diff options
author | Georgios Pinitas <georgios.pinitas@arm.com> | 2021-06-25 05:42:57 +0100 |
---|---|---|
committer | Georgios Pinitas <georgios.pinitas@arm.com> | 2021-06-29 12:38:33 +0000 |
commit | 5fdde99f4271891a40c02cd1e89f1344aa84583a (patch) | |
tree | 35944b8bb0eee6aa9bbca08c38325f10cf66370c /src/core/cpu/kernels/CpuElementwiseKernel.cpp | |
parent | 4a95bba6ca61ce99995ece6fd237b5498c9f322c (diff) | |
download | ComputeLibrary-5fdde99f4271891a40c02cd1e89f1344aa84583a.tar.gz |
Improve selection speed of CPU implementations
CPU micro-kernel to be used was picked during kernel execution.
Move selection during configuration to reduce runtime overhead.
Standardize kernel names as follows:
<simd_tech>_<data_type>_<data_layout>_<kernel_name>
e.g. sve_fp32_nhwc_scale
Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com>
Change-Id: I544f1c08c8fef0f130a3bde61882ccb9a1f47f21
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5855
Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src/core/cpu/kernels/CpuElementwiseKernel.cpp')
-rw-r--r-- | src/core/cpu/kernels/CpuElementwiseKernel.cpp | 103 |
1 files changed, 54 insertions, 49 deletions
diff --git a/src/core/cpu/kernels/CpuElementwiseKernel.cpp b/src/core/cpu/kernels/CpuElementwiseKernel.cpp index dc0c5b210d..dc574fce65 100644 --- a/src/core/cpu/kernels/CpuElementwiseKernel.cpp +++ b/src/core/cpu/kernels/CpuElementwiseKernel.cpp @@ -58,69 +58,68 @@ struct ElementwiseKernel UKernelType *ukernel; }; -template <ArithmeticOperation op> -std::function<void(const ITensor *, const ITensor *, ITensor *, const Window &)> -configure_arithm_func(const ITensorInfo *src0, const ITensorInfo *src1, ITensorInfo *dst) +template <ArithmeticOperation op> +CpuElementwiseKernel::UKernelInfo configure_arithm_func(const ITensorInfo *src0, const ITensorInfo *src1, ITensorInfo *dst) { ARM_COMPUTE_UNUSED(src1, dst); static ElementwiseKernel kernels[] = { #if defined(ARM_COMPUTE_ENABLE_SVE) { - "sve_elementwise_fp32", + "sve_fp32_elementwise", [](const ElementwiseSelectorData & data) { return data.dt == DataType::F32 && data.ci.has_sve(); }, REGISTER_FP32_SVE((arm_compute::cpu::elementwise_arithmetic_op<op, float32_t>)) }, { - "sve_elementwise_s32", + "sve_s32_elementwise", [](const ElementwiseSelectorData & data) { return data.dt == DataType::S32 && data.ci.has_sve(); }, REGISTER_INTEGER_SVE((arm_compute::cpu::elementwise_arithmetic_op<op, int32_t>)) }, { - "sve_elementwise_s16", + "sve_s16_elementwise", [](const ElementwiseSelectorData & data) { return data.dt == DataType::S16 && data.ci.has_sve(); }, REGISTER_INTEGER_SVE((arm_compute::cpu::elementwise_arithmetic_op<op, int16_t>)) }, #endif /* defined(ARM_COMPUTE_ENABLE_SVE) */ #if defined(ARM_COMPUTE_ENABLE_NEON) { - "neon_elementwise_f32", + "neon_fp32_elementwise", [](const ElementwiseSelectorData & data) { return data.dt == DataType::F32; }, REGISTER_FP32_NEON((arm_compute::cpu::elementwise_arithm_op<op, typename wrapper::traits::neon_vector<float, 4>>)) }, { - "neon_elementwise_s32", + "neon_s32_elementwise", [](const ElementwiseSelectorData & data) { return data.dt == DataType::S32; }, REGISTER_INTEGER_NEON((arm_compute::cpu::elementwise_arithm_op<op, typename wrapper::traits::neon_vector<int32_t, 4>>)) }, #endif /* defined(ARM_COMPUTE_ENABLE_NEON) */ #if defined(ARM_COMPUTE_ENABLE_SVE2) { - "sve2_elementwise_qu8", + "sve2_qu8_elementwise", [](const ElementwiseSelectorData & data) { return data.dt == DataType::QASYMM8 && data.ci.has_sve2(); }, REGISTER_QASYMM8_SVE((arm_compute::cpu::elementwise_arithmetic_quantized_op<op, uint8_t>)) }, { - "sve2_elementwise_qs8", + "sve2_qs8_elementwise", [](const ElementwiseSelectorData & data) { return data.dt == DataType::QASYMM8_SIGNED && data.ci.has_sve2(); }, REGISTER_QASYMM8_SIGNED_SVE((arm_compute::cpu::elementwise_arithmetic_quantized_op<op, int8_t>)) }, #endif /* defined(ARM_COMPUTE_ENABLE_SVE2) */ #if defined(ARM_COMPUTE_ENABLE_NEON) || defined(ARM_COMPUTE_ENABLE_SVE) { - "neon_elementwise_qu8", + "neon_qu8_elementwise", [](const ElementwiseSelectorData & data) { return data.dt == DataType::QASYMM8; }, REGISTER_QASYMM8_NEON((arm_compute::cpu::elementwise_arithm_op_quantized<op>)) }, { - "neon_elementwise_qs8", + "neon_qs8_elementwise", [](const ElementwiseSelectorData & data) { return data.dt == DataType::QASYMM8_SIGNED; }, REGISTER_QASYMM8_SIGNED_NEON((arm_compute::cpu::elementwise_arithm_op_quantized_signed<op>)) }, #endif /* defined(ARM_COMPUTE_ENABLE_NEON) || defined(ARM_COMPUTE_ENABLE_SVE) */ #if defined(ARM_COMPUTE_ENABLE_SVE) { - "sve_elementwise_f16", + "sve_fp16_elementwise", [](const ElementwiseSelectorData & data) { return data.dt == DataType::F16 && data.ci.has_sve(); }, REGISTER_FP16_SVE((arm_compute::cpu::elementwise_arithmetic_op<op, float16_t>)) }, @@ -128,13 +127,13 @@ configure_arithm_func(const ITensorInfo *src0, const ITensorInfo *src1, ITensorI #if defined(ARM_COMPUTE_ENABLE_NEON) #if defined(__ARM_FEATURE_FP16_VECTOR_ARITHMETIC) { - "neon_elementwise_f16", + "neon_fp16_elementwise", [](const ElementwiseSelectorData & data) { return data.dt == DataType::F16 && data.ci.has_fp16(); }, REGISTER_FP16_NEON((arm_compute::cpu::elementwise_arithm_op<op, typename wrapper::traits::neon_vector<float16_t, 8>>)) }, #endif /* defined(__ARM_FEATURE_FP16_VECTOR_ARITHMETIC) */ { - "neon_elementwise_s16", + "neon_s16_elementwise", [](const ElementwiseSelectorData & data) { return data.dt == DataType::S16; }, REGISTER_INTEGER_NEON((arm_compute::cpu::elementwise_arithm_op<op, typename wrapper::traits::neon_vector<int16_t, 8>>)) }, @@ -145,98 +144,97 @@ configure_arithm_func(const ITensorInfo *src0, const ITensorInfo *src1, ITensorI { if(uk.is_selected({ src0->data_type(), CPUInfo::get() })) { - return uk.ukernel; + return { uk.name, uk.ukernel }; } } - return nullptr; + return { "", nullptr }; } -template <ComparisonOperation op> -std::function<void(const ITensor *, const ITensor *, ITensor *, const Window &)> -configure_comp_func(const ITensorInfo *src0, const ITensorInfo *src1, ITensorInfo *dst) +template <ComparisonOperation op> +CpuElementwiseKernel::UKernelInfo configure_comp_func(const ITensorInfo *src0, const ITensorInfo *src1, ITensorInfo *dst) { ARM_COMPUTE_UNUSED(src1, dst); static ElementwiseKernel kernels[] = { #if defined(ARM_COMPUTE_ENABLE_SVE) { - "sve_comparison_u8", + "sve_u8_comparison", [](const ElementwiseSelectorData & data) { return data.dt == DataType::U8 && data.ci.has_sve(); }, REGISTER_INTEGER_SVE((arm_compute::cpu::elementwise_comparison_op<op, uint8_t>)) }, { - "sve_comparison_f32", + "sve_fp32_comparison", [](const ElementwiseSelectorData & data) { return data.dt == DataType::F32 && data.ci.has_sve(); }, REGISTER_FP32_SVE((arm_compute::cpu::elementwise_comparison_op<op, float>)) }, { - "sve_comparison_s16", + "sve_s16_comparison", [](const ElementwiseSelectorData & data) { return data.dt == DataType::S16 && data.ci.has_sve(); }, REGISTER_INTEGER_SVE((arm_compute::cpu::elementwise_comparison_op<op, int16_t>)) }, { - "sve_comparison_s32", + "sve_s32_comparison", [](const ElementwiseSelectorData & data) { return data.dt == DataType::S32 && data.ci.has_sve(); }, REGISTER_INTEGER_SVE((arm_compute::cpu::elementwise_comparison_op<op, int32_t>)) }, #endif /* defined(ARM_COMPUTE_ENABLE_SVE) */ #if defined(ARM_COMPUTE_ENABLE_NEON) { - "neon_comparison_u8", + "neon_u8_comparison", [](const ElementwiseSelectorData & data) { return data.dt == DataType::U8; }, REGISTER_INTEGER_NEON((arm_compute::cpu::elementwise_comp_op_8<op, uint8_t, uint8x16_t>)) }, { - "neon_comparison_f32", + "neon_fp32_comparison", [](const ElementwiseSelectorData & data) { return data.dt == DataType::F32; }, REGISTER_FP32_NEON((arm_compute::cpu::elementwise_comp_op_32<op, float, float32x4_t>)) }, { - "neon_comparison_s16", + "neon_s16_comparison", [](const ElementwiseSelectorData & data) { return data.dt == DataType::S16; }, REGISTER_INTEGER_NEON((arm_compute::cpu::elementwise_comp_op_16<op, int16_t, int16x8_t>)) }, { - "neon_comparison_s32", + "neon_s32_comparison", [](const ElementwiseSelectorData & data) { return data.dt == DataType::S32; }, REGISTER_INTEGER_NEON((arm_compute::cpu::elementwise_comp_op_32<op, int32_t, int32x4_t>)) }, #endif /* defined(ARM_COMPUTE_ENABLE_NEON) */ #if defined(ARM_COMPUTE_ENABLE_SVE2) { - "sve_comparison_qu8", + "sve2_qu8_comparison", [](const ElementwiseSelectorData & data) { return data.dt == DataType::QASYMM8 && data.ci.has_sve2(); }, REGISTER_QASYMM8_SVE((arm_compute::cpu::elementwise_comparison_quantized_op<op, uint8_t>)) }, { - "sve_comparison_qs8", + "sve2_qs8_comparison", [](const ElementwiseSelectorData & data) { return data.dt == DataType::QASYMM8_SIGNED && data.ci.has_sve2(); }, REGISTER_QASYMM8_SIGNED_SVE((arm_compute::cpu::elementwise_comparison_quantized_op<op, int8_t>)) }, #endif /* defined(ARM_COMPUTE_ENABLE_SVE2) */ #if defined(ARM_COMPUTE_ENABLE_NEON) || defined(ARM_COMPUTE_ENABLE_SVE) { - "neon_comparison_qu8", + "neon_qu8_comparison", [](const ElementwiseSelectorData & data) { return data.dt == DataType::QASYMM8; }, REGISTER_QASYMM8_NEON((arm_compute::cpu::elementwise_comp_op_quantized<op>)) }, { - "neon_comparison_qs8", + "neon_qs8_comparison", [](const ElementwiseSelectorData & data) { return data.dt == DataType::QASYMM8_SIGNED; }, REGISTER_QASYMM8_SIGNED_NEON((arm_compute::cpu::elementwise_comp_op_quantized_signed<op>)) }, #endif /* defined(ARM_COMPUTE_ENABLE_NEON) || defined(ARM_COMPUTE_ENABLE_SVE) */ #if defined(ARM_COMPUTE_ENABLE_SVE) { - "sve_comparison_f16", + "sve_fp16_comparison", [](const ElementwiseSelectorData & data) { return data.dt == DataType::F16 && data.ci.has_sve(); }, REGISTER_FP16_SVE((arm_compute::cpu::elementwise_comparison_op<op, float16_t>)) }, #endif /* defined(ARM_COMPUTE_ENABLE_SVE) */ #if defined(ARM_COMPUTE_ENABLE_NEON) && defined(__ARM_FEATURE_FP16_VECTOR_ARITHMETIC) { - "neon_comparison_f16", + "neon_fp16_comparison", [](const ElementwiseSelectorData & data) { return data.dt == DataType::F16 && data.ci.has_fp16(); }, REGISTER_FP16_NEON((arm_compute::cpu::elementwise_comp_op_16<op, float16_t, float16x8_t>)) }, @@ -247,11 +245,11 @@ configure_comp_func(const ITensorInfo *src0, const ITensorInfo *src1, ITensorInf { if(uk.is_selected({ src0->data_type(), CPUInfo::get() })) { - return uk.ukernel; + return { uk.name, uk.ukernel }; } } - return nullptr; + return { "", nullptr }; } } // namespace @@ -278,6 +276,11 @@ void CpuElementwiseKernel::configure_common(const ITensorInfo *src0, const ITens { ARM_COMPUTE_ERROR_ON_NULLPTR(src0, src1, dst); + const auto uk = get_implementation(src0, src1, dst); + + _run_method = uk.ukernel; + _name = std::string("CpuElementwiseKernel").append("/").append(uk.name); + // If any of shapes is dynamic, expect a configured window and dst at run-time. if(src0->is_dynamic() || src1->is_dynamic()) { @@ -292,22 +295,26 @@ void CpuElementwiseKernel::configure_common(const ITensorInfo *src0, const ITens void CpuElementwiseKernel::run_op(ITensorPack &tensors, const Window &window, const ThreadInfo &info) { ARM_COMPUTE_UNUSED(info); + ARM_COMPUTE_ERROR_ON(_run_method == nullptr); auto src0 = tensors.get_const_tensor(TensorType::ACL_SRC_0); auto src1 = tensors.get_const_tensor(TensorType::ACL_SRC_1); auto dst = tensors.get_tensor(TensorType::ACL_DST); - auto function = get_implementation(src0->info(), src1->info(), dst->info()); - ARM_COMPUTE_ERROR_ON(function == nullptr); - function(src0, src1, dst, window); + _run_method(src0, src1, dst, window); +} + +const char *CpuElementwiseKernel::name() const +{ + return _name.c_str(); } /** Arithmetic operators (min, max, squared_diff) */ void CpuArithmeticKernel::configure(ArithmeticOperation op, const ITensorInfo *src0, const ITensorInfo *src1, ITensorInfo *dst) { ARM_COMPUTE_ERROR_THROW_ON(validate_arguments(*src0, *src1, *dst)); - configure_common(src0, src1, dst); _op = op; + configure_common(src0, src1, dst); } Status CpuArithmeticKernel::validate_arguments(const ITensorInfo &src0, const ITensorInfo &src1, const ITensorInfo &dst) @@ -329,8 +336,7 @@ Status CpuArithmeticKernel::validate(ArithmeticOperation op, const ITensorInfo * return Status{}; } -std::function<CpuElementwiseKernel::ElementwiseFunction> -CpuArithmeticKernel::get_implementation(const ITensorInfo *src0, const ITensorInfo *src1, ITensorInfo *dst) +CpuElementwiseKernel::UKernelInfo CpuArithmeticKernel::get_implementation(const ITensorInfo *src0, const ITensorInfo *src1, ITensorInfo *dst) { switch(_op) { @@ -349,7 +355,7 @@ CpuArithmeticKernel::get_implementation(const ITensorInfo *src0, const ITensorIn default: ARM_COMPUTE_ERROR("NOT_SUPPORTED!"); } - return nullptr; + return { "", nullptr }; } /** The division operator */ @@ -357,8 +363,8 @@ CpuArithmeticKernel::get_implementation(const ITensorInfo *src0, const ITensorIn void CpuDivisionKernel::configure(const ITensorInfo *src0, const ITensorInfo *src1, ITensorInfo *dst) { ARM_COMPUTE_ERROR_THROW_ON(validate_arguments(*src0, *src1, *dst)); - configure_common(src0, src1, dst); _op = ArithmeticOperation::DIV; + configure_common(src0, src1, dst); } Status CpuDivisionKernel::validate_arguments(const ITensorInfo &src0, const ITensorInfo &src1, const ITensorInfo &dst) @@ -378,8 +384,8 @@ Status CpuDivisionKernel::validate(const ITensorInfo *src0, const ITensorInfo *s void CpuPowerKernel::configure(const ITensorInfo *src0, const ITensorInfo *src1, ITensorInfo *dst) { ARM_COMPUTE_ERROR_THROW_ON(validate_arguments(*src0, *src1, *dst)); - configure_common(src0, src1, dst); _op = ArithmeticOperation::POWER; + configure_common(src0, src1, dst); } Status CpuPowerKernel::validate_arguments(const ITensorInfo &src0, const ITensorInfo &src1, const ITensorInfo &dst) @@ -399,8 +405,8 @@ Status CpuPowerKernel::validate(const ITensorInfo *src0, const ITensorInfo *src1 void CpuComparisonKernel::configure(ComparisonOperation op, const ITensorInfo *src0, const ITensorInfo *src1, ITensorInfo *dst) { ARM_COMPUTE_ERROR_THROW_ON(validate_arguments(*src0, *src1, *dst)); - configure_common(src0, src1, dst); _op = op; + configure_common(src0, src1, dst); } Status CpuComparisonKernel::validate_arguments(const ITensorInfo &src0, const ITensorInfo &src1, const ITensorInfo &dst) @@ -422,8 +428,7 @@ Status CpuComparisonKernel::validate(ComparisonOperation op, const ITensorInfo * return Status{}; } -std::function<CpuElementwiseKernel::ElementwiseFunction> -CpuComparisonKernel::get_implementation(const ITensorInfo *src0, const ITensorInfo *src1, ITensorInfo *dst) +CpuElementwiseKernel::UKernelInfo CpuComparisonKernel::get_implementation(const ITensorInfo *src0, const ITensorInfo *src1, ITensorInfo *dst) { switch(_op) { @@ -442,7 +447,7 @@ CpuComparisonKernel::get_implementation(const ITensorInfo *src0, const ITensorIn default: ARM_COMPUTE_ERROR("NOT_SUPPORTED!"); } - return nullptr; + return { "", nullptr }; } } // namespace kernels } // namespace cpu |