diff options
| author | Manuel Bottini <manuel.bottini@arm.com> | 2020-04-08 10:15:51 +0100 |
|---|---|---|
| committer | Manuel Bottini <manuel.bottini@arm.com> | 2020-04-17 14:10:38 +0000 |
| commit | 4c6bd514a8d424a29b776754f1b3426fa3a8c339 (patch) | |
| tree | ed20ea238ae069ab138b62ea6879e0c1b955b606 /src/core/CL/kernels/CLAccumulateKernel.cpp | |
| parent | d2f6d96cdc2ad2169c4abd0a8c4884f61ed6d186 (diff) | |
| download | ComputeLibrary-4c6bd514a8d424a29b776754f1b3426fa3a8c339.tar.gz | |
COMPMID-3280: Make all ML primitives for CL use the new interface - Part 1
- Only CLKernels have been updated
Change-Id: Ife55b847c2e39e712a186eb6ca452503d5b66937
Signed-off-by: Manuel Bottini <manuel.bottini@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/3001
Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Reviewed-by: Michalis Spyrou <michalis.spyrou@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src/core/CL/kernels/CLAccumulateKernel.cpp')
| -rw-r--r-- | src/core/CL/kernels/CLAccumulateKernel.cpp | 23 |
1 files changed, 19 insertions, 4 deletions
diff --git a/src/core/CL/kernels/CLAccumulateKernel.cpp b/src/core/CL/kernels/CLAccumulateKernel.cpp index 12ee210243..aa13b4a207 100644 --- a/src/core/CL/kernels/CLAccumulateKernel.cpp +++ b/src/core/CL/kernels/CLAccumulateKernel.cpp @@ -1,5 +1,5 @@ /* - * Copyright (c) 2016-2019 ARM Limited. + * Copyright (c) 2016-2020 ARM Limited. * * SPDX-License-Identifier: MIT * @@ -40,11 +40,16 @@ constexpr unsigned int num_elems_processed_per_iteration = 16; void CLAccumulateKernel::configure(const ICLTensor *input, ICLTensor *accum) { + configure(CLKernelLibrary::get().get_compile_context(), input, accum); +} + +void CLAccumulateKernel::configure(CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *accum) +{ ARM_COMPUTE_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(input, 1, DataType::U8); ARM_COMPUTE_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(accum, 1, DataType::S16); // Create kernel - _kernel = static_cast<cl::Kernel>(CLKernelLibrary::get().create_kernel("accumulate")); + _kernel = create_kernel(compile_context, "accumulate"); // Make sure _kernel is initialized before calling the parent's configure ICLSimple2DKernel::configure(input, accum, num_elems_processed_per_iteration); @@ -52,12 +57,17 @@ void CLAccumulateKernel::configure(const ICLTensor *input, ICLTensor *accum) void CLAccumulateWeightedKernel::configure(const ICLTensor *input, float alpha, ICLTensor *accum) { + configure(CLKernelLibrary::get().get_compile_context(), input, alpha, accum); +} + +void CLAccumulateWeightedKernel::configure(CLCompileContext &compile_context, const ICLTensor *input, float alpha, ICLTensor *accum) +{ ARM_COMPUTE_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(input, 1, DataType::U8); ARM_COMPUTE_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(accum, 1, DataType::U8); ARM_COMPUTE_ERROR_ON(alpha < 0.0 || alpha > 1.0); // Create kernel - _kernel = static_cast<cl::Kernel>(CLKernelLibrary::get().create_kernel("accumulate_weighted")); + _kernel = create_kernel(compile_context, "accumulate_weighted"); // Set static kernel arguments unsigned int idx = 2 * num_arguments_per_2D_tensor(); //Skip the input and output parameters @@ -69,12 +79,17 @@ void CLAccumulateWeightedKernel::configure(const ICLTensor *input, float alpha, void CLAccumulateSquaredKernel::configure(const ICLTensor *input, uint32_t shift, ICLTensor *accum) { + configure(CLKernelLibrary::get().get_compile_context(), input, shift, accum); +} + +void CLAccumulateSquaredKernel::configure(CLCompileContext &compile_context, const ICLTensor *input, uint32_t shift, ICLTensor *accum) +{ ARM_COMPUTE_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(input, 1, DataType::U8); ARM_COMPUTE_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(accum, 1, DataType::S16); ARM_COMPUTE_ERROR_ON(shift > 15); // Create kernel - _kernel = static_cast<cl::Kernel>(CLKernelLibrary::get().create_kernel("accumulate_squared")); + _kernel = create_kernel(compile_context, "accumulate_squared"); // Set static kernel arguments unsigned int idx = 2 * num_arguments_per_2D_tensor(); //Skip the input and output parameters |