Use the stable CKW API in the GPU dynamic fusion backend

- Refactor all kernels to work with the CKW stable API
- Add support for sub-tile in the op_load/op_store CKW operator
- Fix mismatch in resize
- Add comments in all kernels written with CKW to help developers
understand the structure of the code
- Add texture image support in depthwise convolution written with CKW
- Add support for different block sizes in depthwise convolution
- Remove the use of the dynamic fusion helper functions.
- Add support for floor in the op_unary() of CKW

Resolves: COMPMID-6708, COMPMID-6743, COMPMID-6530

Signed-off-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Signed-off-by: Gunes Bayir <gunes.bayir@arm.com>
Signed-off-by: Viet-Hoa Do <viet-hoa.do@arm.com>
Signed-off-by: Jakub Sujak <jakub.sujak@arm.com>

Change-Id: I8104ce4d04a3138a1aeb0b84940e1f1c89e76069
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/10914
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Jakub Sujak <jakub.sujak@arm.com>
Reviewed-by: Gunes Bayir <gunes.bayir@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Benchmark: Arm Jenkins <bsgcomp@arm.com>

1 files changed, 23 insertions, 27 deletions

diff --git a/src/dynamic_fusion/sketch/gpu/ckw_driver/GpuCkwComponentArgument.cpp b/src/dynamic_fusion/sketch/gpu/ckw_driver/GpuCkwComponentArgument.cpp
index c4ab110c92..a42b39700c 100644
--- a/src/dynamic_fusion/sketch/gpu/ckw_driver/GpuCkwComponentArgument.cpp
+++ b/src/dynamic_fusion/sketch/gpu/ckw_driver/GpuCkwComponentArgument.cpp
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2023 Arm Limited.
+ * Copyright (c) 2023-2024 Arm Limited.
  *
  * SPDX-License-Identifier: MIT
  *
@@ -24,7 +24,7 @@
 
 #include "src/dynamic_fusion/sketch/gpu/ckw_driver/GpuCkwComponentArgument.h"
 
-#include "ckw/Error.h"
+#include "compute_kernel_writer/include/ckw/Error.h"
 
 namespace arm_compute
 {
@@ -33,75 +33,71 @@ namespace experimental
 namespace dynamic_fusion
 {
 
-GpuCkwComponentArgument::GpuCkwComponentArgument()
+GpuCkwComponentArgument::GpuCkwComponentArgument(ckw::TensorOperand tensor) : _tensor(tensor)
 {
 }
 
-GpuCkwComponentArgument::GpuCkwComponentArgument(ckw::TensorOperand &tensor) : _tensor(&tensor)
-{
-}
-
-GpuCkwComponentArgument &GpuCkwComponentArgument::init_virtual_tensor(ckw::TileOperand             &tile,
-                                                                      const ckw::TensorTileSampler &tile_sampler)
+GpuCkwComponentArgument &GpuCkwComponentArgument::init_virtual_tensor(ckw::TileOperand         &tile,
+                                                                      const ckw::TensorSampler &sampler)
 {
     CKW_ASSERT(_tile == nullptr);
 
-    _tile         = &tile;
-    _tile_sampler = tile_sampler;
+    _tile    = tile;
+    _sampler = sampler;
 
     return *this;
 }
 
 bool GpuCkwComponentArgument::has_tensor() const
 {
-    return _tensor != nullptr;
+    return _tensor.is_valid();
 }
 
 ckw::TensorOperand &GpuCkwComponentArgument::tensor()
 {
-    CKW_ASSERT(_tensor != nullptr);
+    CKW_ASSERT(_tensor.is_valid());
 
-    return *_tensor;
+    return _tensor;
 }
 
 const ckw::TensorOperand &GpuCkwComponentArgument::tensor() const
 {
-    CKW_ASSERT(_tensor != nullptr);
+    CKW_ASSERT(_tensor.is_valid());
 
-    return *_tensor;
+    return _tensor;
 }
 
 bool GpuCkwComponentArgument::has_tile() const
 {
-    return _tile != nullptr;
+    return _tile.is_valid();
 }
 
 ckw::TileOperand &GpuCkwComponentArgument::tile()
 {
-    CKW_ASSERT(_tile != nullptr);
+    CKW_ASSERT(_tile.is_valid());
 
-    return *_tile;
+    return _tile;
 }
 
 const ckw::TileOperand &GpuCkwComponentArgument::tile() const
 {
-    CKW_ASSERT(_tile != nullptr);
+    CKW_ASSERT(_tile.is_valid());
 
-    return *_tile;
+    return _tile;
 }
 
-ckw::TensorTileSampler &GpuCkwComponentArgument::tile_sampler()
+ckw::TensorSampler &GpuCkwComponentArgument::tensor_sampler()
 {
-    CKW_ASSERT(_tile != nullptr);
+    CKW_ASSERT(_tile.is_valid());
 
-    return _tile_sampler;
+    return _sampler;
 }
 
-const ckw::TensorTileSampler &GpuCkwComponentArgument::tile_sampler() const
+const ckw::TensorSampler &GpuCkwComponentArgument::tensor_sampler() const
 {
-    CKW_ASSERT(_tile != nullptr);
+    CKW_ASSERT(_tile.is_valid());
 
-    return _tile_sampler;
+    return _sampler;
 }
 
 } // namespace dynamic_fusion