commit bc5a5d59e07d2365c2c88b3430b4a311d186ebbc
author Tracy Narine <tracy.narine@arm.com> Tue Feb 06 15:22:41 2024 +0000
committer TeresaARM <teresa.charlinreyes@arm.com> Wed Feb 07 21:04:02 2024 +0000
tree 345691d4b83824e13b99348617dd68866e5f530f
parent 7bcae3c835468d9b0770514dc7127f02d47cec5f

IVGCVSW-7622 GpuFsa Op: Add Cast operator

  * Added cast operator support

Signed-off-by: Tracy Narine <tracy.narine@arm.com>
Change-Id: Ie12cb1559a7a059ff35e1c395bc77243499243cd

src/backends/gpuFsa/GpuFsaBackend.cpp

diff --git a/src/backends/gpuFsa/GpuFsaBackend.cpp b/src/backends/gpuFsa/GpuFsaBackend.cpp
index f14687b..4a41030 100644
--- a/src/backends/gpuFsa/GpuFsaBackend.cpp
+++ b/src/backends/gpuFsa/GpuFsaBackend.cpp
@@ -20,6 +20,7 @@
 #include <arm_compute/core/CL/CLKernelLibrary.h>
 #include <arm_compute/runtime/CL/CLBufferAllocator.h>
 
+#include "layers/GpuFsaCast.hpp"
 #include "layers/GpuFsaConvolution2d.hpp"
 #include "layers/GpuFsaDepthwiseConvolution2d.hpp"
 #include "layers/GpuFsaElementwiseBinaryAdd.hpp"
@@ -247,6 +248,13 @@
         // Configure and setup the sketch for each supported op. Their data will be wrapped into a PreCompiled layer
         switch (base.GetType())
         {
+            case (LayerType::Cast):
+            {
+                auto input  = base.GetInputSlot(0).GetConnectedOutputSlot()->GetTensorInfo();
+                auto output = base.GetOutputSlot(0).GetTensorInfo();
+                GpuFsaCastCreateOp(preCompiledBlobPtr, input, output);
+                break;
+            }
             case (LayerType::Convolution2d):
             {
                 auto input = base.GetInputSlot(0).GetConnectedOutputSlot()->GetTensorInfo();