Fix OpenCL direct convolution

- The ARM DOT macro was using wrong variables for performing the dot
  product
- K0 could be a non power of 2 values when IFM was not a multiple of 16
- Refactor the test for direct convolution NHWC

Resolves COMPMID-4135, COMPMID-4155

Change-Id: I3a2dc89ef613ae20245cfc28e76ea36c55eaf81d
Signed-off-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4962
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: TeresaARM <teresa.charlinreyes@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>

1 files changed, 1 insertions, 1 deletions

diff --git a/src/core/CL/kernels/CLDirectConvolutionLayerKernel.cpp b/src/core/CL/kernels/CLDirectConvolutionLayerKernel.cpp
index 3ca72b3e5d..3948261bc3 100644
--- a/src/core/CL/kernels/CLDirectConvolutionLayerKernel.cpp
+++ b/src/core/CL/kernels/CLDirectConvolutionLayerKernel.cpp
@@ -375,7 +375,7 @@ void CLDirectConvolutionLayerKernel::configure(const CLCompileContext &compile_c
 
         const unsigned int n0               = win_config.second.x().step();
         const unsigned int m0               = win_config.second.y().step();
-        const unsigned int k0               = std::min(static_cast<unsigned int>(_input->info()->dimension(channel_idx)), 16u);
+        const unsigned int k0               = adjust_vec_size(16u, _input->info()->dimension(channel_idx));
         const unsigned int partial_store_n0 = _output->info()->dimension(channel_idx) % n0;
         const unsigned int partial_store_m0 = (_output->info()->dimension(width_idx) * _output->info()->dimension(height_idx)) % m0;
         const unsigned int pad_left         = conv_info.pad_left();