openvino/inference-engine/thirdparty/clDNN/src/primitive_inst.cpp

236 lines
9.4 KiB
C++

/*
// Copyright (c) 2016-2019 Intel Corporation
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
*/
///////////////////////////////////////////////////////////////////////////////////////////////////
#include "primitive_inst.h"
#include "data_inst.h"
#include "mutable_data_inst.h"
#include "generic_layer_inst.h"
#include "input_layout_inst.h"
#include "max_unpooling_inst.h"
#include "arg_max_min_inst.h"
#include "fused_conv_eltwise_inst.h"
#include "network_impl.h"
#include "engine_impl.h"
#include "memory_impl.h"
#include "error_handler.h"
#include "json_object.h"
#include <string>
#include <vector>
#include <memory>
#include <algorithm>
namespace cldnn {
uint32_t primitive_inst::get_network_id() const { return _network.get_id(); }
void primitive_inst::check_memory_to_set(const memory_impl& mem, const layout& layout) const {
CLDNN_ERROR_LAYOUT_MISMATCH("network layout",
"set memory layout",
mem.get_layout(),
"expected layout",
layout,
"");
// check shared image/buffer compatibility, if applicable
auto params = mem.get_internal_params();
if (params.mem_type != shared_mem_type::shared_mem_empty) {
if (!mem.is_allocated_by(get_network().get_engine())) {
CLDNN_ERROR_MESSAGE(_node.id(), "Memory object is not suitable");
}
switch (params.mem_type) {
case shared_mem_type::shared_mem_vasurface:
case shared_mem_type::shared_mem_image:
if (!layout.format.is_image_2d())
CLDNN_ERROR_MESSAGE(_node.id(), "Attempt to set user-supplied input or output image instead of a buffer");
break;
case shared_mem_type::shared_mem_buffer:
case shared_mem_type::shared_mem_dxbuffer:
if (layout.format.is_image_2d())
CLDNN_ERROR_MESSAGE(_node.id(), "Attempt to set user-supplied input or output buffer instead of an image");
break;
default:
CLDNN_ERROR_MESSAGE(_node.id(), "Attempt to set user-supplied input or output memory of unknown/invalid type");
break;
}
}
}
void primitive_inst::set_output_memory(memory_impl& mem) {
auto ol = _node.get_output_layout();
check_memory_to_set(mem, ol);
_output = (memory_impl::ptr) &mem;
}
event_impl::ptr primitive_inst::execute(const std::vector<event_impl::ptr>& events) {
const auto primitive_id = id();
CLDNN_ERROR_BOOL(primitive_id,
"Invalid/unset input",
!_has_valid_input,
"Cannot execute primitive " + primitive_id + " with invalid/unset input");
on_execute();
if (_exec_deps.empty())
return _impl->execute(events, *this);
std::vector<event_impl::ptr> dependencies;
dependencies.reserve(_exec_deps.size());
for (auto& input : _exec_deps) {
auto id = input->id();
try {
// if the requested event deos not exits it means that it has not been executed, so the processing_order is
// wrong or synchronization failed.
auto ev = get_network().get_primitive_event(id);
dependencies.emplace_back(ev);
} catch (const std::out_of_range& oor) {
std::string temp = std::string("internal CLDNN error: execution order corrupted.") + std::string("\n") +
std::string(oor.what() + std::string("\n"));
CLDNN_ERROR_MESSAGE(id, temp);
}
}
return _impl->execute(dependencies, *this);
}
void primitive_inst::set_arguments() {
const auto primitive_id = id();
CLDNN_ERROR_BOOL(primitive_id,
"Invalid/unset input",
!_has_valid_input,
"Cannot set arguments for primitive " + primitive_id + " with invalid/unset input");
_impl->set_arguments(*this);
}
void primitive_inst::cleanup() {
_impl->cleanup(*this);
}
void primitive_inst::build_deps() {
if (_deps.empty() && !_node.get_dependencies().empty()) {
_deps = _network.get_primitives(_node.get_dependencies());
_exec_deps = build_exec_deps(_deps);
}
}
primitive_inst::primitive_inst(network_impl& network, program_node const& node, bool allocate_memory)
: _network(network), _node(node), _impl(node.get_selected_impl()), _output(), _output_changed(false) {
if (allocate_memory) {
// In case when output is mutable_data primitive, and other users dependencies are only used for
// suychronization, The output memory of such primitive will be fused with mutable_data
auto users = node.get_users();
auto user_count = users.size();
uint32_t mutable_data_count = 0;
for (auto& user : users) {
// Get mutable_data nodes count from nodes users
if (user->is_type<mutable_data>()) {
mutable_data_count++;
} else if (user->is_type<fused_conv_eltwise>()) {
if (!user->as<fused_conv_eltwise>().get_users().empty() &&
(*user->as<fused_conv_eltwise>().get_users().begin())->is_type<mutable_data>()) {
if (user->as<fused_conv_eltwise>().get_dependency(1).id() == node.id()) {
user_count--;
}
}
}
}
// TODO: Remove WA for arg_max_min node.
// For now it's required to handle the case when only second output of TopK primitive is used in plugin,
// but kernels always write both outputs to the same memory object which leads to wrong result.
if (user_count == 1 && mutable_data_count == 1 && !node.is_type<arg_max_min>()) {
for (auto& user : node.get_users())
if (user->is_type<mutable_data>())
_output = user->as<mutable_data>().get_attached_memory_ptr();
} else {
_output = allocate_output();
}
}
}
memory_impl::ptr primitive_inst::allocate_output() {
auto layout = _node.get_output_layout();
auto net_id = get_network_id();
auto& engine = get_network().get_engine();
// For outputs, cpu prim we want to have lockable alloc type
// Also if the successor of a node is an cpu, then memory needs to be lockable.
auto use_lockable_memory = _node.is_output() || _node.get_selected_impl()->is_cpu()
|| std::any_of(_node.get_users().begin(), _node.get_users().end(),
[](const program_node* n) {return n->get_selected_impl()->is_cpu() || n->can_be_optimized(); })
|| engine.supports_allocation(allocation_type::usm_device) == false;
allocation_type alloc_type = use_lockable_memory ?
engine.get_lockable_preffered_memory_allocation_type(layout.format.is_image_2d())
: allocation_type::usm_device;
if (!_network.is_internal() && (_node.can_be_optimized() || _node.is_type<generic_layer>())) {
return engine.allocate_memory(layout,
_node.id(),
net_id,
_node.get_memory_dependencies(),
alloc_type,
false);
} else if (_network.is_internal() || (!_node.can_share_buffer()) || _node.can_be_optimized() || _node.is_output()) {
return engine.allocate_memory(layout, alloc_type, net_id);
}
return engine.allocate_memory(layout,
_node.id(),
net_id,
_node.get_memory_dependencies(),
alloc_type,
true);
}
std::vector<std::shared_ptr<primitive_inst>> primitive_inst::build_exec_deps(
std::vector<std::shared_ptr<primitive_inst>> const& deps) {
std::vector<std::shared_ptr<primitive_inst>> exec_deps;
exec_deps.reserve(deps.size());
for (auto& dep : deps)
if (dep->get_impl() != nullptr)
exec_deps.push_back(dep);
return exec_deps;
}
std::string primitive_inst::generic_to_string(program_node const& node, const char* type_name) {
auto node_info = node.desc_to_json();
std::stringstream primitive_description;
std::stringstream ss_inputs;
for (size_t i = 0; i < node.get_dependencies().size(); ++i) {
auto& in = node.get_dependency(i);
ss_inputs << in.id();
ss_inputs << ", count: " << in.get_output_layout().count();
i != (node.get_dependencies().size() - 1) ? ss_inputs << ", " : ss_inputs << "";
}
json_composite generic_info;
generic_info.add("type_name", type_name);
generic_info.add("deps count", node.get_dependencies().size());
generic_info.add("deps", ss_inputs.str());
node_info->add("generic info", generic_info);
node_info->dump(primitive_description);
return primitive_description.str();
}
} // namespace cldnn