commit b84e25313e5dc7acbc03623e1e071e845047c111
author Viet-Hoa Do <viet-hoa.do@arm.com> Tue Dec 13 13:09:10 2022 +0000
committer Viet-Hoa Do <viet-hoa.do@arm.com> Fri Dec 16 15:17:51 2022 +0000
tree fbee083f1262017555c64c3280da45e2b638992e
parent a0ae8d2e6c57fd95c0edaf659b9df8b8c540d051

Add output operator for dynamic fusion

* The output of the fused operator must be explicitly specified
  using GpuOutput operator.
* Any temporary tensors used to connect the output of an operator
  to the input of another operator will be marked as no-alloc
  and won't be allocated as a tensor in the memory.

Resolves: COMPMID-5771
Signed-off-by: Viet-Hoa Do <viet-hoa.do@arm.com>
Change-Id: I5ae8e800f8f737db23a055a92b01c4f1d78c3bb8
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/8794
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: SiCong Li <sicong.li@arm.com>
Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Reviewed-by: Gunes Bayir <gunes.bayir@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Benchmark: Arm Jenkins <bsgcomp@arm.com>

tests/validation/fixtures/dynamic_fusion/gpu/cl/DirectConv2dFixture.h

diff --git a/tests/validation/fixtures/dynamic_fusion/gpu/cl/DirectConv2dFixture.h b/tests/validation/fixtures/dynamic_fusion/gpu/cl/DirectConv2dFixture.h
index e437c44..1a2676c 100644
--- a/tests/validation/fixtures/dynamic_fusion/gpu/cl/DirectConv2dFixture.h
+++ b/tests/validation/fixtures/dynamic_fusion/gpu/cl/DirectConv2dFixture.h
@@ -32,6 +32,7 @@
 #include "arm_compute/dynamic_fusion/sketch/OperatorAttributes.h"
 #include "arm_compute/dynamic_fusion/sketch/gpu/GpuWorkloadSketch.h"
 #include "arm_compute/dynamic_fusion/sketch/gpu/operators/GpuConv2d.h"
+#include "arm_compute/dynamic_fusion/sketch/gpu/operators/GpuOutput.h"
 
 #include "tests/CL/CLAccessor.h"
 #include "tests/framework/Fixture.h"
@@ -113,7 +114,11 @@
         auto weight_info = sketch.create_tensor_info(TensorInfo(weights_shape, 1, _data_type, _data_layout));
         auto bias_info   = sketch.create_tensor_info(TensorInfo(bias_shape, 1, _data_type, _data_layout));
         auto dst_info    = sketch.create_tensor_info();
-        FunctionType::create_op(sketch, &input_info, &weight_info, &bias_info, &dst_info, conv2d_attr);
+
+        auto ans_info    = sketch.create_tensor_info();
+
+        FunctionType::create_op(sketch, &input_info, &weight_info, &bias_info, &ans_info, conv2d_attr);
+        GpuOutput::create_op(sketch, &ans_info, &dst_info);
 
         // Configure runtime
         ClWorkloadRuntime runtime;