Make CpuPixelWiseMultiplicationKernel stateless

Resolves: COMPMID-4183

Signed-off-by: Sheri Zhang <sheri.zhang@arm.com>
Change-Id: Ie535c4129a6164b879fb5c4acb15f2be58ee8b6c
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5325
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Michalis Spyrou <michalis.spyrou@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>

3 files changed, 226 insertions, 52 deletions

diff --git a/src/runtime/NEON/functions/NEPixelWiseMultiplication.cpp b/src/runtime/NEON/functions/NEPixelWiseMultiplication.cpp
index 179bcdaf3e..4d7fef89ed 100644
--- a/src/runtime/NEON/functions/NEPixelWiseMultiplication.cpp
+++ b/src/runtime/NEON/functions/NEPixelWiseMultiplication.cpp
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2016-2020 Arm Limited.
+ * Copyright (c) 2016-2021 Arm Limited.
  *
  * SPDX-License-Identifier: MIT
  *
@@ -24,64 +24,30 @@
 #include "arm_compute/runtime/NEON/functions/NEPixelWiseMultiplication.h"
 
 #include "arm_compute/core/ITensor.h"
-#include "src/core/NEON/kernels/NEPixelWiseMultiplicationKernel.h"
+#include "src/runtime/cpu/operators/CpuPixelWiseMultiplication.h"
 
 #include <utility>
 
 namespace arm_compute
 {
-namespace experimental
-{
-void NEPixelWiseMultiplication::configure(ITensorInfo *input1, ITensorInfo *input2, ITensorInfo *output, float scale, ConvertPolicy overflow_policy, RoundingPolicy rounding_policy,
-                                          const ActivationLayerInfo &act_info)
-{
-    ARM_COMPUTE_UNUSED(act_info);
-    auto k = std::make_unique<NEPixelWiseMultiplicationKernel>();
-    k->configure(input1, input2, output, scale, overflow_policy, rounding_policy);
-    _kernel = std::move(k);
-}
-Status NEPixelWiseMultiplication::validate(const ITensorInfo *input1, const ITensorInfo *input2, const ITensorInfo *output, float scale, ConvertPolicy overflow_policy, RoundingPolicy rounding_policy,
-                                           const ActivationLayerInfo &act_info)
-{
-    ARM_COMPUTE_RETURN_ERROR_ON(act_info.enabled());
-    return NEPixelWiseMultiplicationKernel::validate(input1, input2, output, scale, overflow_policy, rounding_policy);
-}
-
-void NEComplexPixelWiseMultiplication::configure(ITensorInfo *input1, ITensorInfo *input2, ITensorInfo *output, const ActivationLayerInfo &act_info)
-{
-    ARM_COMPUTE_UNUSED(act_info);
-    auto k = std::make_unique<NEComplexPixelWiseMultiplicationKernel>();
-    k->configure(input1, input2, output);
-    _kernel = std::move(k);
-}
-
-Status NEComplexPixelWiseMultiplication::validate(const ITensorInfo *input1, const ITensorInfo *input2, const ITensorInfo *output, const ActivationLayerInfo &act_info)
-{
-    ARM_COMPUTE_RETURN_ERROR_ON(act_info.enabled());
-    return NEComplexPixelWiseMultiplicationKernel::validate(input1, input2, output);
-}
-} // namespace experimental
-
 struct NEPixelWiseMultiplication::Impl
 {
-    const ITensor                                           *src_0{ nullptr };
-    const ITensor                                           *src_1{ nullptr };
-    ITensor                                                 *dst{ nullptr };
-    std::unique_ptr<experimental::NEPixelWiseMultiplication> op{ nullptr };
+    const ITensor                                   *src_0{ nullptr };
+    const ITensor                                   *src_1{ nullptr };
+    ITensor                                         *dst{ nullptr };
+    std::unique_ptr<cpu::CpuPixelWiseMultiplication> op{ nullptr };
 };
 
 NEPixelWiseMultiplication::NEPixelWiseMultiplication()
     : _impl(std::make_unique<Impl>())
 {
 }
-NEPixelWiseMultiplication::NEPixelWiseMultiplication(NEPixelWiseMultiplication &&) = default;
-NEPixelWiseMultiplication &NEPixelWiseMultiplication::operator=(NEPixelWiseMultiplication &&) = default;
-NEPixelWiseMultiplication::~NEPixelWiseMultiplication()                                       = default;
+NEPixelWiseMultiplication::~NEPixelWiseMultiplication() = default;
 
 Status NEPixelWiseMultiplication::validate(const ITensorInfo *input1, const ITensorInfo *input2, const ITensorInfo *output, float scale, ConvertPolicy overflow_policy, RoundingPolicy rounding_policy,
                                            const ActivationLayerInfo &act_info)
 {
-    return experimental::NEPixelWiseMultiplication::validate(input1, input2, output, scale, overflow_policy, rounding_policy, act_info);
+    return cpu::CpuPixelWiseMultiplication::validate(input1, input2, output, scale, overflow_policy, rounding_policy, act_info);
 }
 
 void NEPixelWiseMultiplication::configure(const ITensor *input1, const ITensor *input2, ITensor *output, float scale, ConvertPolicy overflow_policy, RoundingPolicy rounding_policy,
@@ -90,7 +56,7 @@ void NEPixelWiseMultiplication::configure(const ITensor *input1, const ITensor *
     _impl->src_0 = input1;
     _impl->src_1 = input2;
     _impl->dst   = output;
-    _impl->op    = std::make_unique<experimental::NEPixelWiseMultiplication>();
+    _impl->op    = std::make_unique<cpu::CpuPixelWiseMultiplication>();
     _impl->op->configure(input1->info(), input2->info(), output->info(), scale, overflow_policy, rounding_policy, act_info);
 }
 
@@ -105,23 +71,21 @@ void NEPixelWiseMultiplication::run()
 
 struct NEComplexPixelWiseMultiplication::Impl
 {
-    ITensor                                                        *src_0{ nullptr };
-    ITensor                                                        *src_1{ nullptr };
-    ITensor                                                        *dst{ nullptr };
-    std::unique_ptr<experimental::NEComplexPixelWiseMultiplication> op{ nullptr };
+    ITensor                                                *src_0{ nullptr };
+    ITensor                                                *src_1{ nullptr };
+    ITensor                                                *dst{ nullptr };
+    std::unique_ptr<cpu::CpuComplexPixelWiseMultiplication> op{ nullptr };
 };
 
 NEComplexPixelWiseMultiplication::NEComplexPixelWiseMultiplication()
     : _impl(std::make_unique<Impl>())
 {
 }
-NEComplexPixelWiseMultiplication::NEComplexPixelWiseMultiplication(NEComplexPixelWiseMultiplication &&) = default;
-NEComplexPixelWiseMultiplication &NEComplexPixelWiseMultiplication::operator=(NEComplexPixelWiseMultiplication &&) = default;
-NEComplexPixelWiseMultiplication::~NEComplexPixelWiseMultiplication()                                              = default;
+NEComplexPixelWiseMultiplication::~NEComplexPixelWiseMultiplication() = default;
 
 Status NEComplexPixelWiseMultiplication::validate(const ITensorInfo *input1, const ITensorInfo *input2, const ITensorInfo *output, const ActivationLayerInfo &act_info)
 {
-    return experimental::NEComplexPixelWiseMultiplication::validate(input1, input2, output, act_info);
+    return cpu::CpuComplexPixelWiseMultiplication::validate(input1, input2, output, act_info);
 }
 
 void NEComplexPixelWiseMultiplication::configure(ITensor *input1, ITensor *input2, ITensor *output, const ActivationLayerInfo &act_info)
@@ -129,7 +93,7 @@ void NEComplexPixelWiseMultiplication::configure(ITensor *input1, ITensor *input
     _impl->src_0 = input1;
     _impl->src_1 = input2;
     _impl->dst   = output;
-    _impl->op    = std::make_unique<experimental::NEComplexPixelWiseMultiplication>();
+    _impl->op    = std::make_unique<cpu::CpuComplexPixelWiseMultiplication>();
     _impl->op->configure(input1->info(), input2->info(), output->info(), act_info);
 }
 
diff --git a/src/runtime/cpu/operators/CpuPixelWiseMultiplication.cpp b/src/runtime/cpu/operators/CpuPixelWiseMultiplication.cpp
new file mode 100644
index 0000000000..2e560d7490
--- /dev/null
+++ b/src/runtime/cpu/operators/CpuPixelWiseMultiplication.cpp
@@ -0,0 +1,77 @@
+/*
+ * Copyright (c) 2016-2021 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.
+ */
+#include "src/runtime/cpu/operators/CpuPixelWiseMultiplication.h"
+
+#include "arm_compute/core/TensorInfo.h"
+#include "arm_compute/core/Validate.h"
+#include "arm_compute/runtime/NEON/NEScheduler.h"
+#include "src/core/cpu/kernels/CpuPixelWiseMultiplicationKernel.h"
+
+namespace arm_compute
+{
+namespace cpu
+{
+Status CpuPixelWiseMultiplication::validate(const ITensorInfo *input1, const ITensorInfo *input2, const ITensorInfo *output, float scale, ConvertPolicy overflow_policy, RoundingPolicy rounding_policy,
+                                            const ActivationLayerInfo &act_info)
+{
+    ARM_COMPUTE_RETURN_ERROR_ON(act_info.enabled());
+    return kernels::CpuPixelWiseMultiplicationKernel::validate(input1, input2, output, scale, overflow_policy, rounding_policy);
+}
+
+void CpuPixelWiseMultiplication::configure(ITensorInfo *input1, ITensorInfo *input2, ITensorInfo *output, float scale, ConvertPolicy overflow_policy, RoundingPolicy rounding_policy,
+                                           const ActivationLayerInfo &act_info)
+{
+    ARM_COMPUTE_UNUSED(act_info);
+    auto k = std::make_unique<kernels::CpuPixelWiseMultiplicationKernel>();
+    k->configure(input1, input2, output, scale, overflow_policy, rounding_policy);
+    _kernel = std::move(k);
+}
+
+void CpuPixelWiseMultiplication::run(ITensorPack &tensors)
+{
+    ARM_COMPUTE_ERROR_ON_MSG(tensors.empty(), "No inputs provided");
+    NEScheduler::get().schedule_op(_kernel.get(), Window::DimY, _kernel->window(), tensors);
+}
+
+Status CpuComplexPixelWiseMultiplication::validate(const ITensorInfo *input1, const ITensorInfo *input2, const ITensorInfo *output, const ActivationLayerInfo &act_info)
+{
+    ARM_COMPUTE_RETURN_ERROR_ON(act_info.enabled());
+    return kernels::CpuComplexPixelWiseMultiplicationKernel::validate(input1, input2, output);
+}
+
+void CpuComplexPixelWiseMultiplication::configure(ITensorInfo *input1, ITensorInfo *input2, ITensorInfo *output, const ActivationLayerInfo &act_info)
+{
+    ARM_COMPUTE_UNUSED(act_info);
+    auto k = std::make_unique<kernels::CpuComplexPixelWiseMultiplicationKernel>();
+    k->configure(input1, input2, output);
+    _kernel = std::move(k);
+}
+
+void CpuComplexPixelWiseMultiplication::run(ITensorPack &tensors)
+{
+    ARM_COMPUTE_ERROR_ON_MSG(tensors.empty(), "No inputs provided");
+    NEScheduler::get().schedule_op(_kernel.get(), Window::DimY, _kernel->window(), tensors);
+}
+} // namespace cpu
+} // namespace arm_compute
\ No newline at end of file
diff --git a/src/runtime/cpu/operators/CpuPixelWiseMultiplication.h b/src/runtime/cpu/operators/CpuPixelWiseMultiplication.h
new file mode 100644
index 0000000000..b2cd7d529b
--- /dev/null
+++ b/src/runtime/cpu/operators/CpuPixelWiseMultiplication.h
@@ -0,0 +1,133 @@
+/*
+ * Copyright (c) 2016-2021 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_CPU_PIXELWISEMULTIPLICATION_H
+#define ARM_COMPUTE_CPU_PIXELWISEMULTIPLICATION_H
+
+#include "arm_compute/core/ITensorInfo.h"
+#include "arm_compute/core/experimental/Types.h"
+#include "src/core/cpu/ICpuKernel.h"
+#include "src/runtime/cpu/ICpuOperator.h"
+
+#include <memory>
+
+namespace arm_compute
+{
+namespace cpu
+{
+/** Basic function to run @ref kernels::CpuPixelWiseMultiplicationKernel */
+class CpuPixelWiseMultiplication : public ICpuOperator
+{
+public:
+    /** Default Constructor */
+    CpuPixelWiseMultiplication() = default;
+    /** Initialise the kernel's inputs, output and convertion policy.
+     *
+     * @note For @p scale equal to 1/255 only round to nearest even (implemented as round half up) is supported.
+     *       For all other scale values only round to zero (implemented as round towards minus infinity) is supported.
+     *
+     * @param[in, out] input1          First input tensor info. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/S32/QSYMM16/F16/F32
+     *                                 This input tensor is [in, out] because its TensorInfo might be modified inside the kernel in case of broadcasting of dimension 0.
+     * @param[in, out] input2          Second input tensor info. Data types supported: U8, QASYMM8 (only if @p input1 is QASYMM8), QASYMM8_SIGNED (only if @p input1 is QASYMM8_SIGNED), S16, S32, QSYMM16 (only if @p input1 is QSYMM16), F16 (only if @p input1 is F16), F32 (only if @p input1 is F32).
+     *                                 This input tensor is [in, out] because its TensorInfo might be modified inside the kernel in case of broadcasting of dimension 0.
+     * @param[out]     output          Output tensor info. Data types supported:
+     *                                 - U8, only if both inputs are U8.
+     *                                 - QASYMM8, only if both inputs are QASYMM8.
+     *                                 - QASYMM8_SIGNED, only if @p input1 is QASYMM8_SIGNED.
+     *                                 - S16.
+     *                                 - QSYMM16, only if both inputs are QSYMM16.
+     *                                 - S32, only if both inputs are S32 or both are QSYMM16.
+     *                                 - F16, only if @p input1 is F16.
+     *                                 - F32, only if both inputs are F32.
+     * @param[in]      scale           Scale to apply after multiplication.
+     *                                 Scale must be positive and its value must be either 1/255 or 1/2^n where n is between 0 and 15.
+     *                                 If both @p input1, @p input2 and @p output are of datatype S32, scale cannot be 1/255
+     * @param[in]      overflow_policy Overflow policy. ConvertPolicy cannot be WRAP if any of the inputs is of quantized datatype
+     * @param[in]      rounding_policy Rounding policy.
+     * @param[in]      act_info        (Optional) Activation layer information in case of a fused activation. Currently not supported.
+     */
+    void configure(ITensorInfo *input1, ITensorInfo *input2, ITensorInfo *output, float scale, ConvertPolicy overflow_policy, RoundingPolicy rounding_policy,
+                   const ActivationLayerInfo &act_info = ActivationLayerInfo());
+    /** Static function to check if given info will lead to a valid configuration of @ref CpuPixelWiseMultiplication
+     *
+     * @note For @p scale equal to 1/255 only round to nearest even (implemented as round half up) is supported.
+     *       For all other scale values only round to zero (implemented as round towards minus infinity) is supported.
+     *
+     * @param[in] input1          First input tensor info. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/S32/QSYMM16/F16/F32
+     * @param[in] input2          Second input tensor info. Data types supported: U8, QASYMM8 (only if @p input1 is QASYMM8), QASYMM8_SIGNED (only if @p input1 is QASYMM8_SIGNED), S16, S32, QSYMM16 (only if both inputs are QSYMM16), F16 (only if @p input1 is F16), F32 (only if @p input1 is F32).
+     * @param[in] output          Output tensor info. Data types supported:
+     *                            - U8, only if both inputs are U8.
+     *                            - QASYMM8, only if both inputs are QASYMM8.
+     *                            - QASYMM8_SIGNED, only if @p input1 is QASYMM8_SIGNED.
+     *                            - S16.
+     *                            - QSYMM16, only if both inputs are QSYMM16.
+     *                            - S32, only if both inputs are S32 or both are QSYMM16.
+     *                            - F16, only if @p input1 is F16.
+     *                            - F32, only if both inputs are F32.
+     * @param[in] scale           Scale to apply after multiplication.
+     *                            Scale must be positive and its value must be either 1/255 or 1/2^n where n is between 0 and 15.
+     *                            If both @p input1, @p input2 and @p output are of datatype S32, scale cannot be 1/255
+     * @param[in] overflow_policy Overflow policy. ConvertPolicy cannot be WRAP if any of the inputs is of quantized datatype
+     * @param[in] rounding_policy Rounding policy.
+     * @param[in] act_info        (Optional) Activation layer information in case of a fused activation. Currently not supported.
+     *
+     * @return a status
+     */
+    static Status validate(const ITensorInfo *input1, const ITensorInfo *input2, const ITensorInfo *output, float scale, ConvertPolicy overflow_policy, RoundingPolicy rounding_policy,
+                           const ActivationLayerInfo &act_info = ActivationLayerInfo());
+
+    // Inherited methods overridden:
+    void run(ITensorPack &tensors) override;
+};
+
+/** Basic function to run @ref kernels::CpuComplexPixelWiseMultiplicationKernel. */
+class CpuComplexPixelWiseMultiplication : public ICpuOperator
+{
+public:
+    /** Default Constructor */
+    CpuComplexPixelWiseMultiplication() = default;
+    /** Initialise the kernel's inputs, output.
+     *
+     * @param[in, out] input1   First input tensor. Data types supported: F32. Number of channels supported: 2 (complex tensor).
+     *                          The input tensor is [in, out] because its TensorInfo might be modified inside the kernel in case of broadcasting of dimension 0.
+     * @param[in, out] input2   Second input tensor. Data types supported: same as @p input1. Number of channels supported: same as @p input1.
+     *                          The input tensor is [in, out] because its TensorInfo might be modified inside the kernel in case of broadcasting of dimension 0.
+     * @param[out]     output   The output tensor. Data types supported: same as @p input1. Number of channels: same as @p input1.
+     * @param[in]      act_info (Optional) Activation layer information in case of a fused activation. Currently not supported.
+     */
+    void configure(ITensorInfo *input1, ITensorInfo *input2, ITensorInfo *output, const ActivationLayerInfo &act_info = ActivationLayerInfo());
+    /** Static function to check if given info will lead to a valid configuration of @ref CpuComplexPixelWiseMultiplication
+     *
+     * @param[in] input1   First input tensor info. Data types supported: F32. Number of channels supported: 2 (complex tensor).
+     * @param[in] input2   Second input tensor info. Data types supported: same as @p input1. Number of channels supported: same as @p input1.
+     * @param[in] output   The output tensor info. Data types supported: same as @p input1. Number of channels supported: same as @p input1.
+     * @param[in] act_info (Optional) Activation layer information in case of a fused activation. Currently not supported.
+     */
+    static Status validate(const ITensorInfo *input1, const ITensorInfo *input2, const ITensorInfo *output, const ActivationLayerInfo &act_info = ActivationLayerInfo());
+
+    // Inherited methods overridden:
+    void run(ITensorPack &tensors) override;
+};
+} // namespace cpu
+} // namespace arm_compute
+#endif /* ARM_COMPUTE_CPU_PIXELWISEMULTIPLICATION_H */
\ No newline at end of file