COMPMID-477 - Optimized batched case in CLConvolutionLayer
Change-Id: I4ef18f49f1da0cb816aaa0762466b940792c15ed
Reviewed-on: http://mpd-gerrit.cambridge.arm.com/84162
Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
diff --git a/src/runtime/CL/functions/CLConvolutionLayer.cpp b/src/runtime/CL/functions/CLConvolutionLayer.cpp
index b1b8398..0bbec94 100644
--- a/src/runtime/CL/functions/CLConvolutionLayer.cpp
+++ b/src/runtime/CL/functions/CLConvolutionLayer.cpp
@@ -197,9 +197,12 @@
// Configure kernels
_input_im2col_kernel.configure(input, &_input_im2col_reshaped, Size2D(kernel_width, kernel_height), conv_info, _has_bias);
+
+ // Configure matrix multiply
if(_is_fully_connected_convolution)
{
- _mm_kernel.configure(&_input_im2col_reshaped, weights, &_gemm_output, 1.0f);
+ // The matrix A and Matrix B have not been reshaped
+ _mm_kernel.configure(&_input_im2col_reshaped, weights, &_gemm_output, 1.0f, false);
}
else
{