diff options
| author |  A. Unique TensorFlower <gardener@tensorflow.org> | 2016-10-11 20:46:35 -0800 |
|---|---|---|
| committer | TensorFlower Gardener <gardener@tensorflow.org> | 2016-10-11 22:05:10 -0700 |
| commit | 6b905241ecd462f1629ed24d0f9b3746187d5933 (patch) | |
| tree | bf16e86d2da938097a0a498aa7fb5b4a38eebf21 /tensorflow/core/kernels/aggregate_ops.cc | |
| parent | cb57912ac43c29c23ec24a36eaba50a01bef4b9a (diff) | |
Enables AddN for float64.
Change: 135880018
Diffstat (limited to 'tensorflow/core/kernels/aggregate_ops.cc')
| -rw-r--r-- | tensorflow/core/kernels/aggregate_ops.cc | 1 |
1 files changed, 1 insertions, 0 deletions
diff --git a/tensorflow/core/kernels/aggregate_ops.cc b/tensorflow/core/kernels/aggregate_ops.cc index f9a7b3ee52..b41e438b2b 100644 --- a/tensorflow/core/kernels/aggregate_ops.cc +++ b/tensorflow/core/kernels/aggregate_ops.cc @@ -139,6 +139,7 @@ TF_CALL_NUMBER_TYPES(REGISTER_ADDN_CPU); #if GOOGLE_CUDA REGISTER_ADDN(Eigen::half, GPU); REGISTER_ADDN(float, GPU); +REGISTER_ADDN(double, GPU); // A special GPU kernel for int32. // TODO(b/25387198): Also enable int32 in device memory. This kernel |