SIMTization and SIMDization of Verlet list construction for molecular dynamics simulations

Compile and run

$ make hsw
$ ./make_list_cpu_4x1.out

$ make skl
$ ./make_list_cpu_8x1.out

$ make knl
$ ./make_list_mic_8x1.out

$ make gpu
$ ./make_list_gpu_warp_unroll_smem.out

CPUはsingleコア実行での実行時間 (単位はms)

	Teska K40t	Tesla P100	E5-2680 v3 @ 2.50GHz	Phi 7250 @ 1.40GHz	Gold 6148 @ 2.40GHz
density 1.0	995	469	12395	28546	8587
density 0.5	432	223	3531	10220	2466

Name		Name	Last commit message	Last commit date
Latest commit History 80 Commits
bandwidth_test		bandwidth_test
GPU_STREAM_RESULT.txt		GPU_STREAM_RESULT.txt
GTX_TITAN_profile.txt		GTX_TITAN_profile.txt
LICENSE.txt		LICENSE.txt
Makefile		Makefile
Metrics		Metrics
README.md		README.md
STREAM_RESULT.txt		STREAM_RESULT.txt
Tesla_K40t_profile.txt		Tesla_K40t_profile.txt
bench_rawdata_cpu.dat		bench_rawdata_cpu.dat
bench_rawdata_gpu.dat		bench_rawdata_gpu.dat
cuda_ptr.cuh		cuda_ptr.cuh
device_util.cuh		device_util.cuh
dyalloc2d.hpp		dyalloc2d.hpp
kernel_impl.cuh		kernel_impl.cuh
make_list.cpp		make_list.cpp
make_list.cu		make_list.cu
neighlist_cpu.hpp		neighlist_cpu.hpp
neighlist_cpu_avx2.hpp		neighlist_cpu_avx2.hpp
neighlist_cpu_avx512.hpp		neighlist_cpu_avx512.hpp
neighlist_gpu.hpp		neighlist_gpu.hpp
simd_util.hpp		simd_util.hpp