Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Added L0 diagnostics, corrected scaling test, added temperature test,…
… halfduplex measures host-device bandwidth
- Loading branch information
Showing
8 changed files
with
382 additions
and
72 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,142 @@ | ||
| // MGBench: Multi-GPU Computing Benchmark Suite | ||
| // Copyright (c) 2016, Tal Ben-Nun | ||
| // All rights reserved. | ||
| // | ||
| // Redistribution and use in source and binary forms, with or without | ||
| // modification, are permitted provided that the following conditions are met: | ||
| // | ||
| // * Redistributions of source code must retain the above copyright notice, | ||
| // this list of conditions and the following disclaimer. | ||
| // * Redistributions in binary form must reproduce the above copyright notice, | ||
| // this list of conditions and the following disclaimer in the documentation | ||
| // and/or other materials provided with the distribution. | ||
| // * Neither the names of the copyright holders nor the names of its | ||
| // contributors may be used to endorse or promote products derived from this | ||
| // software without specific prior written permission. | ||
| // | ||
| // THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" | ||
| // AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE | ||
| // IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE | ||
| // ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE | ||
| // LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR | ||
| // CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF | ||
| // SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS | ||
| // INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN | ||
| // CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) | ||
| // ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE | ||
| // POSSIBILITY OF SUCH DAMAGE. | ||
|
|
||
| #include <cstdio> | ||
| #include <cstdlib> | ||
| #include <iostream> | ||
|
|
||
| #include <cuda_runtime.h> | ||
|
|
||
| static void HandleError(const char *file, int line, cudaError_t err) | ||
| { | ||
| printf("ERROR in %s:%d: %s (%d)\n", file, line, | ||
| cudaGetErrorString(err), err); | ||
| cudaGetLastError(); | ||
| } | ||
|
|
||
| // CUDA assertions | ||
| #define CUDA_CHECK(err) do { cudaError_t errr = (err); if(errr != cudaSuccess) { HandleError(__FILE__, __LINE__, errr); exit(1); } } while(0) | ||
|
|
||
| // Device capability helper macros | ||
| #define CAP(cap) NCAP(cap, cap) | ||
| #define NCAP(cap, name) ((props.cap) ? (#name " ") : "") | ||
|
|
||
| int main(int argc, char **argv) | ||
| { | ||
| int ndevs = 0; | ||
| if (cudaGetDeviceCount(&ndevs) != cudaSuccess) | ||
| return 1; | ||
|
|
||
| int version = 0; | ||
| CUDA_CHECK(cudaDriverGetVersion(&version)); | ||
| std::cout << "Driver version: " << (version / 1000) << "." | ||
| << ((version % 100) / 10) << std::endl; | ||
|
|
||
| version = 0; | ||
| CUDA_CHECK(cudaRuntimeGetVersion(&version)); | ||
| std::cout << "Runtime version: " << (version / 1000) << "." | ||
| << ((version % 100) / 10) << std::endl; | ||
| std::cout << std::endl; | ||
|
|
||
| // Print information for each GPU | ||
| for (int i = 0; i < ndevs; ++i) | ||
| { | ||
| CUDA_CHECK(cudaSetDevice(i)); | ||
| cudaDeviceProp props; | ||
| CUDA_CHECK(cudaGetDeviceProperties(&props, i)); | ||
|
|
||
| std::cout << "GPU " << (i + 1) << ": " << props.name << " (" | ||
| << props.pciDomainID << "/" << props.pciBusID | ||
| << "/" << props.pciDeviceID << ")" << std::endl | ||
|
|
||
|
|
||
| << "Global memory: " << (props.totalGlobalMem/1024.0/1024.0) | ||
| << " MB" << std::endl | ||
| << "Constant memory: " << props.totalConstMem | ||
| << " bytes" << std::endl | ||
| << "Shared memory: " << props.sharedMemPerBlock | ||
| << " bytes" << std::endl | ||
| << "Registers: " << props.regsPerBlock << std::endl | ||
| << "Warp size: " << props.warpSize << std::endl | ||
| << "Multiprocessors: " << props.multiProcessorCount | ||
| << std::endl | ||
| << "Copy engines: " << props.asyncEngineCount << std::endl | ||
| << "Clock rate: " << (props.clockRate / 1e6) | ||
| << " GHz" << std::endl | ||
| << "Threads per MP: " << props.maxThreadsPerMultiProcessor | ||
| << std::endl | ||
| << "Threads per block: " << props.maxThreadsPerBlock | ||
| << std::endl | ||
| << "Max block size: " << props.maxThreadsDim[0] << "x" | ||
| << props.maxThreadsDim[1] << "x" << props.maxThreadsDim[2] | ||
| << std::endl | ||
| << "Max grid size: " << props.maxGridSize[0] << "x" | ||
| << props.maxGridSize[1] << "x" << props.maxGridSize[2] | ||
| << std::endl | ||
| << "Pitch: " << props.memPitch << " bytes" << std::endl; | ||
|
|
||
| std::cout << "Caps: " << NCAP(ECCEnabled, ecc) | ||
| << NCAP(deviceOverlap, overlap) | ||
| << NCAP(unifiedAddressing, uva) | ||
| << NCAP(kernelExecTimeoutEnabled, timeout) | ||
| << CAP(integrated) << NCAP(canMapHostMemory, hostdma) | ||
| << CAP(surfaceAlignment) << CAP(tccDriver) << std::endl; | ||
| std::cout << std::endl; | ||
| } | ||
|
|
||
| std::cout << "DMA access: " << std::endl; | ||
| int tmp = 0; | ||
|
|
||
| // Print top row | ||
| printf(" | "); | ||
| for (int i = 0; i < ndevs; ++i) | ||
| printf("%2d ", i + 1); | ||
| printf("\n---+"); | ||
| for (int i = 0; i < ndevs; ++i) | ||
| printf("---"); | ||
| printf("\n"); | ||
|
|
||
| for (int i = 0; i < ndevs; ++i) | ||
| { | ||
| printf("%2d | ", i + 1); | ||
| for (int j = 0; j < ndevs; ++j) | ||
| { | ||
| if (i == j) | ||
| { | ||
| printf(" x "); | ||
| continue; | ||
| } | ||
|
|
||
| cudaDeviceCanAccessPeer(&tmp, i, j); | ||
| printf("%2d ", tmp ? 1 : 0); | ||
| } | ||
| printf("\n"); | ||
| } | ||
|
|
||
| return 0; | ||
| } |
File renamed without changes.
Oops, something went wrong.