Initial XLA support for TF eager. This is prerequisite for TF compiler's XLA support.
authorMingsheng Hong <hongm@google.com>
Wed, 7 Feb 2018 20:22:55 +0000 (12:22 -0800)
committerTensorFlower Gardener <gardener@tensorflow.org>
Wed, 7 Feb 2018 20:26:52 +0000 (12:26 -0800)
commit65f676426d808cf20abd6a4a30ad48668ee5fdf2
tree84707b8b38781cb4703e46656f24737d86ac1d2d
parent90ce80131a8b5213d9f3eb9649d63921db7874a4
Initial XLA support for TF eager. This is prerequisite for TF compiler's XLA support.

This CL adds XLA support for the following TFE_Op's:

1. A TF op such as MatMul, with full support of constant and resource params.

2. A TF_Function as TFE_Op, where the function must have no constant and
resource params. Removing this restriction requires more discussion and will be
deferred to a later time.

PiperOrigin-RevId: 184877345
tensorflow/c/BUILD
tensorflow/c/c_test_util.cc
tensorflow/c/c_test_util.h
tensorflow/c/eager/BUILD
tensorflow/c/eager/c_api.cc
tensorflow/c/eager/c_api.h
tensorflow/c/eager/c_api_internal.h
tensorflow/c/eager/c_api_test.cc