From 20c246a60869bada4051bd14eb9a3862be5330d7 Mon Sep 17 00:00:00 2001
From: Georgios Pinitas <georgios.pinitas@arm.com>
Date: Wed, 12 Sep 2018 16:45:53 +0100
Subject: COMPMID-1532: Add DepthwiseConvolution3x3 FP16 on NEON

Change-Id: I780970f317b979b3230e2b471ac01df7fda9ee14
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/148168
Tested-by: bsgcomp <bsgcomp@arm.com>
Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
---
 .../NEON/kernels/NEDirectConvolutionLayerOutputStageKernel.cpp     | 7 +++++++
 1 file changed, 7 insertions(+)

(limited to 'src/core/NEON/kernels/NEDirectConvolutionLayerOutputStageKernel.cpp')

diff --git a/src/core/NEON/kernels/NEDirectConvolutionLayerOutputStageKernel.cpp b/src/core/NEON/kernels/NEDirectConvolutionLayerOutputStageKernel.cpp
index eefbd98dd8..864c63f731 100644
--- a/src/core/NEON/kernels/NEDirectConvolutionLayerOutputStageKernel.cpp
+++ b/src/core/NEON/kernels/NEDirectConvolutionLayerOutputStageKernel.cpp
@@ -451,6 +451,13 @@ void NEDirectConvolutionLayerOutputStageKernel::configure(ITensor *input, const
     {
         switch(input->info()->data_type())
         {
+#ifdef __ARM_FEATURE_FP16_VECTOR_ARITHMETIC
+            case DataType::F16:
+            {
+                _func = (output == nullptr) ? &output_stage_nhwc<float16_t, float16_t, true, true> : &output_stage_nhwc<float16_t, float16_t, false, true>;
+                break;
+            }
+#endif /* __ARM_FEATURE_FP16_VECTOR_ARITHMETIC */
             case DataType::F32:
             {
                 _func = (output == nullptr) ? &output_stage_nhwc<float, float, true, true> : &output_stage_nhwc<float, float, false, true>;
-- 
cgit v1.2.1