ONCPUML-1072: Tuned MWS values (for N1, V1) for binary operators used by oneDNN

Added approximate values for MWS for the following binary operators:
Add, Sub, Mul, Min, Max, Div

Change-Id: I5c4c75511129982a3f44c038ee272f09598469de
Signed-off-by: Fadi Arafeh <fadi.arafeh@arm.com>
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/c/VisualCompute/ComputeLibrary/+/459609
Tested-by: bsgcomp <bsgcomp@arm.com>
Reviewed-by: Viet-Hoa Do <viet-hoa.do@arm.com>
Comments-Addressed: bsgcomp <bsgcomp@arm.com>
Signed-off-by: fadara01 <fadi.arafeh@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/8392
Reviewed-by: Gunes Bayir <gunes.bayir@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Benchmark: Arm Jenkins <bsgcomp@arm.com>

1 files changed, 20 insertions, 12 deletions

diff --git a/src/cpu/kernels/CpuElementwiseKernel.h b/src/cpu/kernels/CpuElementwiseKernel.h
index 2785e0a44c..634e38bf9f 100644
--- a/src/cpu/kernels/CpuElementwiseKernel.h
+++ b/src/cpu/kernels/CpuElementwiseKernel.h
@@ -72,8 +72,8 @@ protected:
     static Status validate_arguments_common(const ITensorInfo &src0, const ITensorInfo &src1, const ITensorInfo &dst);
 
 protected:
-    std::function<ElementwiseFunction> _run_method{ nullptr };
-    std::string                        _name{};
+    ElementwiseKernelPtr _run_method{ nullptr };
+    std::string          _name{};
 };
 
 class CpuArithmeticKernel : public CpuElementwiseKernel<CpuArithmeticKernel>
@@ -100,6 +100,15 @@ public:
 
     static const std::vector<CpuElementwiseKernel<CpuArithmeticKernel>::ElementwiseKernel> &get_available_kernels();
 
+    /** Return minimum workload size of the relevant kernel
+     *
+     * @param[in] platform     The CPU platform used to create the context.
+     * @param[in] thread_count Number of threads in the execution.
+     *
+     * @return[out] mws Minimum workload size for requested configuration.
+     */
+    size_t get_mws(const CPUInfo &platform, size_t thread_count) const override;
+
 protected:
     /** Commmon configure function for element-wise operators with no additional options (e.g. Min, Max, SquaredDiff)
      */
@@ -108,16 +117,6 @@ protected:
     static Status validate_arguments(const ITensorInfo &src0, const ITensorInfo &src1, const ITensorInfo &dst);
 
     ArithmeticOperation _op{};
-
-private:
-    /** Function to get the micro kernel implementation
-     *
-     * @param[in] src0 First input tensor information
-     * @param[in] src1 Second input tensor information
-     * @param[in] dst  Output tensor information
-     *
-     * @return the function instance for the micro kernel
-     */
 };
 
 class CpuDivisionKernel : public CpuArithmeticKernel
@@ -141,6 +140,15 @@ public:
      */
     static Status validate(const ITensorInfo *src0, const ITensorInfo *src1, const ITensorInfo *dst);
 
+    /** Return minimum workload size of the relevant kernel
+     *
+     * @param[in] platform     The CPU platform used to create the context.
+     * @param[in] thread_count Number of threads in the execution.
+     *
+     * @return[out] mws Minimum workload size for requested configuration.
+     */
+    size_t get_mws(const CPUInfo &platform, size_t thread_count) const override;
+
 protected:
     // Inherited methods overridden:
     static Status validate_arguments(const ITensorInfo &src0, const ITensorInfo &src1, const ITensorInfo &dst);