IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk...
Transcript of IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk...
![Page 1: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/1.jpg)
IPCC @ RWTH Aachen UniversityOptimization of multibody and long-range solvers in
LAMMPS
Rodrigo Canales William McDoniel Markus HohnerbachAhmed E. Ismail Paolo Bientinesi
IPCC Showcase – November 2016
![Page 2: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/2.jpg)
Team
RWTH
Prof. Paolo Bientinesi Rodrigo Canales William McDoniel Markus Hohnerbach Prof. Ahmed Ismail
Intel
Georg Zitzlsberger Klaus-Dieter Oertel Michael W. Brown
2
![Page 3: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/3.jpg)
Introduction
2015
I May: Kickoff – IPCC @ RWTH AachenOptimizing LAMMPS kernels
I Oct.: First results on Xeon & KNC, @ EMEA IPCC
2016
I Feb.: Showcase 1st year
I March: First results on KNL, @ IPCC & IXPUG Forum
I May: KNL Access
I Nov.: Showcase
2017
I May: End 2nd year
3
![Page 4: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/4.jpg)
Agenda
I Intro to MD, LAMMPS
I Achievements 1st year
I Goals & Progress 2nd yearI AIREBOI REBOI PPPM ElectrostaticsI PPPM Dispersion
I Future Projects
4
![Page 5: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/5.jpg)
LAMMPS
Large-scale Atomic–Molecular Massively Parallel Simulator
I Sandia National Labshttp://lammps.sandia.gov
I Widely used open source MDcode
I Support for OpenMP, Xeon Phi,and GPU (CUDA and OpenCL)
5
![Page 6: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/6.jpg)
Molecular Dynamics
I Many particle systems
I Computes interactions betweenpairs of atoms
ΦLJ = 4ε[( σ
rij
)12−( σrij
)6]
6
![Page 7: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/7.jpg)
First Year
I Pair Potentials
I KNL Ready
7
![Page 8: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/8.jpg)
Buckingham: KNC vs. KNL - Full Node
KNC KNL0
5,000
10,000
15,000
20,000
tau
/day
Default Mode HBM (via numactl)
8
![Page 9: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/9.jpg)
Tersoff: KNC
ThreadCores 1SMT 1Atoms 32.000
CoreCores 1SMT 4Atoms 32.000
FullCores 60SMT 4Atoms 512.000
Measurements in 1000 atom-ns/day/core, SMT minimizes runtime.
Thread Core Full0
20
40
60
80
atom
-ns/day/core
Ref Double Single Mixed
9
![Page 10: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/10.jpg)
Tersoff: KNL
ThreadCores 1SMT 1Atoms 32.000HBM Yes
CoreCores 1SMT 4Atoms 32.000HBM Yes
FullCores 64SMT 4Atoms 512.000HBM Yes
Measurements in 1000 atom-ns/day/core, SMT minimizes runtime.
Thread Core Full0
20
40
60
80
atom
-ns/day/core
Ref Double Single Mixed
9
![Page 11: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/11.jpg)
The Vectorization of the Tersoff Many-Body Potential:An Exercise in Performance Portability
I Initial work: workshop on MD simulation software @ SC’15I Full portability across existing Intel archsI Focus on vector operation wrapper
I Submitted to SC’16 technical programI Additional architecturesI KNL results
I KNL measurements via Mike (Thanks!)
I For submission: NDA waiver
I Accepted
I Best Student Paper Finalist
I (Maybe) part of replication initiative SC’17
10
![Page 12: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/12.jpg)
Second Year (After Q2)
I Multi-body Potentials
I Long Range Interactions
11
![Page 13: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/13.jpg)
Multi-body Potential: REBO
I Similar to Tersoff
I Applicable to Carbohydrates
I Improves Tersoff through additional terms
I Additional neighbor finding routines needed by REBO (Ready)
I Vectorized/Optimized code for KNC/KNL (Ready)
I Optimized code for CPU, same approach as Tersoff (Ready)
I Vectorized/Optimized code for CPU (In Progress)
I Offloading Performance (In Progress)
I Speedup KNL: ca. 2.5x total, and ca. 3x on kernel
I Bottleneck: Neighbor Lists
12
![Page 14: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/14.jpg)
REBO Results – KNL
Ref Opt0
5
10
15
20
25
Tim
e(s
)
Pair Neigh
13
![Page 15: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/15.jpg)
AIREBO
I Based on REBO
I Two additional terms: Torsion and Lennard-Jones
I Torsion: Easy to vectorize (Ready)
I Lennard-Jones: Hard to vectorize (In Progress)
I Search through neighbor list and branch
I Idea: Separate expensive and cheap cases
14
![Page 16: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/16.jpg)
Long Range Interactions: PPPM
I Cutoff distances make pair potential calculations feasible
-1.5
-1
-0.5
0
0.5
1
1.5
2
2.5
3
1 1.5 2 2.5 3 3.5
V / ε
r / σ
Re
pu
lsiv
e
Attractive
Re
pu
lsiv
e
Attractive
Cu
toff r
cI Long-range calculations can still be important:
I ElectrostaticsI Interfaces
I Particle-Particle Particle-Mesh (PPPM) approximateslong-range forces without requiring pair-wise calculations
15
![Page 17: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/17.jpg)
PPPM
Four Steps:
1. Determine the charge distribution ρ by mapping particlecharges to a grid
2. Take the Fourier transform of the charge distribution to findthe potential:
∇2Φ = − ρε0
3. Obtain forces due to all interactions by inverse Fouriertransform:
~F = −∇Φ
4. Map forces back to the particles
16
![Page 18: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/18.jpg)
PPPM: Charge Mapping
17
![Page 19: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/19.jpg)
PPPM: Charge Mapping
17
![Page 20: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/20.jpg)
PPPM: Charge Mapping
17
![Page 21: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/21.jpg)
PPPM: Charge Mapping
17
![Page 22: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/22.jpg)
PPPM: Charge Mapping
17
![Page 23: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/23.jpg)
PPPM: Charge Mapping
17
![Page 24: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/24.jpg)
PPPM: Charge Mapping
17
![Page 25: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/25.jpg)
PPPM: Charge Mapping
17
![Page 26: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/26.jpg)
PPPM: Charge Mapping
17
![Page 27: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/27.jpg)
PPPM: Charge Mapping
17
![Page 28: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/28.jpg)
PPPM: Charge Mapping
17
![Page 29: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/29.jpg)
PPPM: Charge Mapping
17
![Page 30: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/30.jpg)
PPPM: Charge Mapping
17
![Page 31: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/31.jpg)
PPPM: Charge Mapping
17
![Page 32: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/32.jpg)
PPPM: Charge Mapping
17
![Page 33: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/33.jpg)
PPPM: Charge Mapping
17
![Page 34: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/34.jpg)
PPPM: Charge Mapping
17
![Page 35: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/35.jpg)
PPPM: Charge Mapping
17
![Page 36: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/36.jpg)
PPPM: Charge Mapping
17
![Page 37: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/37.jpg)
PPPM: Charge Mapping
17
![Page 38: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/38.jpg)
PPPM: Charge Mapping
17
![Page 39: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/39.jpg)
PPPM: Charge Mapping
17
![Page 40: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/40.jpg)
PPPM: Charge Mapping
17
![Page 41: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/41.jpg)
PPPM: Charge Mapping
17
![Page 42: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/42.jpg)
PPPM: Charge Mapping
17
![Page 43: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/43.jpg)
PPPM: Charge Mapping
17
![Page 44: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/44.jpg)
PPPM: Charge Mapping
17
![Page 45: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/45.jpg)
PPPM: Charge Mapping
17
![Page 46: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/46.jpg)
PPPM: Charge Mapping
17
![Page 47: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/47.jpg)
PPPM: Charge Mapping
17
![Page 48: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/48.jpg)
PPPM: Charge Mapping
17
![Page 49: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/49.jpg)
PPPM: Charge Mapping
17
![Page 50: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/50.jpg)
PPPM: Charge Mapping
18
![Page 51: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/51.jpg)
PPPM: Charge Mapping
18
![Page 52: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/52.jpg)
PPPM: Charge Mapping
18
![Page 53: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/53.jpg)
PPPM: Charge Mapping
18
![Page 54: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/54.jpg)
PPPM: Charge Mapping
18
![Page 55: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/55.jpg)
PPPM: Charge Mapping
18
![Page 56: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/56.jpg)
PPPM: Charge Mapping
18
![Page 57: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/57.jpg)
PPPM: Charge Mapping
18
![Page 58: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/58.jpg)
PPPM: Charge Mapping
18
![Page 59: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/59.jpg)
PPPM: Charge Mapping
18
![Page 60: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/60.jpg)
PPPM: Charge Mapping
18
![Page 61: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/61.jpg)
PPPM: Charge Mapping
18
![Page 62: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/62.jpg)
PPPM: Charge Mapping
18
![Page 63: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/63.jpg)
PPPM: Charge Mapping
18
![Page 64: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/64.jpg)
PPPM: Charge Mapping
18
![Page 65: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/65.jpg)
PPPM: Charge Mapping
18
![Page 66: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/66.jpg)
PPPM: Charge Mapping
18
![Page 67: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/67.jpg)
PPPM: Charge Mapping
18
![Page 68: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/68.jpg)
PPPM: Charge Mapping
18
![Page 69: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/69.jpg)
PPPM: Charge Mapping
18
![Page 70: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/70.jpg)
PPPM: Charge Mapping
I Threading across atoms instead of across grid points
I Look-up table for stencil coefficients
I Vectorized inner stencil loop
I Larger stencil takes advantage of KNL vector length
19
![Page 71: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/71.jpg)
PPPM: Charge Mapping on KNL 1c/1t
40.5k 81k 162k 324k0
0.5
1
1.5·10−3
Problem size (atoms)
tim
ep
erat
om(s
)
baseline new threading vectorized
20
![Page 72: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/72.jpg)
PPPM: Distributing Forces
I Look-up table for stencil coefficients
I Vectorized inner stencil loop
I Larger stencil takes advantage of KNL vector length
I Repack force data to get multiple components simultaneously
21
![Page 73: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/73.jpg)
PPPM: Distributing Forces on KNL 1c/1t
40.5k 81k 162k 324k0
0.5
1
1.5·10−3
Problem size (atoms)
tim
ep
erat
om(s
)
baseline atom loop vectorized
inner stencil vectorized repacking
22
![Page 74: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/74.jpg)
PPPM: FFTs
I Larger stencil allows coarser grid while preserving accuracy
I Reduce communication by doing:2D → remap → 1D → remapinstead of:1D → remap → 1D → remap → 1D → remap
I Vectorization elsewhere makes ad differentiation relativelymore appealing than ik differentiation – half as many FFTs inexchange for more work in stencil loops
23
![Page 75: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/75.jpg)
Water Benchmark – KNL 1c/1t
4 5 6 7 80
20
40
60
Cutoff (angstroms)
Tim
e(s
)
Reference
PPPM non-FFT PPPM FFT Pair Other
4 5 6 7 80
20
40
60
3.5X
3.4X3.1X3.0X
2.4X
Cutoff (angstroms)
Tim
e(s
)
Optimized
PPPM non-FFT PPPM FFT Pair Other
24
![Page 76: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/76.jpg)
Water Benchmark – KNL 64c/1t
4 5 6 7 8 90
20
40
60
80
100
Cutoff (angstroms)
Tim
e(s
)
Reference
PPPM non-FFT PPPM FFT Pair Other
4 5 6 7 8 90
20
40
60
80
100
2.8X2.9X
2.5X2.3X
1.9X
1.5X
Cutoff (angstroms)
Tim
e(s
)
Optimized
PPPM non-FFT PPPM FFT Pair Other
25
![Page 77: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/77.jpg)
PPPM Dispersion
Similar particle mapping concept, but with two potentials:
I Electrostatics ∼ 1r
I Dispersion interactions ∼ 1r6
I Optimize compatible pair potentials (Ready)I Buckingham (buck/long/coul/long)I Lennard Jones (lj/long/coul/long)
I Optimize PPPM-dispersion solver (In Progress)
26
![Page 78: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/78.jpg)
Buckingham – Dispersion
I SiO2 model, 19200 atoms - coulomb and buck potentials
I KNL Cache mode
I Reference : USER OMP
1 Thr 2 Thr 64 × 2 Thr
0
2
4
Speedup on KNL
ref double single
mixed
27
![Page 79: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/79.jpg)
PPPM Dispersion: Components
Having multiple types of forces requires different mixing rules:
I Equivalent routines operate on different stencilsI 2 versions of particle mappingI 4 versions of charge densityI 12 versions of force distribution & poisson solver
I We use templates, optimizing only onceI Minimize control structures
28
![Page 80: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/80.jpg)
PPPM Dispersion: Results
I Optimized charge & particle mappingdouble precision + single precision FFTs
I Between 1.4X and 1.6X speedup on K-space
I Potential speedups for poisson & force
Base Opt0
20
40
60
80
Tim
e(m
s) poissonforce dist.part. mappingcharge density
29
![Page 81: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/81.jpg)
Code Availability
Code Github1 LAMMPS
Tersoff X X
Buckingham X XBuckingham Coul Long X XBuckingham Long Coul Long X . . .
Lennard-Jones Long Coul Long X . . .
PPPM . . . . . .PPPM Dispersion X . . .
REBO . . . . . .AIREBO . . . . . .
1Our group’s repositories are at github.com/HPAC.
30
![Page 82: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/82.jpg)
Dissemination and Community Involvement
I SIAM CSE 2017, Atlanta: MD Exascale Mini-SymposiumI Bientinesi (Aachen), McDoniel (Aachen), Tchipev (Munchen)
I ISC’17 Paper
I SC’16 Technical Program Talk
I IPCC Meeting Toulouse Talk
I Paper for IXPUG Workshop @ ISC’16:“Dynamic SIMD Lane Scheduling”
I Krzikalla (Dresden), Wende (Berlin), Hohnerbach (Aachen)
I ISC’16 Booth and IPCC Meeting TalksI Parallel’16 Talk
I Krzikalla (Dresden), Hohnerbach (Aachen)
I IPCC Meeting Ostrava Talk
I SC’15 Workshop Talk
I IPCC Meeting Munchen Code Dungeon
31
![Page 83: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/83.jpg)
Other activities & future work
32
![Page 84: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/84.jpg)
Other research activities
I Tensors operationsI Tensor transposition, summations, contractions
I Applications from Chemistry and Machine Learning
I Collaboration with IPCC UT Austin
I BLASI Idea: CPU + stream to Phi
I MKL – limited functionality
I Application in Density Functional Theory
I Initial results: 1610 vs 1350 GFLOPS/s (MKL)
33
![Page 85: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/85.jpg)
LAMMPS
I Continue collaboration with Mike BrownI Additional Long-Ranged Solvers
I Multi-Level Summation (MSM): O(n) algorithmI 2/3rd of routines similar to PPPM (particle to grid and back)I 1/3rd: Stencil application (Research topic)I MSM Dispersion solver developed by our group
I Gaussian split Ewald: Mesh-based real/frequency spaceI Might provide better accuracy than MSMI First implementation into LAMMPS
I Extend KOKKOS to enable vector classes (avoid GPU bias)
DSMC
I Particle-based method for rarefied gas
I Similar to molecular dynamics and LAMMPS
34
![Page 86: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/86.jpg)
![Page 87: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/87.jpg)
Buckingham: vectorization, single thread
buck/cut buck/coul/cut buck/cut/coul/long
0
2
4
6
8
10
12
Speedup on KNL (1 Thread)
Double Single Mixed
36
![Page 88: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/88.jpg)
Buckingham: vectorization, full node
buck/cut buck/coul/cut buck/cut/coul/long
0
2
4
6
8
10
12
Speedup on KNL (MPI + multithread)
Double Single Mixed
37
![Page 89: IPCC @ RWTH Aachen Universityhpac.rwth-aachen.de/ipcc/showcase_nov16.pdfI IPCC Meeting Toulouse Talk I Paper for IXPUG Workshop @ ISC’16: \Dynamic SIMD Lane Scheduling" I Krzikalla](https://reader035.fdocuments.us/reader035/viewer/2022062607/60453c9c29235619ad1a4906/html5/thumbnails/89.jpg)
Part 2: Tersoff multibody potential Markus Hohnerbach
for i in local atoms of the current thread dofor j in atoms neighboring i do
ζij ← 0;for k in atoms neighboring i do
ζij ← ζij + ζ(i , j , k);
E ← E + V (i , j , ζij);Fi ← Fi − ∂xiV (i , j , ζij);Fj ← Fj − ∂xjV (i , j , ζij);δζ ← ∂ζV (i , j , ζij);for k in atoms neighboring i do
Fi ← Fi − δζ · ∂xi ζ(i , j , k);Fj ← Fj − δζ · ∂xj ζ(i , j , k);Fk ← Fk − δζ · ∂xk ζ(i , j , k);
38