IVGCVSW-6493 Bug Fix on RHS permute GpuAcc Batch MatMul workload Fp32
Signed-off-by: Teresa Charlin <teresa.charlinreyes@arm.com>
Change-Id: I60e9284b90467f58e0acd74d3f1493546b6f1b9b
diff --git a/src/backends/cl/workloads/ClBatchMatMulWorkload.cpp b/src/backends/cl/workloads/ClBatchMatMulWorkload.cpp
index 4acdef5..ece87c2 100644
--- a/src/backends/cl/workloads/ClBatchMatMulWorkload.cpp
+++ b/src/backends/cl/workloads/ClBatchMatMulWorkload.cpp
@@ -1,5 +1,5 @@
//
-// Copyright © 2022 Arm Ltd and Contributors. All rights reserved.
+// Copyright © 2022-2023 Arm Ltd and Contributors. All rights reserved.
// SPDX-License-Identifier: MIT
//
@@ -158,13 +158,13 @@
if (descriptor.m_Parameters.m_TransposeY == true)
{
armnn::PermutationVector permutationYVector
- = GeneratePermutationVectorOnLastTwoDimensions(info.m_InputTensorInfos[0].GetNumDimensions());
- const TensorInfo permutedYInfo = armnnUtils::Permuted(info.m_InputTensorInfos[0], permutationYVector);
+ = GeneratePermutationVectorOnLastTwoDimensions(info.m_InputTensorInfos[1].GetNumDimensions());
+ const TensorInfo permutedYInfo = armnnUtils::Permuted(info.m_InputTensorInfos[1], permutationYVector);
const auto aclPermutationYVector = armcomputetensorutils::BuildArmComputePermutationVector(permutationYVector);
armcomputetensorutils::BuildArmComputeTensor(m_PermutedTensorY, permutedYInfo);
armcomputetensorutils::InitialiseArmComputeTensorEmpty(m_PermutedTensorY);
- std::unique_ptr<arm_compute::CLPermute> permuteLayerY(new arm_compute::CLPermute());
+ auto permuteLayerY = std::make_unique<arm_compute::CLPermute>();
permuteLayerY->configure(clCompileContext,
&inputY,
&m_PermutedTensorY,