From 23ac91b6ba235e67847802d4b49e494fa5bedbb6 Mon Sep 17 00:00:00 2001
From: "ASIAPAC\\steli01" <stephen.li@arm.com>
Date: Tue, 7 Nov 2017 16:14:44 +0800
Subject: APPBROWSER-290: DC5x5 optimization for FP16

Change-Id: I0833437d8353515a29aad69c15c99c7b0fd65156
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/111631
Tested-by: BSG Visual Compute Jenkins server to access repositories on http://mpd-gerrit.cambridge.arm.com <bsgcomp@arm.com>
Reviewed-by: Joel Liang <joel.liang@arm.com>
Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
---
 src/core/GLES_COMPUTE/kernels/GCDirectConvolutionLayerKernel.cpp | 1 +
 1 file changed, 1 insertion(+)

(limited to 'src/core/GLES_COMPUTE/kernels/GCDirectConvolutionLayerKernel.cpp')

diff --git a/src/core/GLES_COMPUTE/kernels/GCDirectConvolutionLayerKernel.cpp b/src/core/GLES_COMPUTE/kernels/GCDirectConvolutionLayerKernel.cpp
index 962f044214..b032bc5668 100644
--- a/src/core/GLES_COMPUTE/kernels/GCDirectConvolutionLayerKernel.cpp
+++ b/src/core/GLES_COMPUTE/kernels/GCDirectConvolutionLayerKernel.cpp
@@ -211,6 +211,7 @@ void GCDirectConvolutionLayerKernel<kernel_size>::configure(const IGCTensor *inp
         switch(input->info()->data_type())
         {
             case DataType::F16:
+                options.emplace("#define PROCESS_4X_1Y_1Z");
                 num_elems_read_per_iteration_x    = 8;
                 num_elems_written_per_iteration_x = 4;
 
-- 
cgit v1.2.1