diff options
author | Deven Desai <deven.desai.amd@gmail.com> | 2018-11-19 18:13:59 +0000 |
---|---|---|
committer | Deven Desai <deven.desai.amd@gmail.com> | 2018-11-19 18:13:59 +0000 |
commit | e7e6809e6b38a5928efc0b5ca9520258e4d1fb3a (patch) | |
tree | 500d4209bdf3236178e5d9e9c7a6051d9c2d02d2 /unsupported/Eigen/CXX11/src/Tensor/TensorDeviceThreadPool.h | |
parent | 6a510fe69c3d8ec0cdfa3e0f54a68c07ede68620 (diff) |
ROCm/HIP specfic fixes + updates
1. Eigen/src/Core/arch/GPU/Half.h
Updating the HIPCC implementation half so that it can declared as a __shared__ variable
2. Eigen/src/Core/util/Macros.h, Eigen/src/Core/util/Memory.h
introducing a EIGEN_USE_STD(func) macro that calls
- std::func be default
- ::func when eigen is being compiled with HIPCC
This change was requested in the previous HIP PR
(https://bitbucket.org/eigen/eigen/pull-requests/518/pr-with-hip-specific-fixes-for-the-eigen/diff)
3. unsupported/Eigen/CXX11/src/Tensor/TensorDeviceThreadPool.h
Removing EIGEN_DEVICE_FUNC attribute from pure virtual methods as it is not supported by HIPCC
4. unsupported/Eigen/CXX11/src/Tensor/TensorReduction.h
Disabling the template specializations of InnerMostDimReducer as they run into HIPCC link errors
Diffstat (limited to 'unsupported/Eigen/CXX11/src/Tensor/TensorDeviceThreadPool.h')
-rw-r--r-- | unsupported/Eigen/CXX11/src/Tensor/TensorDeviceThreadPool.h | 4 |
1 files changed, 2 insertions, 2 deletions
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorDeviceThreadPool.h b/unsupported/Eigen/CXX11/src/Tensor/TensorDeviceThreadPool.h index 3b87b114d..bb330a77b 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorDeviceThreadPool.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorDeviceThreadPool.h @@ -45,8 +45,8 @@ static EIGEN_STRONG_INLINE void wait_until_ready(SyncType* n) { class Allocator { public: virtual ~Allocator() {} - EIGEN_DEVICE_FUNC virtual void* allocate(size_t num_bytes) const = 0; - EIGEN_DEVICE_FUNC virtual void deallocate(void* buffer) const = 0; + virtual void* allocate(size_t num_bytes) const = 0; + virtual void deallocate(void* buffer) const = 0; }; // Build a thread pool device on top the an existing pool of threads. |