diff options
author | Gunes Bayir <gunes.bayir@arm.com> | 2022-12-18 21:31:29 +0000 |
---|---|---|
committer | Gunes Bayir <gunes.bayir@arm.com> | 2022-12-28 13:36:03 +0000 |
commit | aecb5d93225deace061cb7038d5844eec32c4e52 (patch) | |
tree | 111e0cc4b61b9024d47b14c551424c97a2e9dece /src/dynamic_fusion/sketch/gpu/components/cl/ClComponentLogits1DMaxShiftExpSum.cpp | |
parent | 8ae571454792327fc40641c72fe0b8de1e7d334f (diff) | |
download | ComputeLibrary-aecb5d93225deace061cb7038d5844eec32c4e52.tar.gz |
Implement Logits1DMaxShiftExpSum kernel component in dynamic fusion
Resolves: COMPMID-5719
Change-Id: I2f0911ffccce2b42a9a63fe6826eaa5d2cad06ba
Signed-off-by: Gunes Bayir <gunes.bayir@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/8831
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Reviewed-by: Viet-Hoa Do <viet-hoa.do@arm.com>
Benchmark: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src/dynamic_fusion/sketch/gpu/components/cl/ClComponentLogits1DMaxShiftExpSum.cpp')
-rw-r--r-- | src/dynamic_fusion/sketch/gpu/components/cl/ClComponentLogits1DMaxShiftExpSum.cpp | 97 |
1 files changed, 97 insertions, 0 deletions
diff --git a/src/dynamic_fusion/sketch/gpu/components/cl/ClComponentLogits1DMaxShiftExpSum.cpp b/src/dynamic_fusion/sketch/gpu/components/cl/ClComponentLogits1DMaxShiftExpSum.cpp new file mode 100644 index 0000000000..8ab1853e84 --- /dev/null +++ b/src/dynamic_fusion/sketch/gpu/components/cl/ClComponentLogits1DMaxShiftExpSum.cpp @@ -0,0 +1,97 @@ +/* + * Copyright (c) 2022 Arm Limited. + * + * SPDX-License-Identifier: MIT + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to + * deal in the Software without restriction, including without limitation the + * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or + * sell copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in all + * copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ + +#include "src/dynamic_fusion/sketch/gpu/components/cl/ClComponentLogits1DMaxShiftExpSum.h" + +#include "arm_compute/core/CL/CLHelpers.h" +#include "arm_compute/core/Validate.h" +#include "arm_compute/core/utils/misc/ShapeCalculator.h" +#include "arm_compute/dynamic_fusion/sketch/attributes/DepthwiseConv2dAttributes.h" +#include "src/core/CL/CLValidate.h" +#include "src/dynamic_fusion/sketch/gpu/template_writer/cl/ClTemplateDepthwiseConv2d.h" +#include "src/dynamic_fusion/sketch/gpu/template_writer/cl/ClTemplateLogits1DMaxShiftExpSum.h" + +namespace arm_compute +{ +namespace experimental +{ +namespace dynamic_fusion +{ +Status ClComponentLogits1DMaxShiftExpSum::validate( + const Properties &properties, + const ArgumentPack<ITensorInfo> &tensors, + const Attributes &attributes) +{ + ARM_COMPUTE_UNUSED(properties, attributes); + + const ITensorInfo *src = tensors.get_const_tensor(TensorType::ACL_SRC_0); + const ITensorInfo *sum = tensors.get_const_tensor(TensorType::ACL_DST_0); + const ITensorInfo *dst = tensors.get_const_tensor(TensorType::ACL_DST_1); + + ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(src, sum, dst); + + // 1. Check validity + // All tensor infos are initialized + ARM_COMPUTE_RETURN_ERROR_ON(src->tensor_shape().total_size() == 0); + ARM_COMPUTE_RETURN_ERROR_ON(sum->tensor_shape().total_size() == 0); + ARM_COMPUTE_RETURN_ERROR_ON(dst->tensor_shape().total_size() == 0); + + // Check for mismatches in shapes and data types + ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(src, dst, sum); + ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_SHAPES(src, dst); + + const uint32_t src_dimensions = src->tensor_shape().num_dimensions(); + const uint32_t sum_dimensions = sum->tensor_shape().num_dimensions(); + + ARM_COMPUTE_RETURN_ERROR_ON(src_dimensions != sum_dimensions + 1); + + // Device requirements are met + ARM_COMPUTE_RETURN_ERROR_ON_F16_UNSUPPORTED(src); + + // 2. Check support level + ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(src, 1, DataType::F16, DataType::F32); + + return Status{}; +} + +ClComponentLogits1DMaxShiftExpSum::ClComponentLogits1DMaxShiftExpSum(ComponentId id, + const Properties &properties, + const ArgumentPack<ITensorInfo> &tensors, + const Attributes &attributes) + : IGpuKernelComponent{ id, properties, tensors }, + _component_writer{ std::make_unique<ClTemplateLogits1DMaxShiftExpSum>(id, tensors, attributes) } +{ +} + +ClComponentLogits1DMaxShiftExpSum::~ClComponentLogits1DMaxShiftExpSum() +{ +} + +const IGpuTemplateComponentWriter *ClComponentLogits1DMaxShiftExpSum::template_writer() const +{ + return _component_writer.get(); +} +} // namespace dynamic_fusion +} // namespace experimental +} // namespace arm_compute
\ No newline at end of file |