Add CLAMP operator to Dynamic Fusion interface
Add the CLAMP activation function for GPU backend with generic activation Component and TemplateWriter modules.
CLAMP is internally implemented as LU_BOUNDED_RELU activation function with the alpha and beta variables swapped.
We do NOT consider in-place computation cases in this patch.
* CLAMP operator for GPU backend.
* Activation Component and TemplateWriter for CL backend.
* TemplateWriter generates tiled kernel code.
* Supported data types: F16, F32.
* Validation tests for CLAMP operation.
Resolves: COMPMID-5519
Change-Id: Ieb097d6b1e6a7ed2b882518e88314454efb402f6
Signed-off-by: Jakub Sujak <jakub.sujak@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/8762
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Gunes Bayir <gunes.bayir@arm.com>
Reviewed-by: SiCong Li <sicong.li@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Benchmark: Arm Jenkins <bsgcomp@arm.com>
diff --git a/arm_compute/dynamic_fusion/sketch/gpu/operators/GpuClamp.h b/arm_compute/dynamic_fusion/sketch/gpu/operators/GpuClamp.h
new file mode 100644
index 0000000..66d6c5f
--- /dev/null
+++ b/arm_compute/dynamic_fusion/sketch/gpu/operators/GpuClamp.h
@@ -0,0 +1,94 @@
+/*
+ * Copyright (c) 2022 Arm Limited.
+ *
+ * SPDX-License-Identifier: MIT
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in all
+ * copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+#ifndef ARM_COMPUTE_DYNAMIC_FUSION_SKETCH_GPU_OPERATORS_GPUCLAMP
+#define ARM_COMPUTE_DYNAMIC_FUSION_SKETCH_GPU_OPERATORS_GPUCLAMP
+
+#include "arm_compute/dynamic_fusion/sketch/attributes/ClampAttributes.h"
+
+namespace arm_compute
+{
+namespace experimental
+{
+namespace dynamic_fusion
+{
+/** Forward declaration */
+class GpuWorkloadContext;
+class GpuWorkloadSketch;
+
+/** Operator interface. */
+class GpuClamp final
+{
+public:
+ /** Attributes are a set of backend-agnostic parameters that define what an operator does */
+ using Attributes = ClampAttributes;
+
+ /** Create an operator and fuse it into the workload sketch.
+ * @note If @ref validate_op() fails, the creation also fails and may throw an error.
+ * @note If @ref validate_op() fails, @p sketch remains unchanged and valid.
+ *
+ * Valid data type configurations:
+ * |src |dst |
+ * |:--------------|:--------------|
+ * |F16 |F16 |
+ * |F32 |F32 |
+ *
+ * Valid data layouts:
+ * - All
+ *
+ * @param[in, out] sketch Workload sketch into which the operator will be fused
+ * @param[in] src Source tensor info. Data types supported: F16/F32.
+ * @param[out] dst Destination tensor info. Data types supported: F16/F32.
+ * If an uninitialized ITensorInfo is passed in, it will be auto-initialized
+ * @param[in] attributes Operator attributes
+ */
+ static void create_op(GpuWorkloadSketch &sketch,
+ ITensorInfo *src,
+ ITensorInfo *dst,
+ const Attributes &attributes);
+
+ /** Check if the operator configuration is supported, irrespective of fusion
+ *
+ * @param[in] context Workload context within which the operator is running
+ * @param[in] src Source tensor info. Data types supported: F16/F32.
+ * @param[in] dst Destination tensor info. Data types supported: F16/F32.
+ * If an uninitialized ITensorInfo is passed in, it will be auto-initialized
+ * @param[in] attributes Operator attributes
+ */
+ static Status is_supported_op(const GpuWorkloadContext &context,
+ const ITensorInfo *src,
+ const ITensorInfo *dst,
+ const Attributes &attributes);
+
+ /** Validate the operator and check if it can be fused into the workload sketch.
+ * Similar to @ref GpuClamp::create_op()
+ */
+ static Status validate_op(const GpuWorkloadSketch &sketch,
+ const ITensorInfo *src,
+ const ITensorInfo *dst,
+ const Attributes &attributes);
+};
+} // namespace dynamic_fusion
+} // namespace experimental
+} // namespace arm_compute
+#endif //ARM_COMPUTE_DYNAMIC_FUSION_SKETCH_GPU_OPERATORS_GPUCLAMP