提交 13024c58 编写于 作者: M Manjunath Kudlur 提交者: TensorFlower Gardener

Changed the cuda_crosstool_condition to check for a define of using_cuda.

Checking for a specific crosstool_top directory doesn't work when TensorFlow
is a sub-module for a different project.
Change: 116592676
上级 d82cc200
......@@ -10,14 +10,14 @@ package(default_visibility = ["//visibility:public"])
config_setting(
name = "cuda_crosstool_condition",
values = {"crosstool_top": "//third_party/gpus/crosstool"},
values = {"define": "using_cuda=true"},
visibility = ["//visibility:public"],
)
config_setting(
name = "using_gcudacc",
values = {
"crosstool_top": "//third_party/gpus/crosstool",
"define": "using_cuda=true",
"copt": "--use_gcudacc",
},
visibility = ["//visibility:public"],
......@@ -26,7 +26,7 @@ config_setting(
config_setting(
name = "using_nvcc",
values = {
"crosstool_top": "//third_party/gpus/crosstool",
"define": "using_cuda=true",
"copt": "--use_nvcc",
},
)
......
build:cuda --crosstool_top=//third_party/gpus/crosstool
build:cuda --define=using_cuda=true
build --force_python=py$PYTHON_MAJOR_VERSION
build --python$PYTHON_MAJOR_VERSION_path=$PYTHON_BINARY
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册