From 4692e11e5af29b97748a1585a092df6800a0a831 Mon Sep 17 00:00:00 2001 From: Kevin May Date: Mon, 18 Oct 2021 14:41:50 +0100 Subject: IVGCVSW-6440 Print new Optimize and LoadedNetwork profiling points * Add parent LoadedNetwork profiling point * Make generic populateParent function to print new descendents in json Signed-off-by: Kevin May Change-Id: I41dc876bffae88e61a16d07fb13b062c321e78a6 --- .../cl/workloads/ClConvolution2dWorkload.cpp | 29 ++++++++++++++-------- 1 file changed, 18 insertions(+), 11 deletions(-) (limited to 'src/backends/cl/workloads/ClConvolution2dWorkload.cpp') diff --git a/src/backends/cl/workloads/ClConvolution2dWorkload.cpp b/src/backends/cl/workloads/ClConvolution2dWorkload.cpp index 12a47dcd94..a5f4c7f254 100644 --- a/src/backends/cl/workloads/ClConvolution2dWorkload.cpp +++ b/src/backends/cl/workloads/ClConvolution2dWorkload.cpp @@ -70,6 +70,7 @@ ClConvolution2dWorkload::ClConvolution2dWorkload(const Convolution2dQueueDescrip : BaseWorkload(descriptor, info) , m_ConvolutionLayer(memoryManager) { + ARMNN_SCOPED_PROFILING_EVENT(Compute::Undefined, "ClConvolution2dWorkload"); const TensorInfo& weightInfo = m_Data.m_Weight->GetTensorInfo(); m_KernelTensor = std::make_unique(); @@ -97,16 +98,19 @@ ClConvolution2dWorkload::ClConvolution2dWorkload(const Convolution2dQueueDescrip const arm_compute::ActivationLayerInfo activationInfo = ConvertAdditionalInfoToAclActivationLayerInfo(descriptor); - m_ConvolutionLayer.configure(clCompileContext, - &input, - m_KernelTensor.get(), - m_BiasTensor.get(), - &output, - padStrideInfo, - arm_compute::WeightsInfo(), - aclDilationInfo, - activationInfo, - isFastMathEnabled); + { + ARMNN_SCOPED_PROFILING_EVENT(Compute::Undefined, "ClConvolution2dWorkload_configure"); + m_ConvolutionLayer.configure(clCompileContext, + &input, + m_KernelTensor.get(), + m_BiasTensor.get(), + &output, + padStrideInfo, + arm_compute::WeightsInfo(), + aclDilationInfo, + activationInfo, + isFastMathEnabled); + } m_ConvolutionMethod = m_ConvolutionLayer.get_convolution_method(input.info(), @@ -146,7 +150,10 @@ ClConvolution2dWorkload::ClConvolution2dWorkload(const Convolution2dQueueDescrip // Force Compute Library to perform the necessary copying and reshaping, after which // delete all the input tensors that will no longer be needed - m_ConvolutionLayer.prepare(); + { + ARMNN_SCOPED_PROFILING_EVENT(Compute::Undefined, "ClConvolution2dWorkload_prepare"); + m_ConvolutionLayer.prepare(); + } FreeUnusedTensors(); } -- cgit v1.2.1