mirror of
https://github.com/saymrwulf/onnxruntime.git
synced 2026-07-08 17:17:15 +00:00
2679711bee
Work on minimizing memory management calls by reducing number of allocations and copies. Replace std::unordered_set to InlinedHashSet and add usage of InlinedVector. Employ std::move() to minimize copying and memory allocations. Remove copying of the const shared data into each of the PropagateCast transformer instances. Move inlined_containers.h header to include/common Adjust AsSpan imlementation for C++ < 17
75 lines
No EOL
3.2 KiB
C++
75 lines
No EOL
3.2 KiB
C++
// Copyright (c) Microsoft Corporation. All rights reserved.
|
|
// Licensed under the MIT License.
|
|
|
|
#include "core/optimizer/rule_based_graph_transformer.h"
|
|
|
|
#include "gtest/gtest.h"
|
|
|
|
#include "asserts.h"
|
|
#include "core/graph/graph_viewer.h"
|
|
#include "core/graph/model.h"
|
|
#include "core/optimizer/graph_transformer.h"
|
|
#include "core/optimizer/graph_transformer_mgr.h"
|
|
#include "dummy_graph_transformer.h"
|
|
#include "test/framework/test_utils.h"
|
|
#include "test/test_environment.h"
|
|
|
|
using namespace std;
|
|
using namespace ONNX_NAMESPACE;
|
|
|
|
namespace onnxruntime {
|
|
namespace test {
|
|
|
|
TEST(RuleBasedGraphTransformerTest, TestCompatibleProviders) {
|
|
auto model_uri = ORT_TSTR("testdata/transform/fusion/fuse-conv-bn-mul-add-unsqueeze.onnx");
|
|
|
|
std::shared_ptr<Model> model;
|
|
ASSERT_TRUE(Model::Load(model_uri, model, nullptr,
|
|
DefaultLoggingManager().DefaultLogger())
|
|
.IsOK());
|
|
Graph& graph = model->MainGraph();
|
|
|
|
// Create rule based transformer with a dummy rewrite rule and register it with Cuda as compatible provider
|
|
InlinedHashSet<std::string_view> compatible_provider{onnxruntime::kCudaExecutionProvider, onnxruntime::kRocmExecutionProvider};
|
|
auto dummy_rule = std::make_unique<DummyRewriteRule>("DummyRule");
|
|
const auto* dummy_rule_ptr = dummy_rule.get();
|
|
|
|
auto graph_transformer = std::make_unique<RuleBasedGraphTransformer>("CUDATopDownTransformer", compatible_provider);
|
|
ASSERT_STATUS_OK(graph_transformer->Register(std::move(dummy_rule)));
|
|
|
|
// Create rule based transformer with a dummy rewrite rule and register it with CPU as compatible provider
|
|
auto dummy_rule1 = std::make_unique<DummyRewriteRule>("DummyRule1");
|
|
const auto* dummy_rule1_ptr = dummy_rule1.get();
|
|
|
|
auto graph_transformer1 = std::make_unique<RuleBasedGraphTransformer>("CPUTopDownTransformer");
|
|
|
|
ASSERT_STATUS_OK(graph_transformer1->Register(std::move(dummy_rule1)));
|
|
|
|
onnxruntime::GraphTransformerManager graph_transformation_mgr{5};
|
|
ASSERT_STATUS_OK(graph_transformation_mgr.Register(std::move(graph_transformer), TransformerLevel::Level2));
|
|
ASSERT_STATUS_OK(graph_transformation_mgr.Register(std::move(graph_transformer1), TransformerLevel::Level2));
|
|
|
|
ASSERT_STATUS_OK(graph_transformation_mgr.ApplyTransformers(graph, TransformerLevel::Level2,
|
|
DefaultLoggingManager().DefaultLogger()));
|
|
|
|
// Validate transformer registered with CUDA as compatible provider is not called.
|
|
ASSERT_FALSE(dummy_rule_ptr->IsRewriteRuleInvoked());
|
|
|
|
// Validate transformer registered as global is called.
|
|
ASSERT_TRUE(dummy_rule1_ptr->IsRewriteRuleInvoked());
|
|
}
|
|
|
|
TEST(RuleBasedGraphTransformerTest, TestSettingStepsInGraphTransformerManager) {
|
|
// steps provided at object construction time
|
|
onnxruntime::GraphTransformerManager graph_transformation_mgr{5};
|
|
unsigned steps_queried;
|
|
ASSERT_STATUS_OK(graph_transformation_mgr.GetSteps(steps_queried));
|
|
ASSERT_EQ(steps_queried, static_cast<unsigned>(5));
|
|
|
|
// steps updated
|
|
ASSERT_STATUS_OK(graph_transformation_mgr.SetSteps(10));
|
|
ASSERT_STATUS_OK(graph_transformation_mgr.GetSteps(steps_queried));
|
|
ASSERT_EQ(steps_queried, static_cast<unsigned> (10));
|
|
}
|
|
} // namespace test
|
|
} // namespace onnxruntime
|