diff options
| author |  Rasmus Larsen <rmlarsen@google.com> | 2019-12-04 00:45:42 +0000 |
|---|---|---|
| committer | Rasmus Larsen <rmlarsen@google.com> | 2019-12-04 00:45:42 +0000 |
| commit | cacf433975fa8f9fc01e926cdd001c5dab31210e (patch) | |
| tree | abb2e7c78e302dde0760755a8a834e6796040bdf /Eigen/src/Core/util | |
| parent | 4e696901f873a2347f76d931cf2f701e31e15d05 (diff) | |
| parent | c1a67cb5aff74094ce8061626074f2d184316c52 (diff) | |
Merged in anshuljl/eigen-2/Anshul-Jaiswal/update-configurevectorizationh-to-not-op-1573079916090 (pull request PR-754)
Update ConfigureVectorization.h to not optimize fp16 routines when compiling with cuda.
Approved-by: Deven Desai <deven.desai.amd@gmail.com>
Diffstat (limited to 'Eigen/src/Core/util')
| -rw-r--r-- | Eigen/src/Core/util/ConfigureVectorization.h | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/Eigen/src/Core/util/ConfigureVectorization.h b/Eigen/src/Core/util/ConfigureVectorization.h index d52805d32..271795a06 100644 --- a/Eigen/src/Core/util/ConfigureVectorization.h +++ b/Eigen/src/Core/util/ConfigureVectorization.h @@ -411,7 +411,7 @@ #endif #endif -#if defined(__F16C__) && (!defined(EIGEN_COMP_CLANG) || EIGEN_COMP_CLANG>=380) +#if defined(__F16C__) && (!defined(EIGEN_GPUCC) && (!defined(EIGEN_COMP_CLANG) || EIGEN_COMP_CLANG>=380)) // We can use the optimized fp16 to float and float to fp16 conversion routines #define EIGEN_HAS_FP16_C |