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

Enable batch matmul for half types again which was accidentally disabled in CL 197137612.

Browse Source 
PiperOrigin-RevId: 208022865
This commit is contained in:
A. Unique TensorFlower 2018-08-09 03:24:10 -07:00 committed by TensorFlower Gardener
parent e2d9cd3fa7
commit cfdc469565
2 changed files with 3 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
tensorflow/core/kernels/BUILD
View File

@ -2854,6 +2854,8 @@ tf_kernel_library(
srcs = [] + if_mkl([ srcs = [] + if_mkl([
"mkl_batch_matmul_op.cc", "mkl_batch_matmul_op.cc",
]), ]),
# <prefix>*impl.h are excluded by default from the CPU build, add explicitly.
hdrs = ["batch_matmul_op_impl.h"],
# Override EIGEN_STRONG_INLINE to inline when --define=override_eigen_strong_inline=true, # Override EIGEN_STRONG_INLINE to inline when --define=override_eigen_strong_inline=true,
# to avoid long compiling time. See https://github.com/tensorflow/tensorflow/issues/10521 # to avoid long compiling time. See https://github.com/tensorflow/tensorflow/issues/10521
copts = if_override_eigen_strong_inline(["/DEIGEN_STRONG_INLINE=inline"]), copts = if_override_eigen_strong_inline(["/DEIGEN_STRONG_INLINE=inline"]),

3
tensorflow/core/kernels/batch_matmul_op_real.cc
View File

@ -31,8 +31,7 @@ TF_CALL_int32(REGISTER_BATCH_MATMUL_CPU);
#if GOOGLE_CUDA #if GOOGLE_CUDA
TF_CALL_float(REGISTER_BATCH_MATMUL_GPU); TF_CALL_float(REGISTER_BATCH_MATMUL_GPU);
TF_CALL_double(REGISTER_BATCH_MATMUL_GPU); TF_CALL_double(REGISTER_BATCH_MATMUL_GPU);
// TODO(csigg): Implement Stream::ThenBlasGemv for Eigen::half and uncomment. TF_CALL_half(REGISTER_BATCH_MATMUL_GPU);
// TF_CALL_half(REGISTER_BATCH_MATMUL_GPU);
#endif // GOOGLE_CUDA #endif // GOOGLE_CUDA
#ifdef TENSORFLOW_USE_SYCL #ifdef TENSORFLOW_USE_SYCL