diff options
| author | Gian Marco Iodice <gianmarco.iodice@arm.com> | 2022-12-30 16:07:45 +0000 |
|---|---|---|
| committer | Gian Marco Iodice <gianmarco.iodice@arm.com> | 2023-01-10 09:57:51 +0000 |
| commit | 3cce35dcad8bc8f53a1e6613f719af9ab04feda6 (patch) | |
| tree | e1015566852ebce4af897db37cf5cb1989c29924 /src/dynamic_fusion/runtime/gpu/cl/ClKernelRuntime.cpp | |
| parent | d2d9361a0a338bce478f7d85b4af70d1ed20f26c (diff) | |
| download | ComputeLibrary-3cce35dcad8bc8f53a1e6613f719af9ab04feda6.tar.gz | |
Extend cl image support to input and output tensors
- Add support for texture image to input and output of direct
convolution
- Extend T_LOAD2D_INDIRECT macro to read values from cl image storages
Resolves COMPMID-5715
Signed-off-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Change-Id: Idb0410f53f6d0763cd9e39895a7cbf9bc826d33a
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/8904
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Viet-Hoa Do <viet-hoa.do@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Benchmark: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src/dynamic_fusion/runtime/gpu/cl/ClKernelRuntime.cpp')
| -rw-r--r-- | src/dynamic_fusion/runtime/gpu/cl/ClKernelRuntime.cpp | 8 |
1 files changed, 4 insertions, 4 deletions
diff --git a/src/dynamic_fusion/runtime/gpu/cl/ClKernelRuntime.cpp b/src/dynamic_fusion/runtime/gpu/cl/ClKernelRuntime.cpp index 93fbdfed63..022d4685fe 100644 --- a/src/dynamic_fusion/runtime/gpu/cl/ClKernelRuntime.cpp +++ b/src/dynamic_fusion/runtime/gpu/cl/ClKernelRuntime.cpp @@ -1,5 +1,5 @@ /* - * Copyright (c) 2022 Arm Limited. + * Copyright (c) 2022-2023 Arm Limited. * * SPDX-License-Identifier: MIT * @@ -89,7 +89,7 @@ inline void ClKernelRuntime::add_tensor_argument(unsigned int &idx, const GpuKer { const TensorShape shape2d(tensor->info()->dimension(0) / 4, tensor->info()->dimension(1) * tensor->info()->dimension(2) * tensor->info()->dimension(3)); const size_t image_row_pitch = tensor->info()->strides_in_bytes()[1]; - cl::Image2D tensor_image2d = create_image2d_from_buffer(CLKernelLibrary::get().context(), tensor->cl_buffer(), shape2d, tensor->info()->data_type(), image_row_pitch); + cl::Image2D tensor_image2d = create_image2d_from_buffer(CLKernelLibrary::get().context(), tensor->cl_buffer(), shape2d, tensor->info()->data_type(), image_row_pitch, CLImage2DType::ReadOnly); cl_images.push_back(tensor_image2d); _kernel.setArg(idx++, tensor_image2d); break; @@ -105,7 +105,7 @@ inline void ClKernelRuntime::add_tensor_argument(unsigned int &idx, const GpuKer { const TensorShape shape2d(tensor->info()->dimension(0) / 4, tensor->info()->dimension(1) * tensor->info()->dimension(2) * tensor->info()->dimension(3)); const size_t image_row_pitch = tensor->info()->strides_in_bytes()[1]; - cl::Image2D tensor_image2d = create_image2d_from_buffer(CLKernelLibrary::get().context(), tensor->cl_buffer(), shape2d, tensor->info()->data_type(), image_row_pitch); + cl::Image2D tensor_image2d = create_image2d_from_buffer(CLKernelLibrary::get().context(), tensor->cl_buffer(), shape2d, tensor->info()->data_type(), image_row_pitch, CLImage2DType::ReadOnly); cl_images.push_back(tensor_image2d); _kernel.setArg(idx++, tensor_image2d); _kernel.setArg<cl_uint>(idx++, static_cast<unsigned int>(tensor->info()->strides_in_bytes()[2])); @@ -135,7 +135,7 @@ inline void ClKernelRuntime::add_tensor_argument(unsigned int &idx, const GpuKer const size_t image_stride_y = tensor->info()->strides_in_bytes()[1]; cl::Image2D tensor_image2d = create_image2d_from_buffer(CLKernelLibrary::get().context(), tensor->cl_buffer(), - TensorShape(image_w, image_h), tensor->info()->data_type(), image_stride_y); + TensorShape(image_w, image_h), tensor->info()->data_type(), image_stride_y, CLImage2DType::ReadOnly); cl_images.push_back(tensor_image2d); _kernel.setArg(idx++, tensor_image2d); |