devtools/eigen
1
0
Fork 0
mirror of https://gitlab.com/libeigen/eigen.git synced 2026-04-10 11:34:33 +08:00
Code Issues Packages Projects Releases Wiki Activity

[SYCL] This PR adds the minimum modifications to the Eigen unsupported module required to run it on devices supporting SYCL.

Browse Source 
* Abstracting the pointer type so that both SYCL memory and pointer can be captured.
* Converting SYCL virtual pointer to SYCL device memory in Eigen evaluator class.
* Binding SYCL placeholder accessor to command group handler by using bind method in Eigen evaluator node.
* Adding SYCL macro for controlling loop unrolling.
* Modifying the TensorDeviceSycl.h and SYCL executor method to adopt the above changes.
This commit is contained in:
Mehdi Goli
2019-06-28 10:08:23 +01:00
parent 16a56b2ddd
commit 7d08fa805a
47 changed files with 1983 additions and 951 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
unsupported/Eigen/CXX11/src/Tensor/TensorConvolution.h
View File

@@ -303,6 +303,8 @@ struct TensorEvaluator<const TensorConvolutionOp<Indices, InputArgType, KernelAr
typedef typename XprType::CoeffReturnType CoeffReturnType;
typedef typename PacketType<CoeffReturnType, Device>::type PacketReturnType;
static const int PacketSize = PacketType<CoeffReturnType, Device>::size;
typedef StorageMemory<Scalar, Device> Storage;
typedef typename Storage::Type EvaluatorPointerType;
enum {
IsAligned = TensorEvaluator<InputArgType, Device>::IsAligned & TensorEvaluator<KernelArgType, Device>::IsAligned,
@@ -469,7 +471,7 @@ struct TensorEvaluator<const TensorConvolutionOp<Indices, InputArgType, KernelAr
PacketSize));
}
EIGEN_DEVICE_FUNC typename Eigen::internal::traits<XprType>::PointerType data() const { return NULL; }
EIGEN_DEVICE_FUNC EvaluatorPointerType data() const { return NULL; }
private:
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE Index firstInput(Index index) const {
@@ -525,7 +527,7 @@ struct TensorEvaluator<const TensorConvolutionOp<Indices, InputArgType, KernelAr
m_local_kernel = false;
} else {
size_t kernel_sz = m_kernelImpl.dimensions().TotalSize() * sizeof(Scalar);
Scalar* local = (Scalar*)m_device.allocate(kernel_sz);
Scalar* local = (Scalar*)m_device.allocate_temp(kernel_sz);
typedef TensorEvalToOp<const KernelArgType> EvalTo;
EvalTo evalToTmp(local, m_kernelArg);
const bool Vectorize = internal::IsVectorizable<Device, KernelArgType>::value;
@@ -548,7 +550,7 @@ struct TensorEvaluator<const TensorConvolutionOp<Indices, InputArgType, KernelAr
KernelArgType m_kernelArg;
const Scalar* m_kernel;
bool m_local_kernel;
const Device& m_device;
const Device EIGEN_DEVICE_REF m_device;
};