @ 16:10 Talk based on: Duo

725 views • 31 slides
WITH CUDA C/C++ Pedro Mario Cruz e Silva, Solutions Architect Manager ELEVEN YEARS OF GPU

WITH CUDA C/C++ Pedro Mario Cruz e Silva, Solutions Architect Manager ELEVEN YEARS OF GPU

ACCELERATING APPLICATIONS WITH CUDA C/C++ Pedro Mario Cruz e Silva, Solutions Architect Manager ELEVEN YEARS OF GPU COMPUTING GPU-Trained AI Machine Beats World Champion in Go Worlds First Atomic Model of HIV Capsid Oak Ridge Deploys

562 views • 23 slides
GPGPU 03 NVIDIA case study GeForce 7800 (2006) GeForce 7800 Impossible to maximize

GPGPU 03 NVIDIA case study GeForce 7800 (2006) GeForce 7800 Impossible to maximize

GPGPU 03 NVIDIA case study GeForce 7800 (2006) GeForce 7800 Impossible to maximize throughput with such a rigid architecture: you cant keep vertex and fragment shading units busy all the time As a result, many bottlenecks in the

1.4k views • 108 slides
Common Subexpression Convergence (CSC) Sana Damani and Vivek Sarkar Habanero Extreme Scale

Common Subexpression Convergence (CSC) Sana Damani and Vivek Sarkar Habanero Extreme Scale

Common Subexpression Convergence (CSC) Sana Damani and Vivek Sarkar Habanero Extreme Scale Software Research Lab Georgia Institute of Technology Short paper at LCPC 19, Atlanta, GA Agenda Motivation Common Subexpression Convergence

494 views • 23 slides
Discourse Structure & Wrap-up: Q-A Ling571 Deep Processing Techniques for NLP March 9, 2016

Discourse Structure & Wrap-up: Q-A Ling571 Deep Processing Techniques for NLP March 9, 2016

Discourse Structure & Wrap-up: Q-A Ling571 Deep Processing Techniques for NLP March 9, 2016 TextTiling Segmentation Depth score: Difference between position and adjacent peaks E.g., (y a1 -y a2 )+(y a3 -y a2 ) Evaluation

783 views • 40 slides
The Present and Absent Lord 2019 TRINITY LECTURE 2 30 JULY 2019 MARKUS BOCKMUEHL, UNIVERSITY

The Present and Absent Lord 2019 TRINITY LECTURE 2 30 JULY 2019 MARKUS BOCKMUEHL, UNIVERSITY

St Paul on the Absence and Presence of Jesus The Present and Absent Lord 2019 TRINITY LECTURE 2 30 JULY 2019 MARKUS BOCKMUEHL, UNIVERSITY OF OXFORD Introduction Lecture 1: despite a number of seemingly straightforward Pauline

460 views • 14 slides
r r tr

r r tr

r r tr r r sr t ts void put(x) Val

484 views • 35 slides
Two-Player Perfect Information Games: A Brief Survey Tsan-sheng Hsu tshsu@iis.sinica.edu.tw

Two-Player Perfect Information Games: A Brief Survey Tsan-sheng Hsu tshsu@iis.sinica.edu.tw

Two-Player Perfect Information Games: A Brief Survey Tsan-sheng Hsu tshsu@iis.sinica.edu.tw http://www.iis.sinica.edu.tw/~tshsu 1 Abstract Domain: two-player games. Which game characters are predominant when the solution of a game is the

805 views • 49 slides
2 3 4 5 6 7 8 Each of our trips are planned and

2 3 4 5 6 7 8 Each of our trips are planned and

2 3 4 5 6 7 8 Each of our trips are planned and measured on some or all of the following: 9 Meetings & Meals Panels 11

704 views • 25 slides
Detect ctor Installation in the cr cryostat Jim Stewart August 20 2019 FS Installation

Detect ctor Installation in the cr cryostat Jim Stewart August 20 2019 FS Installation

Detect ctor Installation in the cr cryostat Jim Stewart August 20 2019 FS Installation Workshop SURF -SD Outline Starting Conditions E-Endwall installation APA-CPA installation W-Endwall and TCO closing 8/12/19 J Stewart |

271 views • 23 slides
AFRICAN DUST OUTBREAKS A satellite perspective of temporal and spatial variability over the

AFRICAN DUST OUTBREAKS A satellite perspective of temporal and spatial variability over the

AFRICAN DUST OUTBREAKS A satellite perspective of temporal and spatial variability over the tropical Atlantic Ocean Jingfeng Huang 1,2 , Chidong Zhang 3 , Joseph M. Prospero 3 1 UMBC GEST; 2 NASA GSFC; 3 University of Miami, RSMAS Thanks to:

787 views • 16 slides
Automated Reasoning 2nd Coursework Elaine Murphy Oct 30th Elaine Murphy Automated Reasoning 2nd

Automated Reasoning 2nd Coursework Elaine Murphy Oct 30th Elaine Murphy Automated Reasoning 2nd

Automated Reasoning 2nd Coursework Elaine Murphy Oct 30th Elaine Murphy Automated Reasoning 2nd Coursework Oct 30th 1 Information Demonstrator: Elaine Murphy s0094243@sms.ed.ac.uk Lab sessions: AT level 5 South lab. Mon 14.10 - 15.00 Wed

345 views • 20 slides
CS 188: Artificial Intelligence Markov Decision Processes Instructor: Anca Dragan University of

CS 188: Artificial Intelligence Markov Decision Processes Instructor: Anca Dragan University of

CS 188: Artificial Intelligence Markov Decision Processes Instructor: Anca Dragan University of California, Berkeley [These slides adapted from Dan Klein and Pieter Abbeel] Non-Deterministic Search Example: Grid World A maze-like problem

882 views • 52 slides
Complete the Student Project for this module You may work in a group, but you are responsible for

Complete the Student Project for this module You may work in a group, but you are responsible for

Complete the Student Project for this module You may work in a group, but you are responsible for your own work Turn in the completed assignment for review

393 views • 23 slides
Agile in Practice Duke CompSci Fall 2019 Presented by Kimberly Andrikaitis 1 ABOUT ME

Agile in Practice Duke CompSci Fall 2019 Presented by Kimberly Andrikaitis 1 ABOUT ME

Agile in Practice Duke CompSci Fall 2019 Presented by Kimberly Andrikaitis 1 ABOUT ME Profession: Agile Transformation Coach, Vaco Career Path: Graphic Designer, Tester, Business Analyst, Product Owner, Scrum Master

453 views • 22 slides

More recommend