experiments/STT-tensorflow
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

[XLA:GPU] Add sm_75 to LLVM supported procs list

Browse Source 
- This is required to prevent fallback to the default sm version when
  running on GPUs with the Turing architecture.
This commit is contained in:
Ben Barsdell 2018-12-13 15:09:54 -08:00
parent 1182414979
commit 1674bd08ce
1 changed files with 1 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
tensorflow/compiler/xla/service/gpu/llvm_gpu_backend/nvptx_backend_lib.cc
View File

@ -125,6 +125,7 @@ static string GetSmName(std::pair<int, int> compute_capability) {
{{6, 2}, 62}, {{6, 2}, 62},
{{7, 0}, 70}, {{7, 0}, 70},
{{7, 2}, 72}, {{7, 2}, 72},
{{7, 5}, 75},
}); });
int sm_version = 30; int sm_version = 30;
auto it = m->find(compute_capability); auto it = m->find(compute_capability);