diff options
| author |  Antonio Sanchez <cantonios@google.com> | 2020-12-01 14:27:52 -0800 |
|---|---|---|
| committer | Antonio Sanchez <cantonios@google.com> | 2020-12-01 14:36:52 -0800 |
| commit | ddd48b242cbf3aa79ad13668b66089cece6d1ea0 (patch) | |
| tree | 2c429ccfd6dbf35843ecc8ff692942e763fb456c /Eigen/src/Core/arch/NEON | |
| parent | e57281a7412f82899cabf63968558b0969d174b6 (diff) | |
Implement CUDA __shfl* for Eigen::half
Prior to this fix, `TensorContractionGpu` and the `cxx11_tensor_of_float16_gpu`
test are broken, as well as several ops in Tensorflow. The gpu functions
`__shfl*` became ambiguous now that `Eigen::half` implicitly converts to float.
Here we add the required specializations.
Diffstat (limited to 'Eigen/src/Core/arch/NEON')
0 files changed, 0 insertions, 0 deletions