Add new operator AddMulAdd for Neon™ backend for Float/Quantized types

This is a fused operator that merges Add + Mul + Add [+ Relu-based-Activation] layers and have an intermediate output after the first Add. It's supported for FP16/32/QASYMM8/QASYMM8_SIGNED data types.

The subsequent Add and Mul are intended for scaling and the coefficients only have one dimension (per channel).

The inputs are
     - input1    : nD tensor   [X, Y, Z, W, ..]
     - input2    : nD tensor   [X, Y, Z, W, ..]
     - add_coef  : 1D tensor   [X]
     - mul_coef  : 1D tensor   [X]

The outputs are
     - out1          : nD tensor (intermediate output)  [X, Y, Z, W, ..]
     - out2          : nD tensor (final output)  [X, Y, Z, W, ..]

The operation can be summarized as follows:
     out1 <- input1 + input2
     out2 <- Act(out1 * mul_coef + add_coef)

The activation function can be Identity, Relu, Bounded Relu or Lower/Upper Bounded Relu. The intermediate output can be skipped by providing a nullptr.

The reason of providing this operator is to be able to fuse in case of Residual network patterns and save computations by reducing memory back and forward.

Resolves: COMPMID-5463

Signed-off-by: Gunes Bayir <gunes.bayir@arm.com>
Change-Id: I8ef577aa623b036e9a9f655cc088493fd19a6109
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/9055
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Jakub Sujak <jakub.sujak@arm.com>
Reviewed-by: Viet-Hoa Do <viet-hoa.do@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Benchmark: Arm Jenkins <bsgcomp@arm.com>

1 files changed, 82 insertions, 0 deletions

diff --git a/src/runtime/NEON/functions/NEAddMulAdd.cpp b/src/runtime/NEON/functions/NEAddMulAdd.cpp
new file mode 100644
index 0000000000..55008de5d6
--- /dev/null
+++ b/src/runtime/NEON/functions/NEAddMulAdd.cpp
@@ -0,0 +1,82 @@
+/*
+ * Copyright (c) 2023 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 "arm_compute/runtime/NEON/functions/NEAddMulAdd.h"
+
+#include "src/common/utils/Log.h"
+#include "src/core/helpers/MemoryHelpers.h"
+#include "src/cpu/operators/CpuAddMulAdd.h"
+
+namespace arm_compute
+{
+struct NEAddMulAdd::Impl
+{
+    std::unique_ptr<cpu::CpuAddMulAdd> op{ nullptr };
+    WorkspaceData<Tensor>              workspace_tensors{};
+    ITensorPack                        run_pack{};
+    MemoryGroup                        memory_group{};
+};
+
+NEAddMulAdd::NEAddMulAdd(std::shared_ptr<IMemoryManager> memory_manager)
+    : _impl(std::make_unique<Impl>())
+{
+    _impl->memory_group = MemoryGroup(std::move(memory_manager));
+}
+
+NEAddMulAdd::~NEAddMulAdd() = default;
+
+void NEAddMulAdd::configure(ITensor *input1, ITensor *input2, ITensor *bn_mul, ITensor *bn_add, ITensor *add_output,
+                            ITensor *final_output, const ConvertPolicy policy, const ActivationLayerInfo &act_info)
+{
+    ARM_COMPUTE_LOG_PARAMS(input1, input2, bn_mul, bn_add, add_output, final_output, policy, act_info);
+
+    _impl->op           = std::make_unique<cpu::CpuAddMulAdd>();
+    _impl->op->configure(input1->info(), input2->info(), bn_mul->info(),
+                         bn_add->info(), add_output != nullptr ? add_output->info() : nullptr, final_output->info(), policy, act_info);
+
+    _impl->run_pack =
+    {
+        { TensorType::ACL_SRC_0, input1 },
+        { TensorType::ACL_SRC_1, input2 },
+        { TensorType::ACL_SRC_2, bn_mul },
+        { TensorType::ACL_SRC_3, bn_add },
+        { TensorType::ACL_DST_0, add_output },
+        { TensorType::ACL_DST_1, final_output },
+    };
+
+    _impl->workspace_tensors = manage_workspace<Tensor>(_impl->op->workspace(), _impl->memory_group, _impl->run_pack);
+}
+
+Status NEAddMulAdd::validate(const ITensorInfo *input1, const ITensorInfo *input2, const ITensorInfo *bn_mul,
+                             const ITensorInfo *bn_add, const ITensorInfo *add_output, const ITensorInfo *final_output,
+                             ConvertPolicy policy, const ActivationLayerInfo &act_info)
+{
+    return cpu::CpuAddMulAdd::validate(input1, input2, bn_mul, bn_add, add_output, final_output, policy, act_info);
+}
+
+void NEAddMulAdd::run()
+{
+    _impl->op->run(_impl->run_pack);
+}
+} // namespace arm_compute