diff options
Diffstat (limited to 'src/backends/cl/workloads/ClReduceWorkload.cpp')
-rw-r--r-- | src/backends/cl/workloads/ClReduceWorkload.cpp | 22 |
1 files changed, 14 insertions, 8 deletions
diff --git a/src/backends/cl/workloads/ClReduceWorkload.cpp b/src/backends/cl/workloads/ClReduceWorkload.cpp index c9c2e0a400..1a7bc64420 100644 --- a/src/backends/cl/workloads/ClReduceWorkload.cpp +++ b/src/backends/cl/workloads/ClReduceWorkload.cpp @@ -17,28 +17,28 @@ using namespace armcomputetensorutils; arm_compute::Status ClReduceWorkloadValidate(const TensorInfo& input, const TensorInfo& output, - const ReduceDescriptor& desc) + const ReduceDescriptor& descriptor) { - if ( desc.m_vAxis.size()==1 || desc.m_vAxis.empty()) + if (descriptor.m_vAxis.size() == 1 || descriptor.m_vAxis.empty()) { - const arm_compute::TensorInfo aclInputInfo = armcomputetensorutils::BuildArmComputeTensorInfo(input); + const arm_compute::TensorInfo aclInputInfo = armcomputetensorutils::BuildArmComputeTensorInfo(input); const arm_compute::TensorInfo aclOutputInfo = armcomputetensorutils::BuildArmComputeTensorInfo(output); arm_compute::Coordinates coords = BuildArmComputeReductionCoordinates(aclInputInfo.num_dimensions(), input.GetNumDimensions(), - desc.m_vAxis); + descriptor.m_vAxis); return arm_compute::CLReductionOperation::validate(&aclInputInfo, &aclOutputInfo, static_cast<unsigned int>(coords[0]), - ConvertReductionOperationToAcl(desc), - desc.m_KeepDims); + ConvertReductionOperationToAcl(descriptor), + descriptor.m_KeepDims); } else { // Validate layer if there are multiple axes. arm_compute::Status status; - IS_MULTI_AXES_REDUCE_SUPPORTED(ClReduceWorkloadValidate, input, desc, status); + IS_MULTI_AXES_REDUCE_SUPPORTED(ClReduceWorkloadValidate, input, descriptor, status); return status; } } @@ -46,6 +46,12 @@ arm_compute::Status ClReduceWorkloadValidate(const TensorInfo& input, ClReduceWorkload::ClReduceWorkload(const ReduceQueueDescriptor& descriptor, const WorkloadInfo& info) : BaseWorkload<ReduceQueueDescriptor>(descriptor, info) { + // Report Profiling Details + ARMNN_REPORT_PROFILING_WORKLOAD_DESC("ClReduceWorkload_Construct", + descriptor.m_Parameters, + info, + this->GetGuid()); + m_Data.ValidateInputsOutputs("ClReduceWorkload", 1, 1); arm_compute::ICLTensor& input = static_cast<IClTensorHandle*>(m_Data.m_Inputs[0])->GetTensor(); @@ -63,7 +69,7 @@ ClReduceWorkload::ClReduceWorkload(const ReduceQueueDescriptor& descriptor, cons void ClReduceWorkload::Execute() const { - ARMNN_SCOPED_PROFILING_EVENT_CL("ClReduceWorkload_Execute"); + ARMNN_SCOPED_PROFILING_EVENT_CL_GUID("ClReduceWorkload_Execute", this->GetGuid()); m_Layer.run(); } |