tu-c0r3n00 - 0 device=0 binding=--interleave=0,1 tu-c0r3n00 - 1 device=1 binding=--interleave=2,3 tu-c0r3n00 - 2 device=2 binding=--interleave=4,5 tu-c0r3n00 - 3 device=3 binding=--interleave=6,7 tu-c0r3n06 - 1 device=1 binding=--interleave=2,3 tu-c0r3n06 - 3 device=3 binding=--interleave=6,7 tu-c0r3n06 - 0 device=0 binding=--interleave=0,1 tu-c0r3n06 - 2 device=2 binding=--interleave=4,5 tu-c0r3n03 - 1 device=1 binding=--interleave=2,3 tu-c0r3n03 - 2 device=2 binding=--interleave=4,5 tu-c0r3n03 - 0 device=0 binding=--interleave=0,1 tu-c0r3n03 - 3 device=3 binding=--interleave=6,7 tu-c0r3n09 - 0 device=0 binding=--interleave=0,1 tu-c0r3n09 - 1 device=1 binding=--interleave=2,3 tu-c0r3n09 - 2 device=2 binding=--interleave=4,5 tu-c0r3n09 - 3 device=3 binding=--interleave=6,7 OPENMPI detected AcceleratorCudaInit: using default device AcceleratorCudaInit: assume user either uses AcceleratorCudaInit: a) IBM jsrun, or AcceleratorCudaInit: b) invokes through a wrapping script to set CUDA_VISIBLE_DEVICES, UCX_NET_DEVICES, and numa binding AcceleratorCudaInit: Configure options --enable-setdevice=no OPENMPI detected AcceleratorCudaInit[0]: ======================== AcceleratorCudaInit[0]: Device Number : 0 AcceleratorCudaInit[0]: ======================== AcceleratorCudaInit[0]: Device identifier: NVIDIA A100-SXM4-40GB AcceleratorCudaInit[0]: totalGlobalMem: 42505273344 AcceleratorCudaInit[0]: managedMemory: 1 AcceleratorCudaInit[0]: isMultiGpuBoard: 0 AcceleratorCudaInit[0]: warpSize: 32 AcceleratorCudaInit[0]: pciBusID: 3 AcceleratorCudaInit[0]: pciDeviceID: 0 AcceleratorCudaInit[0]: maxGridSize (2147483647,65535,65535) AcceleratorCudaInit: using default device AcceleratorCudaInit: assume user either uses AcceleratorCudaInit: a) IBM jsrun, or AcceleratorCudaInit: b) invokes through a wrapping script to set CUDA_VISIBLE_DEVICES, UCX_NET_DEVICES, and numa binding AcceleratorCudaInit: Configure options --enable-setdevice=no OPENMPI detected AcceleratorCudaInit: using default device AcceleratorCudaInit: assume user either uses AcceleratorCudaInit: a) IBM jsrun, or AcceleratorCudaInit: b) invokes through a wrapping script to set CUDA_VISIBLE_DEVICES, UCX_NET_DEVICES, and numa binding AcceleratorCudaInit: Configure options --enable-setdevice=no OPENMPI detected AcceleratorCudaInit: using default device AcceleratorCudaInit: assume user either uses AcceleratorCudaInit: a) IBM jsrun, or AcceleratorCudaInit: b) invokes through a wrapping script to set CUDA_VISIBLE_DEVICES, UCX_NET_DEVICES, and numa binding AcceleratorCudaInit: Configure options --enable-setdevice=no OPENMPI detected AcceleratorCudaInit: using default device AcceleratorCudaInit: assume user either uses AcceleratorCudaInit: a) IBM jsrun, or AcceleratorCudaInit: b) invokes through a wrapping script to set CUDA_VISIBLE_DEVICES, UCX_NET_DEVICES, and numa binding AcceleratorCudaInit: Configure options --enable-setdevice=no OPENMPI detected OPENMPI detected AcceleratorCudaInit: using default device AcceleratorCudaInit: assume user either uses AcceleratorCudaInit: a) IBM jsrun, or AcceleratorCudaInit: b) invokes through a wrapping script to set CUDA_VISIBLE_DEVICES, UCX_NET_DEVICES, and numa binding AcceleratorCudaInit: Configure options --enable-setdevice=no OPENMPI detected AcceleratorCudaInit: using default device AcceleratorCudaInit: assume user either uses AcceleratorCudaInit: a) IBM jsrun, or AcceleratorCudaInit: b) invokes through a wrapping script to set CUDA_VISIBLE_DEVICES, UCX_NET_DEVICES, and numa binding AcceleratorCudaInit: Configure options --enable-setdevice=no AcceleratorCudaInit[0]: ======================== AcceleratorCudaInit[0]: Device Number : 0 AcceleratorCudaInit[0]: ======================== AcceleratorCudaInit[0]: Device identifier: NVIDIA A100-SXM4-40GB AcceleratorCudaInit[0]: totalGlobalMem: 42505273344 AcceleratorCudaInit[0]: managedMemory: 1 AcceleratorCudaInit[0]: isMultiGpuBoard: 0 AcceleratorCudaInit[0]: warpSize: 32 AcceleratorCudaInit[0]: pciBusID: 3 AcceleratorCudaInit[0]: pciDeviceID: 0 AcceleratorCudaInit[0]: maxGridSize (2147483647,65535,65535) AcceleratorCudaInit: using default device AcceleratorCudaInit: assume user either uses AcceleratorCudaInit: a) IBM jsrun, or AcceleratorCudaInit: b) invokes through a wrapping script to set CUDA_VISIBLE_DEVICES, UCX_NET_DEVICES, and numa binding AcceleratorCudaInit: Configure options --enable-setdevice=no local rank 1 device 0 bus id: 0000:44:00.0 AcceleratorCudaInit: ================================================ local rank 0 device 0 bus id: 0000:03:00.0 AcceleratorCudaInit: ================================================ AcceleratorCudaInit: ================================================ AcceleratorCudaInit: ================================================ AcceleratorCudaInit: ================================================ AcceleratorCudaInit: ================================================ local rank 0 device 0 bus id: 0000:03:00.0 AcceleratorCudaInit: ================================================ AcceleratorCudaInit: ================================================ local rank 2 device 0 bus id: 0000:84:00.0 SharedMemoryMpi: World communicator of size 16 SharedMemoryMpi: Node communicator of size 4 0SharedMemoryMpi: SharedMemoryMPI.cc acceleratorAllocDevice 2147483648bytes at 0x153960000000 for comms buffers Setting up IPC __|__|__|__|__|__|__|__|__|__|__|__|__|__|__ __|__|__|__|__|__|__|__|__|__|__|__|__|__|__ __|_ | | | | | | | | | | | | _|__ __|_ _|__ __|_ GGGG RRRR III DDDD _|__ __|_ G R R I D D _|__ __|_ G R R I D D _|__ __|_ G GG RRRR I D D _|__ __|_ G G R R I D D _|__ __|_ GGGG R R III DDDD _|__ __|_ _|__ __|__|__|__|__|__|__|__|__|__|__|__|__|__|__ __|__|__|__|__|__|__|__|__|__|__|__|__|__|__ | | | | | | | | | | | | | | Copyright (C) 2015 Peter Boyle, Azusa Yamaguchi, Guido Cossu, Antonin Portelli and other authors This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. Current Grid git commit hash=da06d15f73184ceb15d66d4e7e702b02fed7b940: (HEAD -> feature/dirichlet, develop) uncommited changes Grid : Message : ================================================ Grid : Message : MPI is initialised and logging filters activated Grid : Message : ================================================ Grid : Message : Requested 2147483648 byte stencil comms buffers Grid : Message : MemoryManager Cache 34004218675 bytes Grid : Message : MemoryManager::Init() setting up Grid : Message : MemoryManager::Init() cache pool for recent allocations: SMALL 8 LARGE 2 Grid : Message : MemoryManager::Init() Non unified: Caching accelerator data in dedicated memory Grid : Message : MemoryManager::Init() Using cudaMalloc Grid : Message : 1.875883 s : Grid Layout Grid : Message : 1.875893 s : Global lattice size : 64 64 64 64 Grid : Message : 1.875897 s : OpenMP threads : 4 Grid : Message : 1.875898 s : MPI tasks : 2 2 2 2 Grid : Message : 1.993571 s : Initialising 4d RNG Grid : Message : 2.881990 s : Intialising parallel RNG with unique string 'The 4D RNG' Grid : Message : 2.882370 s : Seed SHA256: 49db4542db694e3b1a74bf2592a8c1b83bfebbe18401693c2609a4c3af1 Grid : Message : 2.495044 s : Initialising 5d RNG Grid : Message : 4.120900 s : Intialising parallel RNG with unique string 'The 5D RNG' Grid : Message : 4.121350 s : Seed SHA256: b6316f2fac44ce14111f93e0296389330b077bfd0a7b359f781c58589f8a Grid : Message : 15.268010 s : Drawing gauge field Grid : Message : 16.234025 s : Random gauge initialised Grid : Message : 16.234057 s : Applying BCs Grid : Message : 16.365565 s : Setting up Cshift based reference Grid : Message : 44.512418 s : ***************************************************************** Grid : Message : 44.512448 s : * Kernel options --dslash-generic, --dslash-unroll, --dslash-asm Grid : Message : 44.512450 s : ***************************************************************** Grid : Message : 44.512451 s : ***************************************************************** Grid : Message : 44.512452 s : * Benchmarking DomainWallFermionR::Dhop Grid : Message : 44.512453 s : * Vectorising space-time by 8 Grid : Message : 44.512454 s : * VComplexF size is 64 B Grid : Message : 44.512456 s : * SINGLE precision Grid : Message : 44.512459 s : * Using Overlapped Comms/Compute Grid : Message : 44.512460 s : * Using GENERIC Nc WilsonKernels Grid : Message : 44.512461 s : ***************************************************************** Grid : Message : 46.389070 s : Called warmup Grid : Message : 49.211265 s : Called Dw 300 times in 2.82203e+06 us Grid : Message : 49.211295 s : mflop/s = 3.76681e+07 Grid : Message : 49.211297 s : mflop/s per rank = 2.35425e+06 Grid : Message : 49.211299 s : mflop/s per node = 9.41702e+06 Grid : Message : 49.211301 s : RF GiB/s (base 2) = 76540.6 Grid : Message : 49.211308 s : mem GiB/s (base 2) = 47837.9 Grid : Message : 49.214868 s : norm diff 1.06409e-13 Grid : Message : 92.647781 s : Compare to naive wilson implementation Dag to verify correctness Grid : Message : 92.647816 s : Called DwDag Grid : Message : 92.647817 s : norm dag result 12.0421 Grid : Message : 92.801806 s : norm dag ref 12.0421 Grid : Message : 92.817724 s : norm dag diff 7.21921e-14 Grid : Message : 92.858973 s : Calling Deo and Doe and //assert Deo+Doe == Dunprec Grid : Message : 93.210378 s : src_e0.499997 Grid : Message : 93.583286 s : src_o0.500003 Grid : Message : 93.682468 s : ********************************************************* Grid : Message : 93.682471 s : * Benchmarking DomainWallFermionF::DhopEO Grid : Message : 93.682472 s : * Vectorising space-time by 8 Grid : Message : 93.682473 s : * SINGLE precision Grid : Message : 93.682475 s : * Using Overlapped Comms/Compute Grid : Message : 93.682476 s : * Using GENERIC Nc WilsonKernels Grid : Message : 93.682477 s : ********************************************************* Grid : Message : 95.162342 s : Deo mflop/s = 3.92487e+07 Grid : Message : 95.162387 s : Deo mflop/s per rank 2.45305e+06 Grid : Message : 95.162389 s : Deo mflop/s per node 9.81219e+06 Grid : Message : 95.232801 s : r_e6.02111 Grid : Message : 95.240061 s : r_o6.02102 Grid : Message : 95.245975 s : res12.0421 Grid : Message : 95.833402 s : norm diff 0 Grid : Message : 96.573829 s : norm diff even 0 Grid : Message : 96.868272 s : norm diff odd 0 Dirichlet block [0 64 64 32 32] Grid : Message : 97.756909 s : Grid Layout Grid : Message : 97.756911 s : Global lattice size : 64 64 64 64 Grid : Message : 97.756921 s : OpenMP threads : 4 Grid : Message : 97.756922 s : MPI tasks : 2 2 2 2 Grid : Message : 97.897085 s : Initialising 4d RNG Grid : Message : 97.965061 s : Intialising parallel RNG with unique string 'The 4D RNG' Grid : Message : 97.965097 s : Seed SHA256: 49db4542db694e3b1a74bf2592a8c1b83bfebbe18401693c2609a4c3af1 Grid : Message : 98.367431 s : Initialising 5d RNG Grid : Message : 99.752745 s : Intialising parallel RNG with unique string 'The 5D RNG' Grid : Message : 99.752790 s : Seed SHA256: b6316f2fac44ce14111f93e0296389330b077bfd0a7b359f781c58589f8a Grid : Message : 111.290148 s : Drawing gauge field Grid : Message : 112.349289 s : Random gauge initialised Grid : Message : 112.349320 s : Applying BCs Grid : Message : 113.948740 s : Setting up Cshift based reference Grid : Message : 140.320415 s : ***************************************************************** Grid : Message : 140.320443 s : * Kernel options --dslash-generic, --dslash-unroll, --dslash-asm Grid : Message : 140.320444 s : ***************************************************************** Grid : Message : 140.320445 s : ***************************************************************** Grid : Message : 140.320446 s : * Benchmarking DomainWallFermionR::Dhop Grid : Message : 140.320447 s : * Vectorising space-time by 8 Grid : Message : 140.320448 s : * VComplexF size is 64 B Grid : Message : 140.320450 s : * SINGLE precision Grid : Message : 140.320451 s : * Using Overlapped Comms/Compute Grid : Message : 140.320452 s : * Using GENERIC Nc WilsonKernels Grid : Message : 140.320453 s : ***************************************************************** Grid : Message : 142.296150 s : Called warmup Grid : Message : 144.397678 s : Called Dw 300 times in 2.36719e+06 us Grid : Message : 144.397700 s : mflop/s = 4.49058e+07 Grid : Message : 144.397702 s : mflop/s per rank = 2.80661e+06 Grid : Message : 144.397704 s : mflop/s per node = 1.12265e+07 Grid : Message : 144.397706 s : RF GiB/s (base 2) = 91247.6 Grid : Message : 144.397708 s : mem GiB/s (base 2) = 57029.7 Grid : Message : 144.401269 s : norm diff 9.78944e-14 Grid : Message : 186.885460 s : Compare to naive wilson implementation Dag to verify correctness Grid : Message : 186.885492 s : Called DwDag Grid : Message : 186.885493 s : norm dag result 10.4157 Grid : Message : 186.897154 s : norm dag ref 11.2266 Grid : Message : 186.912538 s : norm dag diff 0.484633