TheRock/rocm-libraries/projects/hipdnn/tests/frontend/IntegrationConvolutionWgradDescriptorLowering.cpp

// Copyright © Advanced Micro Devices, Inc., or its affiliates.
// SPDX-License-Identifier:  MIT

#include <gtest/gtest.h>
#include <hip/hip_runtime.h>
#include <memory>
#include <unordered_map>
#include <unordered_set>
#include <vector>

#include <hipdnn_flatbuffers_sdk/data_objects/convolution_wrw_attributes_generated.h>
#include <hipdnn_flatbuffers_sdk/data_objects/graph_generated.h>
#include <hipdnn_frontend.hpp>
#include <hipdnn_test_sdk/constants/ConvWgradConstants.hpp>
#include <hipdnn_test_sdk/utilities/IntegrationTestFixture.hpp>
#include <hipdnn_test_sdk/utilities/LoweringTestHelpers.hpp>
#include <hipdnn_test_sdk/utilities/TestUtilities.hpp>
#include <hipdnn_test_sdk/utilities/TestableGraph.hpp>
#include <hipdnn_test_sdk/utilities/ToVec.hpp>

#include "test_plugins/TestPluginConstants.hpp"

using namespace hipdnn_frontend;
using namespace hipdnn_frontend::graph;
using namespace hipdnn_tests::constants;
using hipdnn_tests::toVec;
using DataTypeSdk = hipdnn_flatbuffers_sdk::data_objects::DataType;
using NodeAttrType = hipdnn_flatbuffers_sdk::data_objects::NodeAttributes;
using ConvModeSdk = hipdnn_flatbuffers_sdk::data_objects::ConvMode;
using hipdnn_tests::buildTensorMap;
using hipdnn_tests::IntegrationTestFixture;
using hipdnn_tests::lowerAndDeserialize;
using hipdnn_tests::TestableGraphLowering;

namespace
{

// -- Test constants for AutoAssignedUidsPreservedInRoundTrip --

constexpr std::array<int64_t, 4> K_AUTO_X_DIMS = {1, 3, 8, 8};
constexpr std::array<int64_t, 4> K_AUTO_X_STRIDES = {192, 64, 8, 1};
constexpr std::array<int64_t, 4> K_AUTO_DY_DIMS = {1, 16, 6, 6};
constexpr std::array<int64_t, 4> K_AUTO_DY_STRIDES = {576, 36, 6, 1};
constexpr std::array<int64_t, 4> K_AUTO_DW_DIMS = {16, 3, 3, 3};
constexpr std::array<int64_t, 4> K_AUTO_DW_STRIDES = {27, 9, 3, 1};

constexpr std::array<int64_t, 2> K_AUTO_PADDING = {0, 0};
constexpr std::array<int64_t, 2> K_AUTO_STRIDE = {1, 1};
constexpr std::array<int64_t, 2> K_AUTO_DILATION = {1, 1};

// Lowers a frontend graph via build_operation_graph_via_descriptors, then
// retrieves the serialized graph and deserializes it for verification.
class IntegrationConvolutionWgradDescriptorLowering : public IntegrationTestFixture
{
};

// Builds a conv_wgrad graph via the frontend API, lowers it to the backend
// via build_operation_graph_via_descriptors, retrieves the serialized graph,
// and verifies all tensor and operation attributes match the values set
// in the frontend.
TEST_F(IntegrationConvolutionWgradDescriptorLowering, ConvWgradGraphRoundTrip)
{
    auto graph = std::make_shared<TestableGraphLowering>();
    graph->set_name("TestConvWgradGraph")
        .set_io_data_type(DataType::FLOAT)
        .set_intermediate_data_type(DataType::FLOAT)
        .set_compute_data_type(DataType::FLOAT);

    auto x = std::make_shared<TensorAttributes>();
    x->set_uid(K_WGRAD_TENSOR_X_UID).set_name("X").set_data_type(DataType::FLOAT);
    x->set_dim(toVec(K_WGRAD_TENSOR_X_DIMS)).set_stride(toVec(K_WGRAD_TENSOR_X_STRIDES));

    auto dy = std::make_shared<TensorAttributes>();
    dy->set_uid(K_WGRAD_TENSOR_DY_UID).set_name("DY").set_data_type(DataType::FLOAT);
    dy->set_dim(toVec(K_WGRAD_TENSOR_DY_DIMS)).set_stride(toVec(K_WGRAD_TENSOR_DY_STRIDES));

    ConvWgradAttributes convAttrs;
    convAttrs.set_name("conv_wgrad_op");
    convAttrs.set_pre_padding(toVec(K_WGRAD_CONV_PADDING));
    convAttrs.set_post_padding(toVec(K_WGRAD_CONV_PADDING));
    convAttrs.set_stride(toVec(K_WGRAD_CONV_STRIDE));
    convAttrs.set_dilation(toVec(K_WGRAD_CONV_DILATION));
    convAttrs.set_convolution_mode(ConvolutionMode::CROSS_CORRELATION);

    auto dw = graph->conv_wgrad(dy, x, convAttrs);
    dw->set_dim(toVec(K_WGRAD_TENSOR_DW_DIMS));
    dw->set_uid(K_WGRAD_TENSOR_DW_UID).set_output(true).set_name("DW");

    auto graphT = lowerAndDeserialize(*graph, _handle);

    // -- Verify graph-level attributes --
    EXPECT_EQ(graphT.compute_data_type, DataTypeSdk::FLOAT);
    EXPECT_EQ(graphT.intermediate_data_type, DataTypeSdk::FLOAT);
    EXPECT_EQ(graphT.io_data_type, DataTypeSdk::FLOAT);

    // -- Verify tensors --
    ASSERT_EQ(graphT.tensors.size(), 3u);

    auto tensorMap = buildTensorMap(graphT);

    // Verify X tensor
    ASSERT_NE(tensorMap.count(K_WGRAD_TENSOR_X_UID), 0u);
    auto* xT = tensorMap[K_WGRAD_TENSOR_X_UID];
    EXPECT_EQ(xT->name, "X");
    EXPECT_EQ(xT->data_type, DataTypeSdk::FLOAT);
    EXPECT_EQ(xT->dims, toVec(K_WGRAD_TENSOR_X_DIMS));
    EXPECT_EQ(xT->strides, toVec(K_WGRAD_TENSOR_X_STRIDES));
    EXPECT_FALSE(xT->virtual_);

    // Verify DY tensor
    ASSERT_NE(tensorMap.count(K_WGRAD_TENSOR_DY_UID), 0u);
    auto* dyT = tensorMap[K_WGRAD_TENSOR_DY_UID];
    EXPECT_EQ(dyT->name, "DY");
    EXPECT_EQ(dyT->data_type, DataTypeSdk::FLOAT);
    EXPECT_EQ(dyT->dims, toVec(K_WGRAD_TENSOR_DY_DIMS));
    EXPECT_EQ(dyT->strides, toVec(K_WGRAD_TENSOR_DY_STRIDES));
    EXPECT_FALSE(dyT->virtual_);

    // Verify DW tensor
    ASSERT_NE(tensorMap.count(K_WGRAD_TENSOR_DW_UID), 0u);
    auto* dwT = tensorMap[K_WGRAD_TENSOR_DW_UID];
    EXPECT_EQ(dwT->name, "DW");
    EXPECT_EQ(dwT->data_type, DataTypeSdk::FLOAT);
    EXPECT_FALSE(dwT->virtual_);
    EXPECT_EQ(dwT->dims, toVec(K_WGRAD_TENSOR_DW_DIMS));
    EXPECT_EQ(dwT->strides, toVec(K_WGRAD_TENSOR_DW_STRIDES));

    // -- Verify conv wrw operation node --
    ASSERT_EQ(graphT.nodes.size(), 1u);
    auto& node = graphT.nodes[0];
    EXPECT_EQ(node->name, "conv_wgrad_op");
    EXPECT_EQ(node->compute_data_type, DataTypeSdk::FLOAT);
    EXPECT_EQ(node->attributes.type, NodeAttrType::ConvolutionWrwAttributes);

    auto* convWrw = node->attributes.AsConvolutionWrwAttributes();
    ASSERT_NE(convWrw, nullptr);

    EXPECT_EQ(convWrw->x_tensor_uid, K_WGRAD_TENSOR_X_UID);
    EXPECT_EQ(convWrw->dy_tensor_uid, K_WGRAD_TENSOR_DY_UID);
    EXPECT_EQ(convWrw->dw_tensor_uid, K_WGRAD_TENSOR_DW_UID);
    EXPECT_EQ(convWrw->pre_padding, toVec(K_WGRAD_CONV_PADDING));
    EXPECT_EQ(convWrw->post_padding, toVec(K_WGRAD_CONV_PADDING));
    EXPECT_EQ(convWrw->stride, toVec(K_WGRAD_CONV_STRIDE));
    EXPECT_EQ(convWrw->dilation, toVec(K_WGRAD_CONV_DILATION));
    EXPECT_EQ(convWrw->conv_mode, ConvModeSdk::CROSS_CORRELATION);
}

// Verifies that tensor UIDs auto-assigned by the frontend are preserved
// through the lowering round-trip.
TEST_F(IntegrationConvolutionWgradDescriptorLowering, AutoAssignedUidsPreservedInRoundTrip)
{
    auto graph = std::make_shared<TestableGraphLowering>();
    graph->set_name("AutoUidWgradGraph")
        .set_io_data_type(DataType::FLOAT)
        .set_intermediate_data_type(DataType::FLOAT)
        .set_compute_data_type(DataType::FLOAT);

    auto x = std::make_shared<TensorAttributes>();
    x->set_name("X").set_data_type(DataType::FLOAT);
    x->set_dim(toVec(K_AUTO_X_DIMS)).set_stride(toVec(K_AUTO_X_STRIDES));

    auto dy = std::make_shared<TensorAttributes>();
    dy->set_name("DY").set_data_type(DataType::FLOAT);
    dy->set_dim(toVec(K_AUTO_DY_DIMS)).set_stride(toVec(K_AUTO_DY_STRIDES));

    ConvWgradAttributes convAttrs;
    convAttrs.set_padding(toVec(K_AUTO_PADDING));
    convAttrs.set_stride(toVec(K_AUTO_STRIDE));
    convAttrs.set_dilation(toVec(K_AUTO_DILATION));

    auto dw = graph->conv_wgrad(dy, x, convAttrs);
    dw->set_dim(toVec(K_AUTO_DW_DIMS));
    dw->set_output(true);

    auto graphT = lowerAndDeserialize(*graph, _handle);

    // All tensors should have been auto-assigned unique UIDs
    ASSERT_EQ(graphT.tensors.size(), 3u);
    std::unordered_set<int64_t> uids;
    for(const auto& t : graphT.tensors)
    {
        uids.insert(t->uid);
    }
    EXPECT_EQ(uids.size(), 3u)
        << "Tensor UIDs are not unique"; // NOLINT(readability-implicit-bool-conversion)

    // The conv wrw operation should reference the auto-assigned UIDs
    ASSERT_EQ(graphT.nodes.size(), 1u);
    auto* convWrw = graphT.nodes[0]->attributes.AsConvolutionWrwAttributes();
    ASSERT_NE(convWrw, nullptr);

    // Tensor UIDs in the node should match tensors in the graph
    EXPECT_TRUE(uids.count(convWrw->x_tensor_uid) > 0)
        << "X tensor UID " << convWrw->x_tensor_uid
        << " not found in graph tensors"; // NOLINT(readability-implicit-bool-conversion)
    EXPECT_TRUE(uids.count(convWrw->dy_tensor_uid) > 0)
        << "DY tensor UID " << convWrw->dy_tensor_uid
        << " not found in graph tensors"; // NOLINT(readability-implicit-bool-conversion)
    EXPECT_TRUE(uids.count(convWrw->dw_tensor_uid) > 0)
        << "DW tensor UID " << convWrw->dw_tensor_uid
        << " not found in graph tensors"; // NOLINT(readability-implicit-bool-conversion)

    // All three tensor UIDs referenced by the node should be distinct
    const std::unordered_set<int64_t> nodeUids
        = {convWrw->x_tensor_uid, convWrw->dy_tensor_uid, convWrw->dw_tensor_uid};
    EXPECT_EQ(nodeUids.size(), 3u)
        << "Conv wrw node tensor UIDs are not distinct"; // NOLINT(readability-implicit-bool-conversion)

    auto tensorMap = buildTensorMap(graphT);
    ASSERT_NE(tensorMap.count(convWrw->dw_tensor_uid), 0u);
    EXPECT_EQ(tensorMap[convWrw->dw_tensor_uid]->dims, toVec(K_AUTO_DW_DIMS));
    EXPECT_EQ(tensorMap[convWrw->dw_tensor_uid]->strides, toVec(K_AUTO_DW_STRIDES));
}

} // namespace