src/graph/mutators/NodeFusionMutator.cpp

/*
 * Copyright (c) 2018-2021 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 "arm_compute/graph/mutators/NodeFusionMutator.h"

#include "arm_compute/graph/GraphBuilder.h"
#include "arm_compute/graph/Logger.h"
#include "arm_compute/graph/Utils.h"
#include "arm_compute/graph/backends/BackendRegistry.h"
#include "arm_compute/graph/nodes/FusedConvolutionBatchNormalizationNode.h"
#include "arm_compute/graph/nodes/Nodes.h"

#include "support/Cast.h"

#include <set>

namespace arm_compute
{
namespace graph
{
namespace detail
{
void fuse_convolution_with_batch_normalization(Graph &g, const Edge *output_edge)
{
    ARM_COMPUTE_ERROR_ON(output_edge == nullptr);

    auto *conv_node = arm_compute::utils::cast::polymorphic_downcast<ConvolutionLayerNode *>(output_edge->producer());
    auto *bn_node   = arm_compute::utils::cast::polymorphic_downcast<BatchNormalizationLayerNode *>(output_edge->consumer());

    // Not fusing if number of groups is greater than 1
    if(conv_node->num_groups() > 1)
    {
        return;
    }

    ARM_COMPUTE_LOG_GRAPH_VERBOSE("Fusing convolution node with ID : " << output_edge->producer_id()
                                  << " with BatchNormalization Layer node with ID : " << output_edge->consumer_id() << std::endl);

    // Prevent fusion if fused node has an output accessor
    if(conv_node->output(0)->accessor() == nullptr)
    {
        const Target assigned_target = conv_node->assigned_target();

        // Extract conv inputs
        const auto   conv_input_id   = conv_node->input_edge(0)->producer_id();
        const auto   conv_weights_id = conv_node->input_edge(1)->producer_id();
        const auto   conv_info       = conv_node->convolution_info();
        const auto   conv_method     = conv_node->convolution_method();
        const auto   num_groups      = conv_node->num_groups();
        const auto   act_info        = bn_node->fused_activation();
        FastMathHint fast_math_hint  = conv_node->fast_math_hint();

        // Extract bn inputs
        const auto bn_mean_id = bn_node->input_edge(1)->producer_id();
        const auto bn_var_id  = bn_node->input_edge(2)->producer_id();

        const auto epsilon = bn_node->epsilon();

        // Create the fused node
        const NodeID fused_id = g.add_node<FusedConvolutionBatchNormalizationNode>(epsilon, conv_info, num_groups, conv_method, fast_math_hint, act_info);

        if(conv_node->input_edge(2) != nullptr)
        {
            auto conv_bias_id = conv_node->input_edge(2)->producer_id();
            g.add_connection(conv_bias_id, 0, fused_id, 2);
        }

        // Add connections from the conv/batch_norm inputs to the fused node
        g.add_connection(conv_input_id, 0, fused_id, 0);
        g.add_connection(conv_weights_id, 0, fused_id, 1);
        g.add_connection(bn_mean_id, 0, fused_id, 3);
        g.add_connection(bn_var_id, 0, fused_id, 4);

        if(bn_node->input_edge(3) != nullptr)
        {
            const auto bn_beta_id = bn_node->input_edge(3)->producer_id();
            g.add_connection(bn_beta_id, 0, fused_id, 5);
        }

        if(bn_node->input_edge(4) != nullptr)
        {
            const auto bn_gamma_id = bn_node->input_edge(4)->producer_id();
            g.add_connection(bn_gamma_id, 0, fused_id, 6);
        }

        auto                     fused_node       = g.node(fused_id);
        std::vector<NodeIdxPair> bn_driving_nodes = get_driving_nodes(*bn_node);

        // Extract batch normalization node accessor if any
        auto bn_node_accessor = bn_node->output(0)->extract_accessor();
        auto bn_node_name     = bn_node->name();

        // Remove batch normalization node
        g.remove_node(bn_node->id());

        // Get driving nodes of batch normalization node
        for(auto &driving_node : bn_driving_nodes)
        {
            g.add_connection(fused_id, 0, driving_node.node_id, driving_node.index);
            configure_tensor(fused_node->output(0));
        }
        // Update fused node outputs
        fused_node->output(0)->set_accessor(std::move(bn_node_accessor));
        fused_node->set_assigned_target(assigned_target);
        fused_node->set_common_node_parameters(NodeParams{ conv_node->name() + "+" + bn_node_name, assigned_target });

        // Remove convolution node
        g.remove_node(conv_node->id());
    }
    else
    {
        ARM_COMPUTE_LOG_GRAPH_VERBOSE("Prevented fusion of convolution with batch normalization due to the presence of an output accessor\n");
    }
}

void fuse_depthwise_convolution_with_batch_normalization(Graph &g, const Edge *output_edge)
{
    ARM_COMPUTE_ERROR_ON(output_edge == nullptr);

    auto *depth_conv_node = arm_compute::utils::cast::polymorphic_downcast<DepthwiseConvolutionLayerNode *>(output_edge->producer());
    auto *bn_node         = arm_compute::utils::cast::polymorphic_downcast<BatchNormalizationLayerNode *>(output_edge->consumer());

    ARM_COMPUTE_LOG_GRAPH_VERBOSE("Fusing depthwise convolution node with ID : " << output_edge->producer_id()
                                  << " with BatchNormalization Layer node with ID : " << output_edge->consumer_id() << std::endl);

    // Prevent fusion if fused node has an output accessor
    if(depth_conv_node->output(0)->accessor() == nullptr)
    {
        const Target assigned_target = depth_conv_node->assigned_target();

        // Extract conv inputs
        const auto depth_conv_input_id = depth_conv_node->input_edge(0)->producer_id();
        const auto conv_weights_id     = depth_conv_node->input_edge(1)->producer_id();
        const auto conv_info           = depth_conv_node->convolution_info();
        const auto depth_conv_method   = depth_conv_node->depthwise_convolution_method();
        const auto depth_multiplier    = depth_conv_node->depth_multiplier();
        const auto act_info            = bn_node->fused_activation();

        // Extract bn inputs
        const auto bn_mean_id  = bn_node->input_edge(1)->producer_id();
        const auto bn_var_id   = bn_node->input_edge(2)->producer_id();
        const auto bn_beta_id  = bn_node->input_edge(3)->producer_id();
        const auto bn_gamma_id = bn_node->input_edge(4)->producer_id();
        const auto epsilon     = bn_node->epsilon();

        // Create the fused node
        const NodeID fused_id = g.add_node<FusedDepthwiseConvolutionBatchNormalizationNode>(epsilon, conv_info, depth_multiplier, depth_conv_method, act_info);

        if(depth_conv_node->input_edge(2) != nullptr)
        {
            const auto conv_bias_id = depth_conv_node->input_edge(2)->producer_id();
            g.add_connection(conv_bias_id, 0, fused_id, 2);
        }

        // Add connections from the conv/batch_norm inputs to the fused node
        g.add_connection(depth_conv_input_id, 0, fused_id, 0);
        g.add_connection(conv_weights_id, 0, fused_id, 1);
        g.add_connection(bn_mean_id, 0, fused_id, 3);
        g.add_connection(bn_var_id, 0, fused_id, 4);
        g.add_connection(bn_beta_id, 0, fused_id, 5);
        g.add_connection(bn_gamma_id, 0, fused_id, 6);

        auto                     fused_node       = g.node(fused_id);
        std::vector<NodeIdxPair> bn_driving_nodes = get_driving_nodes(*bn_node);

        // Extract batch normalization node accessor if any
        auto bn_node_accessor = bn_node->output(0)->extract_accessor();
        auto bn_node_name     = bn_node->name();

        // Remove batch normalization node
        g.remove_node(bn_node->id());

        // Get driving nodes of batch normalization node
        for(auto &driving_node : bn_driving_nodes)
        {
            g.add_connection(fused_id, 0, driving_node.node_id, driving_node.index);
            configure_tensor(fused_node->output(0));
        }
        // Update fused node outputs
        fused_node->output(0)->set_accessor(std::move(bn_node_accessor));
        fused_node->set_assigned_target(assigned_target);
        fused_node->set_common_node_parameters(NodeParams{ depth_conv_node->name() + "+" + bn_node_name, assigned_target });

        // Remove convolution node
        g.remove_node(depth_conv_node->id());
    }
    else
    {
        ARM_COMPUTE_LOG_GRAPH_VERBOSE("Prevented fusion of depthwise convolution with batch normalization due to the presence of an output accessor\n");
    }
}

template <typename N>
void fuse_node_with_activation(Graph &g, const Edge *output_edge, const std::set<Activation> &supported_fused_activations)
{
    ARM_COMPUTE_ERROR_ON(output_edge == nullptr);

    auto *n_node   = arm_compute::utils::cast::polymorphic_downcast<N *>(output_edge->producer());
    auto *act_node = arm_compute::utils::cast::polymorphic_downcast<ActivationLayerNode *>(output_edge->consumer());

    ARM_COMPUTE_ERROR_ON(act_node->output(0) == nullptr || n_node->output(0) == nullptr);

    // Check if activation is supported for fusion
    if(supported_fused_activations.count(act_node->activation_info().activation()) == 0)
    {
        return;
    }

    // EltwiseLayerNode can only be fused when dataype is float
    if(n_node->type() == NodeType::EltwiseLayer && !is_data_type_float(n_node->output(0)->desc().data_type))
    {
        return;
    }

    ARM_COMPUTE_LOG_GRAPH_VERBOSE("Fusing node with ID : " << output_edge->producer_id()
                                  << " with Activation Layer node with ID : " << output_edge->consumer_id() << std::endl);

    // Prevent fusion if fused node has an output accessor
    if(n_node->output(0)->accessor() == nullptr)
    {
        // Get driving nodes of activation node
        std::vector<NodeIdxPair> act_driving_nodes = get_driving_nodes(*act_node);

        // Set activation info to fused node
        n_node->set_fused_activation(act_node->activation_info());

        // Extract activation node accessor if any
        auto act_node_accessor = act_node->output(0)->extract_accessor();

        // Remove activation node
        g.remove_node(act_node->id());

        // Update fused node outputs
        for(auto &driving_node : act_driving_nodes)
        {
            g.add_connection(n_node->id(), 0, driving_node.node_id, driving_node.index);
        }

        // Update accessor to fused node
        n_node->output(0)->set_accessor(std::move(act_node_accessor));
    }
    else
    {
        ARM_COMPUTE_LOG_GRAPH_VERBOSE("Prevented fusion of node with activation due to the presence of an output accessor\n");
    }
}

bool check_padding_info(const DataLayout &layout, const PaddingList &padding_list, PaddingInfo &pad_w, PaddingInfo &pad_h)
{
    if(layout == DataLayout::NCHW || layout == DataLayout::NHWC)
    {
        const PaddingInfo zero_padding(0, 0);

        const unsigned int height_index = get_dimension_idx(layout, DataLayoutDimension::HEIGHT);
        const unsigned int width_index  = get_dimension_idx(layout, DataLayoutDimension::WIDTH);

        pad_w = width_index < padding_list.size() ? padding_list[width_index] : zero_padding;
        pad_h = height_index < padding_list.size() ? padding_list[height_index] : zero_padding;

        for(unsigned int i = 0; i < padding_list.size(); i++)
        {
            if(i != height_index && i != width_index && padding_list[i] != zero_padding)
            {
                // if the index is not either height or width, don't fuse
                return false;
            }
        }

        return true;
    }

    return false;
}

template <typename N>
void fuse_pad_with_convolution(Graph &g, const Edge *output_edge)
{
    auto *pad_node  = arm_compute::utils::cast::polymorphic_downcast<PadLayerNode *>(output_edge->producer());
    auto *conv_node = arm_compute::utils::cast::polymorphic_downcast<N *>(output_edge->consumer());

    const Edge *input_edge = pad_node->input_edge(0);
    if(input_edge != nullptr && input_edge->tensor() != nullptr && pad_node->output(0)->accessor() == nullptr
       && pad_node->pad_value().get<float>() == 0.0)
    {
        const DataLayout  layout       = input_edge->tensor()->desc().layout;
        const PaddingList padding_list = pad_node->padding();
        PaddingInfo       pad_w, pad_h;

        if(check_padding_info(layout, padding_list, pad_w, pad_h))
        {
            // Add paddings to the convolution node
            const PadStrideInfo conv_info = conv_node->convolution_info();
            const PadStrideInfo new_conv_info(
                conv_info.stride().first,
                conv_info.stride().second,
                conv_info.pad_left() + pad_w.first,
                conv_info.pad_right() + pad_w.second,
                conv_info.pad_top() + pad_h.first,
                conv_info.pad_bottom() + pad_h.second,
                conv_info.round());
            conv_node->set_convolution_info(new_conv_info);

            // Update drivers of the convolution node
            std::vector<NodeIdxPair> pad_driver_nodes = get_driver_nodes(*pad_node);
            g.remove_node(pad_node->id());

            // Update fused node inputs
            for(auto &driver_node : pad_driver_nodes)
            {
                g.add_connection(driver_node.node_id, driver_node.index, conv_node->id(), 0);
            }
        }
    }
}

template <typename N1, typename N2, typename F, typename... Args>
void fuse_layer(Graph &g, std::function<bool(INode &)> const &prec, const F fuse_fcn, Args &&... optional_arguments)
{
    // Note that fused nodes may be added to the end of the node list.
    // Instead of only looping over the original list of nodes, we loop over the current node list which could be growing.
    // This is intentional as it probes the newly added fused nodes for further fusing opportunities.
    for(unsigned int i = 0; i < g.nodes().size(); ++i)
    {
        auto node = g.node(i);
        // Check if the node is of type N and not a branching node
        if(node && node->type() == N1::node_type && node->output_edges().size() == 1)
        {
            const auto output_edge_id = *node->output_edges().begin();
            const auto output_edge    = g.edge(output_edge_id);

            // Check if following node is an activation layer node
            if((output_edge != nullptr) && (output_edge->consumer() != nullptr) && (output_edge->consumer()->type() == N2::node_type) && prec(*output_edge->producer()))
            {
                fuse_fcn(g, output_edge, optional_arguments...);
            }
        }
    }
}
} // namespace detail

const char *NodeFusionMutator::name()
{
    return "NodeFusionMutator";
}

IGraphMutator::MutationType NodeFusionMutator::type() const
{
    return IGraphMutator::MutationType::Backend;
}

void NodeFusionMutator::mutate(Graph &g)
{
    // Supported activations when fusing
    const std::set<Activation> supported_fused_activations = { Activation::ABS, Activation::BOUNDED_RELU, Activation::ELU,
                                                               Activation::HARD_SWISH, Activation::IDENTITY, Activation::LEAKY_RELU,
                                                               Activation::LINEAR, Activation::LOGISTIC, Activation::LU_BOUNDED_RELU,
                                                               Activation::RELU, Activation::SOFT_RELU, Activation::SQRT,
                                                               Activation::SQUARE, Activation::TANH
                                                             };

    // Preconditions
    auto empty_prec = [](INode &)
    {
        return true;
    };
    auto cl_target_prec = [](INode & n)
    {
        return n.assigned_target() == Target::CL;
    };
    auto qs8_prec = [&g](INode & n)
    {
        ARM_COMPUTE_ERROR_ON(n.output(0) == nullptr);

        const auto output_edge_id = *n.output_edges().begin();
        const auto output_edge    = g.edge(output_edge_id);
        // To perform fusion the two nodes must have same output quantization information
        const bool same_qinfo     = n.output(0)->desc().quant_info == output_edge->producer()->output(0)->desc().quant_info;
        const bool output_qasymm8 = n.output(0)->desc().data_type == DataType::QASYMM8;

        return (output_qasymm8 && same_qinfo) || !output_qasymm8;
    };

    // Fusion mutations
    detail::fuse_layer<PadLayerNode, ConvolutionLayerNode>(g, empty_prec, detail::fuse_pad_with_convolution<ConvolutionLayerNode>);
    detail::fuse_layer<PadLayerNode, DepthwiseConvolutionLayerNode>(g, empty_prec, detail::fuse_pad_with_convolution<DepthwiseConvolutionLayerNode>);
    detail::fuse_layer<BatchNormalizationLayerNode, ActivationLayerNode>(g, empty_prec, detail::fuse_node_with_activation<BatchNormalizationLayerNode>, supported_fused_activations);
    detail::fuse_layer<ConvolutionLayerNode, ActivationLayerNode>(g, empty_prec, detail::fuse_node_with_activation<ConvolutionLayerNode>, supported_fused_activations);
    detail::fuse_layer<DepthwiseConvolutionLayerNode, ActivationLayerNode>(g, qs8_prec, detail::fuse_node_with_activation<DepthwiseConvolutionLayerNode>, supported_fused_activations);
    detail::fuse_layer<FullyConnectedLayerNode, ActivationLayerNode>(g, empty_prec, detail::fuse_node_with_activation<FullyConnectedLayerNode>, supported_fused_activations);
    detail::fuse_layer<EltwiseLayerNode, ActivationLayerNode>(g, cl_target_prec, detail::fuse_node_with_activation<EltwiseLayerNode>, supported_fused_activations);
    detail::fuse_layer<ConvolutionLayerNode, BatchNormalizationLayerNode>(g, empty_prec, detail::fuse_convolution_with_batch_normalization);
    detail::fuse_layer<DepthwiseConvolutionLayerNode, BatchNormalizationLayerNode>(g, empty_prec, detail::fuse_depthwise_convolution_with_batch_normalization);
}
} // namespace graph
} // namespace arm_compute