From c577f2c6a3b4ddb6ba87a882723c53a248afbeba Mon Sep 17 00:00:00 2001 From: telsoa01 Date: Fri, 31 Aug 2018 09:22:23 +0100 Subject: Release 18.08 --- .../NeonMultiplicationFloat32Workload.cpp | 23 ++++++++++++++++++++-- 1 file changed, 21 insertions(+), 2 deletions(-) (limited to 'src/armnn/backends/NeonWorkloads/NeonMultiplicationFloat32Workload.cpp') diff --git a/src/armnn/backends/NeonWorkloads/NeonMultiplicationFloat32Workload.cpp b/src/armnn/backends/NeonWorkloads/NeonMultiplicationFloat32Workload.cpp index 58ce7b74ba..a8a3cd77b4 100644 --- a/src/armnn/backends/NeonWorkloads/NeonMultiplicationFloat32Workload.cpp +++ b/src/armnn/backends/NeonWorkloads/NeonMultiplicationFloat32Workload.cpp @@ -9,9 +9,28 @@ namespace armnn { +arm_compute::Status NeonMultiplicationWorkloadValidate(const TensorInfo& input0, + const TensorInfo& input1, + const TensorInfo& output) +{ + const arm_compute::TensorInfo aclInput1 = armcomputetensorutils::BuildArmComputeTensorInfo(input0); + const arm_compute::TensorInfo aclInput2 = armcomputetensorutils::BuildArmComputeTensorInfo(input1); + const arm_compute::TensorInfo aclOutput = armcomputetensorutils::BuildArmComputeTensorInfo(output); + + // At the time of writing, configure() will fail if a rounding policy other than TO_ZERO is supplied to it, + // when providing a scale of 1.0 for F32 tensors, even though the provided rounding policy appears to be + // ignored for F32 tensors. + return arm_compute::NEPixelWiseMultiplication::validate(&aclInput1, + &aclInput2, + &aclOutput, + 1.0f, + arm_compute::ConvertPolicy::SATURATE, + arm_compute::RoundingPolicy::TO_ZERO); +} + NeonMultiplicationFloat32Workload::NeonMultiplicationFloat32Workload(const MultiplicationQueueDescriptor& descriptor, const WorkloadInfo& info) - : Float32Workload(descriptor, info) + : FloatWorkload(descriptor, info) { m_Data.ValidateInputsOutputs("NeonMultiplicationFloat32Workload", 2, 1); @@ -32,7 +51,7 @@ NeonMultiplicationFloat32Workload::NeonMultiplicationFloat32Workload(const Multi void NeonMultiplicationFloat32Workload::Execute() const { - ARMNN_SCOPED_PROFILING_EVENT(Compute::CpuAcc, "NeonMultiplicationFloat32Workload_Execute"); + ARMNN_SCOPED_PROFILING_EVENT_NEON("NeonMultiplicationFloat32Workload_Execute"); m_PixelWiseMultiplication.run(); } -- cgit v1.2.1