tu-c0r1n24 - 0 device=0 binding=--interleave=0,1 tu-c0r1n24 - 1 device=1 binding=--interleave=2,3 tu-c0r1n24 - 2 device=2 binding=--interleave=4,5 tu-c0r1n24 - 3 device=3 binding=--interleave=6,7 tu-c0r2n33 - 0 device=0 binding=--interleave=0,1 tu-c0r2n33 - 3 device=3 binding=--interleave=6,7 tu-c0r2n33 - 1 device=1 binding=--interleave=2,3 tu-c0r2n27 - 3 device=3 binding=--interleave=6,7 tu-c0r2n27 - 1 device=1 binding=--interleave=2,3 tu-c0r2n33 - 2 device=2 binding=--interleave=4,5 tu-c0r2n27 - 2 device=2 binding=--interleave=4,5 tu-c0r2n27 - 0 device=0 binding=--interleave=0,1 tu-c0r1n66 - 0 device=0 binding=--interleave=0,1 tu-c0r2n66 - 3 device=3 binding=--interleave=6,7 tu-c0r2n66 - 1 device=1 binding=--interleave=2,3 tu-c0r1n66 - 1 device=1 binding=--interleave=2,3 tu-c0r2n30 - 2 device=2 binding=--interleave=4,5 tu-c0r1n66 - 2 device=2 binding=--interleave=4,5 tu-c0r1n66 - 3 device=3 binding=--interleave=6,7 tu-c0r2n60 - 3 device=3 binding=--interleave=6,7 tu-c0r2n66 - 0 device=0 binding=--interleave=0,1 tu-c0r2n66 - 2 device=2 binding=--interleave=4,5 tu-c0r2n30 - 3 device=3 binding=--interleave=6,7 tu-c0r2n30 - 1 device=1 binding=--interleave=2,3 tu-c0r2n30 - 0 device=0 binding=--interleave=0,1 tu-c0r2n24 - 0 device=0 binding=--interleave=0,1 tu-c0r2n24 - 1 device=1 binding=--interleave=2,3 tu-c0r2n60 - 0 device=0 binding=--interleave=0,1 tu-c0r2n24 - 3 device=3 binding=--interleave=6,7 tu-c0r2n60 - 2 device=2 binding=--interleave=4,5 tu-c0r2n60 - 1 device=1 binding=--interleave=2,3 tu-c0r2n24 - 2 device=2 binding=--interleave=4,5 tu-c0r1n27 - 1 device=1 binding=--interleave=2,3 tu-c0r1n27 - 2 device=2 binding=--interleave=4,5 tu-c0r1n27 - 0 device=0 binding=--interleave=0,1 tu-c0r1n27 - 3 device=3 binding=--interleave=6,7 tu-c0r1n69 - 1 device=1 binding=--interleave=2,3 tu-c0r1n69 - 0 device=0 binding=--interleave=0,1 tu-c0r1n69 - 3 device=3 binding=--interleave=6,7 tu-c0r1n69 - 2 device=2 binding=--interleave=4,5 tu-c0r2n63 - 3 device=3 binding=--interleave=6,7 tu-c0r2n63 - 2 device=2 binding=--interleave=4,5 tu-c0r1n63 - 0 device=0 binding=--interleave=0,1 tu-c0r1n63 - 1 device=1 binding=--interleave=2,3 tu-c0r2n63 - 1 device=1 binding=--interleave=2,3 tu-c0r2n63 - 0 device=0 binding=--interleave=0,1 tu-c0r1n63 - 2 device=2 binding=--interleave=4,5 tu-c0r1n63 - 3 device=3 binding=--interleave=6,7 tu-c0r2n69 - 2 device=2 binding=--interleave=4,5 tu-c0r2n69 - 3 device=3 binding=--interleave=6,7 tu-c0r2n69 - 0 device=0 binding=--interleave=0,1 tu-c0r2n69 - 1 device=1 binding=--interleave=2,3 tu-c0r1n60 - 3 device=3 binding=--interleave=6,7 tu-c0r1n60 - 0 device=0 binding=--interleave=0,1 tu-c0r1n60 - 1 device=1 binding=--interleave=2,3 tu-c0r1n60 - 2 device=2 binding=--interleave=4,5 tu-c0r1n30 - 1 device=1 binding=--interleave=2,3 tu-c0r1n30 - 0 device=0 binding=--interleave=0,1 tu-c0r1n30 - 3 device=3 binding=--interleave=6,7 tu-c0r1n30 - 2 device=2 binding=--interleave=4,5 tu-c0r1n33 - 0 device=0 binding=--interleave=0,1 tu-c0r1n33 - 3 device=3 binding=--interleave=6,7 tu-c0r1n33 - 2 device=2 binding=--interleave=4,5 tu-c0r1n33 - 1 device=1 binding=--interleave=2,3 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[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 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: 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: ================================================ local rank 0 device 0 bus id: 0000:03:00.0 AcceleratorCudaInit: ================================================ AcceleratorCudaInit: ================================================ AcceleratorCudaInit: ================================================ local rank 0 device 0 bus id: 0000:03:00.0 AcceleratorCudaInit: ================================================ AcceleratorCudaInit: ================================================ AcceleratorCudaInit: ================================================ AcceleratorCudaInit: ================================================ local rank 3 device 0 bus id: 0000:C4:00.0 local rank 1 device 0 bus id: 0000:44:00.0 local rank 2 device 0 bus id: 0000:84:00.0 SharedMemoryMpi: World communicator of size 64 SharedMemoryMpi: Node communicator of size 4 0SharedMemoryMpi: SharedMemoryMPI.cc acceleratorAllocDevice 2147483648bytes at 0x148160000000 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=188d2c7a4dc77807b545f5f2813cdb589b9e44ca: (HEAD -> develop, gh/develop, gh/HEAD) 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 : 1.660594 s : Grid Layout Grid : Message : 1.660598 s : Global lattice size : 64 64 64 256 Grid : Message : 1.660603 s : OpenMP threads : 4 Grid : Message : 1.660604 s : MPI tasks : 2 2 2 8 Grid : Message : 1.699865 s : Making s innermost grids Grid : Message : 1.758720 s : Initialising 4d RNG Grid : Message : 1.851014 s : Intialising parallel RNG with unique string 'The 4D RNG' Grid : Message : 1.851045 s : Seed SHA256: 49db4542db694e3b1a74bf2592a8c1b83bfebbe18401693c2609a4c3af1 Grid : Message : 3.428220 s : Initialising 5d RNG Grid : Message : 4.860765 s : Intialising parallel RNG with unique string 'The 5D RNG' Grid : Message : 4.860797 s : Seed SHA256: b6316f2fac44ce14111f93e0296389330b077bfd0a7b359f781c58589f8a Grid : Message : 38.127813 s : Initialised RNGs Grid : Message : 42.756686 s : Drawing gauge field Grid : Message : 43.389158 s : Random gauge initialised Grid : Message : 43.400941 s : Setting up Cshift based reference Grid : Message : 72.593110 s : ***************************************************************** Grid : Message : 72.593135 s : * Kernel options --dslash-generic, --dslash-unroll, --dslash-asm Grid : Message : 72.593137 s : ***************************************************************** Grid : Message : 72.593138 s : ***************************************************************** Grid : Message : 72.593139 s : * Benchmarking DomainWallFermionR::Dhop Grid : Message : 72.593140 s : * Vectorising space-time by 8 Grid : Message : 72.593141 s : * VComplexF size is 64 B Grid : Message : 72.593142 s : * SINGLE precision Grid : Message : 72.593146 s : * Using Overlapped Comms/Compute Grid : Message : 72.593147 s : * Using GENERIC Nc WilsonKernels Grid : Message : 72.593148 s : ***************************************************************** Grid : Message : 74.591016 s : Called warmup Grid : Message : 354.958828 s : Called Dw 30000 times in 2.80365e+08 us Grid : Message : 354.958880 s : mflop/s = 1.5166e+08 Grid : Message : 354.958882 s : mflop/s per rank = 2.36969e+06 Grid : Message : 354.958884 s : mflop/s per node = 9.47875e+06 Grid : Message : 354.958886 s : RF GiB/s (base 2) = 308169 Grid : Message : 354.958888 s : mem GiB/s (base 2) = 192606 Grid : Message : 354.962395 s : norm diff 1.06407e-13 Grid : Message : 355.127390 s : #### Dhop calls report Grid : Message : 355.127550 s : WilsonFermion5D Number of DhopEO Calls : 60002 Grid : Message : 355.127590 s : WilsonFermion5D TotalTime /Calls : 4675.78 us Grid : Message : 355.127610 s : WilsonFermion5D CommTime /Calls : 3142.16 us Grid : Message : 355.127630 s : WilsonFermion5D FaceTime /Calls : 468.717 us Grid : Message : 355.127650 s : WilsonFermion5D ComputeTime1/Calls : 5.03673 us Grid : Message : 355.127670 s : WilsonFermion5D ComputeTime2/Calls : 1078.94 us Grid : Message : 355.128060 s : Average mflops/s per call : 1.25821e+11 Grid : Message : 355.128110 s : Average mflops/s per call per rank : 1.96595e+09 Grid : Message : 355.128140 s : Average mflops/s per call per node : 7.86381e+09 Grid : Message : 355.128170 s : Average mflops/s per call (full) : 1.54318e+08 Grid : Message : 355.128200 s : Average mflops/s per call per rank (full): 2.41121e+06 Grid : Message : 355.128230 s : Average mflops/s per call per node (full): 9.64485e+06 Grid : Message : 355.128270 s : WilsonFermion5D Stencil Grid : Message : 355.128290 s : WilsonFermion5D StencilEven Grid : Message : 355.128310 s : WilsonFermion5D StencilOdd Grid : Message : 355.128330 s : WilsonFermion5D Stencil Reporti() Grid : Message : 355.128350 s : WilsonFermion5D StencilEven Reporti() Grid : Message : 355.128370 s : WilsonFermion5D StencilOdd Reporti() Grid : Message : 410.262132 s : Compare to naive wilson implementation Dag to verify correctness Grid : Message : 410.262157 s : Called DwDag Grid : Message : 410.262158 s : norm dag result 12.0421 Grid : Message : 410.275556 s : norm dag ref 12.0421 Grid : Message : 410.291310 s : norm dag diff 7.21924e-14 Grid : Message : 410.332108 s : Calling Deo and Doe and //assert Deo+Doe == Dunprec Grid : Message : 410.709082 s : src_e0.499998 Grid : Message : 411.157130 s : src_o0.500002 Grid : Message : 411.256016 s : ********************************************************* Grid : Message : 411.256025 s : * Benchmarking DomainWallFermionF::DhopEO Grid : Message : 411.256027 s : * Vectorising space-time by 8 Grid : Message : 411.256029 s : * SINGLE precision Grid : Message : 411.256031 s : * Using Overlapped Comms/Compute Grid : Message : 411.256033 s : * Using GENERIC Nc WilsonKernels Grid : Message : 411.256036 s : ********************************************************* Grid : Message : 548.708437 s : Deo mflop/s = 1.54735e+08 Grid : Message : 548.708469 s : Deo mflop/s per rank 2.41774e+06 Grid : Message : 548.708476 s : Deo mflop/s per node 9.67094e+06 Grid : Message : 548.708479 s : #### Dhop calls report Grid : Message : 548.708482 s : WilsonFermion5D Number of DhopEO Calls : 30001 Grid : Message : 548.708485 s : WilsonFermion5D TotalTime /Calls : 4581.46 us Grid : Message : 548.708488 s : WilsonFermion5D CommTime /Calls : 3007.72 us Grid : Message : 548.708491 s : WilsonFermion5D FaceTime /Calls : 570.745 us Grid : Message : 548.708493 s : WilsonFermion5D ComputeTime1/Calls : 6.29339 us Grid : Message : 548.708495 s : WilsonFermion5D ComputeTime2/Calls : 1029.87 us Grid : Message : 548.708520 s : Average mflops/s per call : 1.06812e+11 Grid : Message : 548.708524 s : Average mflops/s per call per rank : 1.66894e+09 Grid : Message : 548.708526 s : Average mflops/s per call per node : 6.67577e+09 Grid : Message : 548.708529 s : Average mflops/s per call (full) : 1.57494e+08 Grid : Message : 548.708532 s : Average mflops/s per call per rank (full): 2.46085e+06 Grid : Message : 548.708535 s : Average mflops/s per call per node (full): 9.8434e+06 Grid : Message : 548.708539 s : WilsonFermion5D Stencil Grid : Message : 548.708540 s : WilsonFermion5D StencilEven Grid : Message : 548.708542 s : WilsonFermion5D StencilOdd Grid : Message : 548.708544 s : WilsonFermion5D Stencil Reporti() Grid : Message : 548.708546 s : WilsonFermion5D StencilEven Reporti() Grid : Message : 548.708549 s : WilsonFermion5D StencilOdd Reporti() Grid : Message : 548.779742 s : r_e6.02108 Grid : Message : 548.786659 s : r_o6.02101 Grid : Message : 548.793024 s : res12.0421 Grid : Message : 549.458970 s : norm diff 0 Grid : Message : 550.459127 s : norm diff even 0 Grid : Message : 550.902468 s : norm diff odd 0