Skip to content

Commit b91375a

Browse files
authored
Update CUDA capabilities to match official TF releases (#243)
* Add staging branch to triggers * Change CUDA capabilities for official ones
1 parent 7761384 commit b91375a

File tree

2 files changed

+2
-1
lines changed

2 files changed

+2
-1
lines changed

.github/workflows/ci.yml

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -3,6 +3,7 @@ on:
33
push:
44
branches:
55
- master
6+
- staging
67
- r[0-9]+.*
78
pull_request:
89
branches:

tensorflow-core/tensorflow-core-api/build.sh

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -20,7 +20,7 @@ fi
2020

2121
if [[ "${EXTENSION:-}" == *gpu* ]]; then
2222
export BUILD_FLAGS="$BUILD_FLAGS --config=cuda"
23-
export TF_CUDA_COMPUTE_CAPABILITIES="${TF_CUDA_COMPUTE_CAPABILITIES:-"3.5,7.0"}"
23+
export TF_CUDA_COMPUTE_CAPABILITIES="${TF_CUDA_COMPUTE_CAPABILITIES:-"sm_35,sm_50,sm_60,sm_70,sm_75,compute_80"}"
2424
if [[ -z ${TF_CUDA_PATHS:-} ]] && [[ -d ${CUDA_PATH:-} ]]; then
2525
# Work around some issue with Bazel preventing it from detecting CUDA on Windows
2626
export TF_CUDA_PATHS="$CUDA_PATH"

0 commit comments

Comments
 (0)