COMPMID-2754: Add support for QASYMM8_SIGNED in NE kernels/functions.

Kernels/Functions extended support:
- NEBatchToSpaceLayerKernel/NEBatchToSpaceLayer
- NEChannelShuffleLayerKernel/NEChannelShuffleLayer
- NECol2ImKernel/NECol2Im
- NEConvertFullyConnectedWeightsKernel/NEConvertFullyConnectedWeights
- NECopyKernel/NECopy
- NEConvolutionLayerReshapeWeights
- NEDepthToSpaceLayerKernel/NEDepthToSpaceLayer
- NEFlattenLayerKernel/NEFlattenLayer
- NEFillBorderKernel
- NEFullyConnectedLayerReshapeWeights
- NEGatherKernel/NEGather
- NEGEMMInterleave4x4Kernel
- NEGEMMTranspose1xWKernel
- NEIm2ColKernel/NEIm2Col
- NEMemsetKernel
- NEPadLayerKernel/NEPadLayer
- NEPermuteKernel/NEPermute
- NEReverseKernel/NEReverse
- NEReorgLayerKernel/NEReorgLayer
- NEReshapeLayerKernel/NEReshapeLayer
- NESplit
- NESlice
- NEStridedSliceKernel/NEStridedSlice
- NESpaceToBatchLayerKernel/NESpaceToBatchLayer
- NESpaceToDepthLayerKernel/NESpaceToDepthLayerKernel
- NEStackLayerKernel/NEStackLayer
- NETileKernel/NETile
- NETransposeKernel/NETranspose
- NEWidthConcatenateLayerKernel/NEHeightConcatenateLayer
- NEHeightConcatenateLayerKernel/NEHeightConcatenateLayer
- NEDepthConcatenateLayerKernel/NEDepthConcatenateLayer
- NEBathConcatenateLayerKernel/NEBatchConcatenateLayer

Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com>
Change-Id: Ia070332ad4c4dbced2541dc46f7f2f3a86833b65
Reviewed-on: https://review.mlplatform.org/c/2442
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>

1 files changed, 13 insertions, 6 deletions

diff --git a/src/core/NEON/kernels/NEHeightConcatenateLayerKernel.cpp b/src/core/NEON/kernels/NEHeightConcatenateLayerKernel.cpp
index 18be1d5bb7..4f7ae8c0b1 100644
--- a/src/core/NEON/kernels/NEHeightConcatenateLayerKernel.cpp
+++ b/src/core/NEON/kernels/NEHeightConcatenateLayerKernel.cpp
@@ -36,8 +36,8 @@
 
 #include <cstdint>
 
-using namespace arm_compute;
-
+namespace arm_compute
+{
 namespace
 {
 std::pair<Status, Window> validate_and_configure_window(ITensorInfo *input, ITensorInfo *output)
@@ -58,10 +58,7 @@ Status validate_arguments(const ITensorInfo *input, unsigned int height_offset,
 {
     ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(input, output);
     // Note: ARM_COMPUTE_RETURN_ERROR_ON_CPU_F16_UNSUPPORTED(input) is not needed here as this kernel doesn't use NEON FP16 instructions.
-    ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(input, 1,
-                                                         DataType::U8, DataType::S8, DataType::QASYMM8,
-                                                         DataType::U16, DataType::S16, DataType::F16,
-                                                         DataType::U32, DataType::F32);
+    ARM_COMPUTE_RETURN_ERROR_ON(input->data_type() == DataType::UNKNOWN);
     ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(input, output);
     ARM_COMPUTE_RETURN_ERROR_ON(input->dimension(Window::DimX) != output->dimension(Window::DimX));
     ARM_COMPUTE_RETURN_ERROR_ON(input->dimension(Window::DimY) + height_offset > output->dimension(Window::DimY));
@@ -128,6 +125,15 @@ void NEHeightConcatenateLayerKernel::run(const Window &window, const ThreadInfo
         },
         input, output);
     }
+    else if(dt == DataType::QASYMM8_SIGNED && input_qinfo != output_qinfo)
+    {
+        execute_window_loop(window, [&](const Coordinates &)
+        {
+            vst1q_s8(reinterpret_cast<int8_t *>(output_ptr + output.offset()),
+                     vquantize_signed(vdequantize(vld1q_s8(reinterpret_cast<int8_t *>(input.ptr())), input_qinfo), output_qinfo));
+        },
+        input, output);
+    }
     else
     {
         execute_window_loop(window, [&](const Coordinates &)
@@ -140,3 +146,4 @@ void NEHeightConcatenateLayerKernel::run(const Window &window, const ThreadInfo
         input, output);
     }
 }
+} // namespace arm_compute