Media
/
compute-runtime
mirror of https://github.com/intel/compute-runtime.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
compute-runtime/opencl/test/unit_test/command_queue/dispatch_walker_tests.cpp

1443 lines
57 KiB
C++
Raw Blame History

/*
* Copyright (C) 2017-2020 Intel Corporation
*
* SPDX-License-Identifier: MIT
*
*/
#include "shared/source/helpers/aligned_memory.h"
#include "shared/source/memory_manager/internal_allocation_storage.h"
#include "shared/source/utilities/tag_allocator.h"
#include "shared/test/unit_test/cmd_parse/hw_parse.h"
#include "shared/test/unit_test/helpers/debug_manager_state_restore.h"
#include "opencl/source/built_ins/aux_translation_builtin.h"
#include "opencl/source/command_queue/gpgpu_walker.h"
#include "opencl/source/command_queue/hardware_interface.h"
#include "opencl/source/event/perf_counter.h"
#include "opencl/source/helpers/hardware_commands_helper.h"
#include "opencl/source/helpers/task_information.h"
#include "opencl/test/unit_test/command_queue/command_queue_fixture.h"
#include "opencl/test/unit_test/fixtures/cl_device_fixture.h"
#include "opencl/test/unit_test/helpers/unit_test_helper.h"
#include "opencl/test/unit_test/mocks/mock_buffer.h"
#include "opencl/test/unit_test/mocks/mock_command_queue.h"
#include "opencl/test/unit_test/mocks/mock_graphics_allocation.h"
#include "opencl/test/unit_test/mocks/mock_kernel.h"
#include "opencl/test/unit_test/mocks/mock_mdi.h"
#include "opencl/test/unit_test/mocks/mock_program.h"
#include "test.h"
using namespace NEO;
struct DispatchWalkerTest : public CommandQueueFixture, public ClDeviceFixture, public ::testing::Test {
using CommandQueueFixture::SetUp;
void SetUp() override {
DebugManager.flags.EnableTimestampPacket.set(0);
ClDeviceFixture::SetUp();
context = std::make_unique<MockContext>(pClDevice);
CommandQueueFixture::SetUp(context.get(), pClDevice, 0);
program = std::make_unique<MockProgram>(*pDevice->getExecutionEnvironment());
memset(&kernelHeader, 0, sizeof(kernelHeader));
kernelHeader.KernelHeapSize = sizeof(kernelIsa);
memset(&dataParameterStream, 0, sizeof(dataParameterStream));
dataParameterStream.DataParameterStreamSize = sizeof(crossThreadData);
executionEnvironment = {};
memset(&executionEnvironment, 0, sizeof(executionEnvironment));
executionEnvironment.CompiledSIMD32 = 1;
executionEnvironment.LargestCompiledSIMDSize = 32;
memset(&threadPayload, 0, sizeof(threadPayload));
threadPayload.LocalIDXPresent = 1;
threadPayload.LocalIDYPresent = 1;
threadPayload.LocalIDZPresent = 1;
samplerArray.BorderColorOffset = 0;
samplerArray.Count = 1;
samplerArray.Offset = 4;
samplerArray.Size = 2;
samplerArray.Token = 0;
kernelInfo.heapInfo.pKernelHeap = kernelIsa;
kernelInfo.heapInfo.KernelHeapSize = sizeof(kernelIsa);
kernelInfo.patchInfo.dataParameterStream = &dataParameterStream;
kernelInfo.patchInfo.executionEnvironment = &executionEnvironment;
kernelInfo.patchInfo.threadPayload = &threadPayload;
kernelInfoWithSampler.heapInfo.pKernelHeap = kernelIsa;
kernelInfoWithSampler.heapInfo.KernelHeapSize = sizeof(kernelIsa);
kernelInfoWithSampler.patchInfo.dataParameterStream = &dataParameterStream;
kernelInfoWithSampler.patchInfo.executionEnvironment = &executionEnvironment;
kernelInfoWithSampler.patchInfo.threadPayload = &threadPayload;
kernelInfoWithSampler.patchInfo.samplerStateArray = &samplerArray;
kernelInfoWithSampler.heapInfo.pDsh = static_cast<const void *>(dsh);
}
void TearDown() override {
CommandQueueFixture::TearDown();
context.reset();
ClDeviceFixture::TearDown();
}
std::unique_ptr<KernelOperation> createBlockedCommandsData(CommandQueue &commandQueue) {
auto commandStream = new LinearStream();
auto &gpgpuCsr = commandQueue.getGpgpuCommandStreamReceiver();
gpgpuCsr.ensureCommandBufferAllocation(*commandStream, 1, 1);
return std::make_unique<KernelOperation>(commandStream, *gpgpuCsr.getInternalAllocationStorage());
}
std::unique_ptr<MockContext> context;
std::unique_ptr<MockProgram> program;
SKernelBinaryHeaderCommon kernelHeader = {};
SPatchDataParameterStream dataParameterStream = {};
SPatchExecutionEnvironment executionEnvironment = {};
SPatchThreadPayload threadPayload = {};
SPatchSamplerStateArray samplerArray = {};
KernelInfo kernelInfo;
KernelInfo kernelInfoWithSampler;
uint32_t kernelIsa[32];
uint32_t crossThreadData[32];
uint32_t dsh[32];
DebugManagerStateRestore dbgRestore;
};
HWTEST_F(DispatchWalkerTest, WhenGettingComputeDimensionsThenCorrectNumberOfDimensionsIsReturned) {
const size_t workItems1D[] = {100, 1, 1};
EXPECT_EQ(1u, computeDimensions(workItems1D));
const size_t workItems2D[] = {100, 100, 1};
EXPECT_EQ(2u, computeDimensions(workItems2D));
const size_t workItems3D[] = {100, 100, 100};
EXPECT_EQ(3u, computeDimensions(workItems3D));
}
HWTEST_F(DispatchWalkerTest, givenSimd1WhenSetGpgpuWalkerThreadDataThenSimdInWalkerIsSetTo32Value) {
uint32_t pCmdBuffer[1024];
MockGraphicsAllocation gfxAllocation((void *)pCmdBuffer, sizeof(pCmdBuffer));
LinearStream linearStream(&gfxAllocation);
using WALKER_TYPE = typename FamilyType::WALKER_TYPE;
WALKER_TYPE *computeWalker = static_cast<WALKER_TYPE *>(linearStream.getSpace(sizeof(WALKER_TYPE)));
*computeWalker = FamilyType::cmdInitGpgpuWalker;
size_t globalOffsets[] = {0, 0, 0};
size_t startWorkGroups[] = {0, 0, 0};
size_t numWorkGroups[] = {1, 1, 1};
size_t localWorkSizesIn[] = {32, 1, 1};
uint32_t simd = 1;
iOpenCL::SPatchThreadPayload threadPayload;
GpgpuWalkerHelper<FamilyType>::setGpgpuWalkerThreadData(
computeWalker, globalOffsets, startWorkGroups, numWorkGroups, localWorkSizesIn, simd, 3, true, false, threadPayload, 5u);
EXPECT_EQ(computeWalker->getSimdSize(), 32 >> 4);
}
HWTEST_F(DispatchWalkerTest, WhenDispatchingWalkerThenCommandStreamMemoryIsntChanged) {
MockKernel kernel(program.get(), kernelInfo, *pClDevice);
ASSERT_EQ(CL_SUCCESS, kernel.initialize());
auto &commandStream = pCmdQ->getCS(4096);
// Consume all memory except what is needed for this enqueue
auto sizeDispatchWalkerNeeds = sizeof(typename FamilyType::WALKER_TYPE) +
HardwareCommandsHelper<FamilyType>::getSizeRequiredCS(&kernel);
//cs has a minimum required size
auto sizeThatNeedsToBeSubstracted = sizeDispatchWalkerNeeds + CSRequirements::minCommandQueueCommandStreamSize;
commandStream.getSpace(commandStream.getMaxAvailableSpace() - sizeThatNeedsToBeSubstracted);
ASSERT_EQ(commandStream.getAvailableSpace(), sizeThatNeedsToBeSubstracted);
auto commandStreamStart = commandStream.getUsed();
auto commandStreamBuffer = commandStream.getCpuBase();
ASSERT_NE(0u, commandStreamStart);
size_t globalOffsets[3] = {0, 0, 0};
size_t workItems[3] = {1, 1, 1};
cl_uint dimensions = 1;
DispatchInfo dispatchInfo(const_cast<MockKernel *>(&kernel), dimensions, workItems, nullptr, globalOffsets);
MultiDispatchInfo multiDispatchInfo;
multiDispatchInfo.push(dispatchInfo);
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
multiDispatchInfo,
CsrDependencies(),
nullptr,
nullptr,
nullptr,
nullptr,
nullptr,
CL_COMMAND_NDRANGE_KERNEL);
EXPECT_EQ(commandStreamBuffer, commandStream.getCpuBase());
EXPECT_LT(commandStreamStart, commandStream.getUsed());
EXPECT_EQ(sizeDispatchWalkerNeeds, commandStream.getUsed() - commandStreamStart);
}
HWTEST_F(DispatchWalkerTest, GivenNoLocalIdsWhenDispatchingWalkerThenWalkerIsDispatched) {
threadPayload.LocalIDXPresent = 0;
threadPayload.LocalIDYPresent = 0;
threadPayload.LocalIDZPresent = 0;
threadPayload.UnusedPerThreadConstantPresent = 1;
MockKernel kernel(program.get(), kernelInfo, *pClDevice);
ASSERT_EQ(CL_SUCCESS, kernel.initialize());
auto &commandStream = pCmdQ->getCS(4096);
// Consume all memory except what is needed for this enqueue
auto sizeDispatchWalkerNeeds = sizeof(typename FamilyType::WALKER_TYPE) +
HardwareCommandsHelper<FamilyType>::getSizeRequiredCS(&kernel);
//cs has a minimum required size
auto sizeThatNeedsToBeSubstracted = sizeDispatchWalkerNeeds + CSRequirements::minCommandQueueCommandStreamSize;
commandStream.getSpace(commandStream.getMaxAvailableSpace() - sizeThatNeedsToBeSubstracted);
ASSERT_EQ(commandStream.getAvailableSpace(), sizeThatNeedsToBeSubstracted);
auto commandStreamStart = commandStream.getUsed();
auto commandStreamBuffer = commandStream.getCpuBase();
ASSERT_NE(0u, commandStreamStart);
size_t globalOffsets[3] = {0, 0, 0};
size_t workItems[3] = {1, 1, 1};
cl_uint dimensions = 1;
DispatchInfo dispatchInfo(const_cast<MockKernel *>(&kernel), dimensions, workItems, nullptr, globalOffsets);
MultiDispatchInfo multiDispatchInfo;
multiDispatchInfo.push(dispatchInfo);
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
multiDispatchInfo,
CsrDependencies(),
nullptr,
nullptr,
nullptr,
nullptr,
nullptr,
CL_COMMAND_NDRANGE_KERNEL);
EXPECT_EQ(commandStreamBuffer, commandStream.getCpuBase());
EXPECT_LT(commandStreamStart, commandStream.getUsed());
EXPECT_EQ(sizeDispatchWalkerNeeds, commandStream.getUsed() - commandStreamStart);
}
HWTEST_F(DispatchWalkerTest, GivenDefaultLwsAlgorithmWhenDispatchingWalkerThenDimensionsAreCorrect) {
MockKernel kernel(program.get(), kernelInfo, *pClDevice);
kernelInfo.workloadInfo.workDimOffset = 0;
ASSERT_EQ(CL_SUCCESS, kernel.initialize());
size_t globalOffsets[3] = {0, 0, 0};
size_t workItems[3] = {1, 1, 1};
for (uint32_t dimension = 1; dimension <= 3; ++dimension) {
workItems[dimension - 1] = 256;
DispatchInfo dispatchInfo(const_cast<MockKernel *>(&kernel), dimension, workItems, nullptr, globalOffsets);
MultiDispatchInfo multiDispatchInfo;
multiDispatchInfo.push(dispatchInfo);
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
multiDispatchInfo,
CsrDependencies(),
nullptr,
nullptr,
nullptr,
nullptr,
nullptr,
CL_COMMAND_NDRANGE_KERNEL);
EXPECT_EQ(dimension, *kernel.workDim);
}
}
HWTEST_F(DispatchWalkerTest, GivenSquaredLwsAlgorithmWhenDispatchingWalkerThenDimensionsAreCorrect) {
DebugManagerStateRestore dbgRestore;
DebugManager.flags.EnableComputeWorkSizeND.set(false);
DebugManager.flags.EnableComputeWorkSizeSquared.set(true);
MockKernel kernel(program.get(), kernelInfo, *pClDevice);
kernelInfo.workloadInfo.workDimOffset = 0;
ASSERT_EQ(CL_SUCCESS, kernel.initialize());
size_t globalOffsets[3] = {0, 0, 0};
size_t workItems[3] = {1, 1, 1};
for (uint32_t dimension = 1; dimension <= 3; ++dimension) {
workItems[dimension - 1] = 256;
DispatchInfo dispatchInfo(const_cast<MockKernel *>(&kernel), dimension, workItems, nullptr, globalOffsets);
MultiDispatchInfo multiDispatchInfo;
multiDispatchInfo.push(dispatchInfo);
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
multiDispatchInfo,
CsrDependencies(),
nullptr,
nullptr,
nullptr,
nullptr,
nullptr,
CL_COMMAND_NDRANGE_KERNEL);
EXPECT_EQ(dimension, *kernel.workDim);
}
}
HWTEST_F(DispatchWalkerTest, GivenNdLwsAlgorithmWhenDispatchingWalkerThenDimensionsAreCorrect) {
DebugManagerStateRestore dbgRestore;
DebugManager.flags.EnableComputeWorkSizeND.set(true);
MockKernel kernel(program.get(), kernelInfo, *pClDevice);
kernelInfo.workloadInfo.workDimOffset = 0;
ASSERT_EQ(CL_SUCCESS, kernel.initialize());
size_t globalOffsets[3] = {0, 0, 0};
size_t workItems[3] = {1, 1, 1};
for (uint32_t dimension = 1; dimension <= 3; ++dimension) {
workItems[dimension - 1] = 256;
DispatchInfo dispatchInfo(const_cast<MockKernel *>(&kernel), dimension, workItems, nullptr, globalOffsets);
MultiDispatchInfo multiDispatchInfo;
multiDispatchInfo.push(dispatchInfo);
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
multiDispatchInfo,
CsrDependencies(),
nullptr,
nullptr,
nullptr,
nullptr,
nullptr,
CL_COMMAND_NDRANGE_KERNEL);
EXPECT_EQ(dimension, *kernel.workDim);
}
}
HWTEST_F(DispatchWalkerTest, GivenOldLwsAlgorithmWhenDispatchingWalkerThenDimensionsAreCorrect) {
DebugManagerStateRestore dbgRestore;
DebugManager.flags.EnableComputeWorkSizeND.set(false);
DebugManager.flags.EnableComputeWorkSizeSquared.set(false);
MockKernel kernel(program.get(), kernelInfo, *pClDevice);
kernelInfo.workloadInfo.workDimOffset = 0;
ASSERT_EQ(CL_SUCCESS, kernel.initialize());
size_t globalOffsets[3] = {0, 0, 0};
size_t workItems[3] = {1, 1, 1};
for (uint32_t dimension = 1; dimension <= 3; ++dimension) {
workItems[dimension - 1] = 256;
DispatchInfo dispatchInfo(const_cast<MockKernel *>(&kernel), dimension, workItems, nullptr, globalOffsets);
MultiDispatchInfo multiDispatchInfo;
multiDispatchInfo.push(dispatchInfo);
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
multiDispatchInfo,
CsrDependencies(),
nullptr,
nullptr,
nullptr,
nullptr,
nullptr,
CL_COMMAND_NDRANGE_KERNEL);
EXPECT_EQ(dimension, *kernel.workDim);
}
}
HWTEST_F(DispatchWalkerTest, GivenNumWorkGroupsWhenDispatchingWalkerThenNumWorkGroupsIsCorrectlySet) {
MockKernel kernel(program.get(), kernelInfo, *pClDevice);
kernelInfo.workloadInfo.numWorkGroupsOffset[0] = 0;
kernelInfo.workloadInfo.numWorkGroupsOffset[1] = 4;
kernelInfo.workloadInfo.numWorkGroupsOffset[2] = 8;
ASSERT_EQ(CL_SUCCESS, kernel.initialize());
size_t globalOffsets[3] = {0, 0, 0};
size_t workItems[3] = {2, 5, 10};
size_t workGroupSize[3] = {1, 1, 1};
cl_uint dimensions = 3;
DispatchInfo dispatchInfo(const_cast<MockKernel *>(&kernel), dimensions, workItems, workGroupSize, globalOffsets);
MultiDispatchInfo multiDispatchInfo;
multiDispatchInfo.push(dispatchInfo);
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
multiDispatchInfo,
CsrDependencies(),
nullptr,
nullptr,
nullptr,
nullptr,
nullptr,
CL_COMMAND_NDRANGE_KERNEL);
EXPECT_EQ(2u, *kernel.numWorkGroupsX);
EXPECT_EQ(5u, *kernel.numWorkGroupsY);
EXPECT_EQ(10u, *kernel.numWorkGroupsZ);
}
HWTEST_F(DispatchWalkerTest, GivenNoLocalWorkSizeAndDefaultAlgorithmWhenDispatchingWalkerThenLwsIsCorrect) {
DebugManagerStateRestore dbgRestore;
DebugManager.flags.EnableComputeWorkSizeND.set(false);
MockKernel kernel(program.get(), kernelInfo, *pClDevice);
kernelInfo.workloadInfo.localWorkSizeOffsets[0] = 0;
kernelInfo.workloadInfo.localWorkSizeOffsets[1] = 4;
kernelInfo.workloadInfo.localWorkSizeOffsets[2] = 8;
ASSERT_EQ(CL_SUCCESS, kernel.initialize());
size_t globalOffsets[3] = {0, 0, 0};
size_t workItems[3] = {2, 5, 10};
cl_uint dimensions = 3;
DispatchInfo dispatchInfo(const_cast<MockKernel *>(&kernel), dimensions, workItems, nullptr, globalOffsets);
MultiDispatchInfo multiDispatchInfo;
multiDispatchInfo.push(dispatchInfo);
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
multiDispatchInfo,
CsrDependencies(),
nullptr,
nullptr,
nullptr,
nullptr,
nullptr,
CL_COMMAND_NDRANGE_KERNEL);
EXPECT_EQ(2u, *kernel.localWorkSizeX);
EXPECT_EQ(5u, *kernel.localWorkSizeY);
EXPECT_EQ(1u, *kernel.localWorkSizeZ);
}
HWTEST_F(DispatchWalkerTest, GivenNoLocalWorkSizeAndNdOnWhenDispatchingWalkerThenLwsIsCorrect) {
DebugManagerStateRestore dbgRestore;
DebugManager.flags.EnableComputeWorkSizeND.set(true);
MockKernel kernel(program.get(), kernelInfo, *pClDevice);
kernelInfo.workloadInfo.localWorkSizeOffsets[0] = 0;
kernelInfo.workloadInfo.localWorkSizeOffsets[1] = 4;
kernelInfo.workloadInfo.localWorkSizeOffsets[2] = 8;
ASSERT_EQ(CL_SUCCESS, kernel.initialize());
size_t globalOffsets[3] = {0, 0, 0};
size_t workItems[3] = {2, 5, 10};
cl_uint dimensions = 3;
DispatchInfo dispatchInfo(const_cast<MockKernel *>(&kernel), dimensions, workItems, nullptr, globalOffsets);
MultiDispatchInfo multiDispatchInfo;
multiDispatchInfo.push(dispatchInfo);
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
multiDispatchInfo,
CsrDependencies(),
nullptr,
nullptr,
nullptr,
nullptr,
nullptr,
CL_COMMAND_NDRANGE_KERNEL);
EXPECT_EQ(2u, *kernel.localWorkSizeX);
EXPECT_EQ(5u, *kernel.localWorkSizeY);
EXPECT_EQ(10u, *kernel.localWorkSizeZ);
}
HWTEST_F(DispatchWalkerTest, GivenNoLocalWorkSizeAndSquaredAlgorithmWhenDispatchingWalkerThenLwsIsCorrect) {
DebugManagerStateRestore dbgRestore;
DebugManager.flags.EnableComputeWorkSizeSquared.set(true);
DebugManager.flags.EnableComputeWorkSizeND.set(false);
MockKernel kernel(program.get(), kernelInfo, *pClDevice);
kernelInfo.workloadInfo.localWorkSizeOffsets[0] = 0;
kernelInfo.workloadInfo.localWorkSizeOffsets[1] = 4;
kernelInfo.workloadInfo.localWorkSizeOffsets[2] = 8;
ASSERT_EQ(CL_SUCCESS, kernel.initialize());
size_t globalOffsets[3] = {0, 0, 0};
size_t workItems[3] = {2, 5, 10};
cl_uint dimensions = 3;
DispatchInfo dispatchInfo(const_cast<MockKernel *>(&kernel), dimensions, workItems, nullptr, globalOffsets);
MultiDispatchInfo multiDispatchInfo;
multiDispatchInfo.push(dispatchInfo);
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
multiDispatchInfo,
CsrDependencies(),
nullptr,
nullptr,
nullptr,
nullptr,
nullptr,
CL_COMMAND_NDRANGE_KERNEL);
EXPECT_EQ(2u, *kernel.localWorkSizeX);
EXPECT_EQ(5u, *kernel.localWorkSizeY);
EXPECT_EQ(1u, *kernel.localWorkSizeZ);
}
HWTEST_F(DispatchWalkerTest, GivenNoLocalWorkSizeAndSquaredAlgorithmOffAndNdOffWhenDispatchingWalkerThenLwsIsCorrect) {
DebugManagerStateRestore dbgRestore;
DebugManager.flags.EnableComputeWorkSizeSquared.set(false);
DebugManager.flags.EnableComputeWorkSizeND.set(false);
MockKernel kernel(program.get(), kernelInfo, *pClDevice);
kernelInfo.workloadInfo.localWorkSizeOffsets[0] = 0;
kernelInfo.workloadInfo.localWorkSizeOffsets[1] = 4;
kernelInfo.workloadInfo.localWorkSizeOffsets[2] = 8;
ASSERT_EQ(CL_SUCCESS, kernel.initialize());
size_t globalOffsets[3] = {0, 0, 0};
size_t workItems[3] = {2, 5, 10};
cl_uint dimensions = 3;
DispatchInfo dispatchInfo(const_cast<MockKernel *>(&kernel), dimensions, workItems, nullptr, globalOffsets);
MultiDispatchInfo multiDispatchInfo;
multiDispatchInfo.push(dispatchInfo);
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
multiDispatchInfo,
CsrDependencies(),
nullptr,
nullptr,
nullptr,
nullptr,
nullptr,
CL_COMMAND_NDRANGE_KERNEL);
EXPECT_EQ(2u, *kernel.localWorkSizeX);
EXPECT_EQ(5u, *kernel.localWorkSizeY);
EXPECT_EQ(1u, *kernel.localWorkSizeZ);
}
HWTEST_F(DispatchWalkerTest, GivenNoLocalWorkSizeWhenDispatchingWalkerThenLwsIsCorrect) {
MockKernel kernel(program.get(), kernelInfo, *pClDevice);
kernelInfo.workloadInfo.localWorkSizeOffsets[0] = 0;
kernelInfo.workloadInfo.localWorkSizeOffsets[1] = 4;
kernelInfo.workloadInfo.localWorkSizeOffsets[2] = 8;
ASSERT_EQ(CL_SUCCESS, kernel.initialize());
size_t globalOffsets[3] = {0, 0, 0};
size_t workItems[3] = {2, 5, 10};
size_t workGroupSize[3] = {1, 2, 3};
cl_uint dimensions = 3;
DispatchInfo dispatchInfo(const_cast<MockKernel *>(&kernel), dimensions, workItems, workGroupSize, globalOffsets);
MultiDispatchInfo multiDispatchInfo;
multiDispatchInfo.push(dispatchInfo);
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
multiDispatchInfo,
CsrDependencies(),
nullptr,
nullptr,
nullptr,
nullptr,
nullptr,
CL_COMMAND_NDRANGE_KERNEL);
EXPECT_EQ(1u, *kernel.localWorkSizeX);
EXPECT_EQ(2u, *kernel.localWorkSizeY);
EXPECT_EQ(3u, *kernel.localWorkSizeZ);
}
HWTEST_F(DispatchWalkerTest, GivenTwoSetsOfLwsOffsetsWhenDispatchingWalkerThenLwsIsCorrect) {
MockKernel kernel(program.get(), kernelInfo, *pClDevice);
kernelInfo.workloadInfo.localWorkSizeOffsets[0] = 0;
kernelInfo.workloadInfo.localWorkSizeOffsets[1] = 4;
kernelInfo.workloadInfo.localWorkSizeOffsets[2] = 8;
kernelInfo.workloadInfo.localWorkSizeOffsets2[0] = 12;
kernelInfo.workloadInfo.localWorkSizeOffsets2[1] = 16;
kernelInfo.workloadInfo.localWorkSizeOffsets2[2] = 20;
ASSERT_EQ(CL_SUCCESS, kernel.initialize());
size_t globalOffsets[3] = {0, 0, 0};
size_t workItems[3] = {2, 5, 10};
size_t workGroupSize[3] = {1, 2, 3};
cl_uint dimensions = 3;
DispatchInfo dispatchInfo(const_cast<MockKernel *>(&kernel), dimensions, workItems, workGroupSize, globalOffsets);
MultiDispatchInfo multiDispatchInfo;
multiDispatchInfo.push(dispatchInfo);
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
multiDispatchInfo,
CsrDependencies(),
nullptr,
nullptr,
nullptr,
nullptr,
nullptr,
CL_COMMAND_NDRANGE_KERNEL);
EXPECT_EQ(1u, *kernel.localWorkSizeX);
EXPECT_EQ(2u, *kernel.localWorkSizeY);
EXPECT_EQ(3u, *kernel.localWorkSizeZ);
EXPECT_EQ(1u, *kernel.localWorkSizeX2);
EXPECT_EQ(2u, *kernel.localWorkSizeY2);
EXPECT_EQ(3u, *kernel.localWorkSizeZ2);
}
HWTEST_F(DispatchWalkerTest, GivenSplitKernelWhenDispatchingWalkerThenLwsIsCorrect) {
MockKernel kernel1(program.get(), kernelInfo, *pClDevice);
kernelInfo.workloadInfo.localWorkSizeOffsets[0] = 0;
kernelInfo.workloadInfo.localWorkSizeOffsets[1] = 4;
kernelInfo.workloadInfo.localWorkSizeOffsets[2] = 8;
ASSERT_EQ(CL_SUCCESS, kernel1.initialize());
MockKernel kernel2(program.get(), kernelInfoWithSampler, *pClDevice);
kernelInfoWithSampler.workloadInfo.localWorkSizeOffsets[0] = 12;
kernelInfoWithSampler.workloadInfo.localWorkSizeOffsets[1] = 16;
kernelInfoWithSampler.workloadInfo.localWorkSizeOffsets[2] = 20;
ASSERT_EQ(CL_SUCCESS, kernel2.initialize());
DispatchInfo di1(&kernel1, 3, {10, 10, 10}, {1, 2, 3}, {0, 0, 0});
DispatchInfo di2(&kernel2, 3, {10, 10, 10}, {4, 5, 6}, {0, 0, 0});
MockMultiDispatchInfo multiDispatchInfo(std::vector<DispatchInfo *>({&di1, &di2}));
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
multiDispatchInfo,
CsrDependencies(),
nullptr,
nullptr,
nullptr,
nullptr,
nullptr,
CL_COMMAND_NDRANGE_KERNEL);
auto dispatchId = 0;
for (auto &dispatchInfo : multiDispatchInfo) {
auto &kernel = *dispatchInfo.getKernel();
if (dispatchId == 0) {
EXPECT_EQ(1u, *kernel.localWorkSizeX);
EXPECT_EQ(2u, *kernel.localWorkSizeY);
EXPECT_EQ(3u, *kernel.localWorkSizeZ);
}
if (dispatchId == 1) {
EXPECT_EQ(4u, *kernel.localWorkSizeX);
EXPECT_EQ(5u, *kernel.localWorkSizeY);
EXPECT_EQ(6u, *kernel.localWorkSizeZ);
}
dispatchId++;
}
}
HWTEST_F(DispatchWalkerTest, GivenSplitWalkerWhenDispatchingWalkerThenLwsIsCorrect) {
MockKernel kernel1(program.get(), kernelInfo, *pClDevice);
MockKernel mainKernel(program.get(), kernelInfo, *pClDevice);
kernelInfo.workloadInfo.localWorkSizeOffsets[0] = 0;
kernelInfo.workloadInfo.localWorkSizeOffsets[1] = 4;
kernelInfo.workloadInfo.localWorkSizeOffsets[2] = 8;
kernelInfo.workloadInfo.localWorkSizeOffsets2[0] = 12;
kernelInfo.workloadInfo.localWorkSizeOffsets2[1] = 16;
kernelInfo.workloadInfo.localWorkSizeOffsets2[2] = 20;
kernelInfo.workloadInfo.numWorkGroupsOffset[0] = 24;
kernelInfo.workloadInfo.numWorkGroupsOffset[1] = 28;
kernelInfo.workloadInfo.numWorkGroupsOffset[2] = 32;
ASSERT_EQ(CL_SUCCESS, kernel1.initialize());
ASSERT_EQ(CL_SUCCESS, mainKernel.initialize());
DispatchInfo di1(&kernel1, 3, {10, 10, 10}, {1, 2, 3}, {0, 0, 0});
DispatchInfo di2(&mainKernel, 3, {10, 10, 10}, {4, 5, 6}, {0, 0, 0});
MultiDispatchInfo multiDispatchInfo(&mainKernel);
multiDispatchInfo.push(di1);
multiDispatchInfo.push(di2);
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
multiDispatchInfo,
CsrDependencies(),
nullptr,
nullptr,
nullptr,
nullptr,
nullptr,
CL_COMMAND_NDRANGE_KERNEL);
for (auto &dispatchInfo : multiDispatchInfo) {
auto &kernel = *dispatchInfo.getKernel();
if (&kernel == &mainKernel) {
EXPECT_EQ(4u, *kernel.localWorkSizeX);
EXPECT_EQ(5u, *kernel.localWorkSizeY);
EXPECT_EQ(6u, *kernel.localWorkSizeZ);
EXPECT_EQ(4u, *kernel.localWorkSizeX2);
EXPECT_EQ(5u, *kernel.localWorkSizeY2);
EXPECT_EQ(6u, *kernel.localWorkSizeZ2);
EXPECT_EQ(3u, *kernel.numWorkGroupsX);
EXPECT_EQ(2u, *kernel.numWorkGroupsY);
EXPECT_EQ(2u, *kernel.numWorkGroupsZ);
} else {
EXPECT_EQ(0u, *kernel.localWorkSizeX);
EXPECT_EQ(0u, *kernel.localWorkSizeY);
EXPECT_EQ(0u, *kernel.localWorkSizeZ);
EXPECT_EQ(1u, *kernel.localWorkSizeX2);
EXPECT_EQ(2u, *kernel.localWorkSizeY2);
EXPECT_EQ(3u, *kernel.localWorkSizeZ2);
EXPECT_EQ(0u, *kernel.numWorkGroupsX);
EXPECT_EQ(0u, *kernel.numWorkGroupsY);
EXPECT_EQ(0u, *kernel.numWorkGroupsZ);
}
}
}
HWTEST_F(DispatchWalkerTest, GivenBlockedQueueWhenDispatchingWalkerThenCommandSteamIsNotConsumed) {
MockKernel kernel(program.get(), kernelInfo, *pClDevice);
ASSERT_EQ(CL_SUCCESS, kernel.initialize());
size_t globalOffsets[3] = {0, 0, 0};
size_t workItems[3] = {1, 1, 1};
size_t workGroupSize[3] = {2, 5, 10};
cl_uint dimensions = 1;
auto blockedCommandsData = createBlockedCommandsData(*pCmdQ);
DispatchInfo dispatchInfo(const_cast<MockKernel *>(&kernel), dimensions, workItems, workGroupSize, globalOffsets);
MultiDispatchInfo multiDispatchInfo;
multiDispatchInfo.push(dispatchInfo);
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
multiDispatchInfo,
CsrDependencies(),
blockedCommandsData.get(),
nullptr,
nullptr,
nullptr,
nullptr,
CL_COMMAND_NDRANGE_KERNEL);
auto &commandStream = pCmdQ->getCS(1024);
EXPECT_EQ(0u, commandStream.getUsed());
EXPECT_NE(nullptr, blockedCommandsData);
EXPECT_NE(nullptr, blockedCommandsData->commandStream);
EXPECT_NE(nullptr, blockedCommandsData->dsh);
EXPECT_NE(nullptr, blockedCommandsData->ioh);
EXPECT_NE(nullptr, blockedCommandsData->ssh);
}
HWTEST_F(DispatchWalkerTest, GivenBlockedQueueWhenDispatchingWalkerThenRequiredHeaSizesAreTakenFromKernel) {
MockKernel kernel(program.get(), kernelInfo, *pClDevice);
ASSERT_EQ(CL_SUCCESS, kernel.initialize());
size_t globalOffsets[3] = {0, 0, 0};
size_t workItems[3] = {1, 1, 1};
size_t workGroupSize[3] = {2, 5, 10};
cl_uint dimensions = 1;
auto blockedCommandsData = createBlockedCommandsData(*pCmdQ);
DispatchInfo dispatchInfo(const_cast<MockKernel *>(&kernel), dimensions, workItems, workGroupSize, globalOffsets);
MultiDispatchInfo multiDispatchInfo(&kernel);
multiDispatchInfo.push(dispatchInfo);
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
multiDispatchInfo,
CsrDependencies(),
blockedCommandsData.get(),
nullptr,
nullptr,
nullptr,
nullptr,
CL_COMMAND_NDRANGE_KERNEL);
Vec3<size_t> localWorkgroupSize(workGroupSize);
auto expectedSizeDSH = HardwareCommandsHelper<FamilyType>::getSizeRequiredDSH(kernel);
auto expectedSizeIOH = HardwareCommandsHelper<FamilyType>::getSizeRequiredIOH(kernel, Math::computeTotalElementsCount(localWorkgroupSize));
auto expectedSizeSSH = HardwareCommandsHelper<FamilyType>::getSizeRequiredSSH(kernel);
EXPECT_LE(expectedSizeDSH, blockedCommandsData->dsh->getMaxAvailableSpace());
EXPECT_LE(expectedSizeIOH, blockedCommandsData->ioh->getMaxAvailableSpace());
EXPECT_LE(expectedSizeSSH, blockedCommandsData->ssh->getMaxAvailableSpace());
}
HWTEST_F(DispatchWalkerTest, givenBlockedEnqueueWhenObtainingCommandStreamThenAllocateEnoughSpaceAndBlockedKernelData) {
DispatchInfo dispatchInfo;
MultiDispatchInfo multiDispatchInfo;
multiDispatchInfo.push(dispatchInfo);
std::unique_ptr<KernelOperation> blockedKernelData;
MockCommandQueueHw<FamilyType> mockCmdQ(nullptr, pClDevice, nullptr);
auto expectedSizeCSAllocation = MemoryConstants::pageSize64k;
auto expectedSizeCS = MemoryConstants::pageSize64k - CSRequirements::csOverfetchSize;
CsrDependencies csrDependencies;
EventsRequest eventsRequest(0, nullptr, nullptr);
auto cmdStream = mockCmdQ.template obtainCommandStream<CL_COMMAND_NDRANGE_KERNEL>(csrDependencies, false, true,
multiDispatchInfo, eventsRequest, blockedKernelData,
nullptr, 0u);
EXPECT_EQ(expectedSizeCS, cmdStream->getMaxAvailableSpace());
EXPECT_EQ(expectedSizeCSAllocation, cmdStream->getGraphicsAllocation()->getUnderlyingBufferSize());
EXPECT_NE(nullptr, blockedKernelData);
EXPECT_EQ(cmdStream, blockedKernelData->commandStream.get());
}
HWTEST_F(DispatchWalkerTest, GivenBlockedQueueWhenDispatchingWalkerThenRequiredHeapSizesAreTakenFromMdi) {
MockKernel kernel(program.get(), kernelInfo, *pClDevice);
ASSERT_EQ(CL_SUCCESS, kernel.initialize());
MockMultiDispatchInfo multiDispatchInfo(&kernel);
auto blockedCommandsData = createBlockedCommandsData(*pCmdQ);
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
multiDispatchInfo,
CsrDependencies(),
blockedCommandsData.get(),
nullptr,
nullptr,
nullptr,
nullptr,
CL_COMMAND_NDRANGE_KERNEL);
auto expectedSizeDSH = HardwareCommandsHelper<FamilyType>::getTotalSizeRequiredDSH(multiDispatchInfo);
auto expectedSizeIOH = HardwareCommandsHelper<FamilyType>::getTotalSizeRequiredIOH(multiDispatchInfo);
auto expectedSizeSSH = HardwareCommandsHelper<FamilyType>::getTotalSizeRequiredSSH(multiDispatchInfo);
EXPECT_LE(expectedSizeDSH, blockedCommandsData->dsh->getMaxAvailableSpace());
EXPECT_LE(expectedSizeIOH, blockedCommandsData->ioh->getMaxAvailableSpace());
EXPECT_LE(expectedSizeSSH, blockedCommandsData->ssh->getMaxAvailableSpace());
}
HWTEST_F(DispatchWalkerTest, givenBlockedQueueWhenDispatchWalkerIsCalledThenCommandStreamHasGpuAddress) {
MockKernel kernel(program.get(), kernelInfo, *pClDevice);
ASSERT_EQ(CL_SUCCESS, kernel.initialize());
MockMultiDispatchInfo multiDispatchInfo(&kernel);
auto blockedCommandsData = createBlockedCommandsData(*pCmdQ);
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
multiDispatchInfo,
CsrDependencies(),
blockedCommandsData.get(),
nullptr,
nullptr,
nullptr,
nullptr,
CL_COMMAND_NDRANGE_KERNEL);
EXPECT_NE(nullptr, blockedCommandsData->commandStream->getGraphicsAllocation());
EXPECT_NE(0ull, blockedCommandsData->commandStream->getGraphicsAllocation()->getGpuAddress());
}
HWTEST_F(DispatchWalkerTest, givenThereAreAllocationsForReuseWhenDispatchWalkerIsCalledThenCommandStreamObtainsReusableAllocation) {
MockKernel kernel(program.get(), kernelInfo, *pClDevice);
ASSERT_EQ(CL_SUCCESS, kernel.initialize());
MockMultiDispatchInfo multiDispatchInfo(&kernel);
auto &csr = pCmdQ->getGpgpuCommandStreamReceiver();
auto allocation = csr.getMemoryManager()->allocateGraphicsMemoryWithProperties({csr.getRootDeviceIndex(), MemoryConstants::pageSize64k + CSRequirements::csOverfetchSize,
GraphicsAllocation::AllocationType::COMMAND_BUFFER});
csr.getInternalAllocationStorage()->storeAllocation(std::unique_ptr<GraphicsAllocation>{allocation}, REUSABLE_ALLOCATION);
ASSERT_FALSE(csr.getInternalAllocationStorage()->getAllocationsForReuse().peekIsEmpty());
auto blockedCommandsData = createBlockedCommandsData(*pCmdQ);
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
multiDispatchInfo,
CsrDependencies(),
blockedCommandsData.get(),
nullptr,
nullptr,
nullptr,
nullptr,
CL_COMMAND_NDRANGE_KERNEL);
EXPECT_TRUE(csr.getInternalAllocationStorage()->getAllocationsForReuse().peekIsEmpty());
EXPECT_EQ(allocation, blockedCommandsData->commandStream->getGraphicsAllocation());
}
HWTEST_F(DispatchWalkerTest, GivenMultipleKernelsWhenDispatchingWalkerThenWorkDimensionsAreCorrect) {
MockKernel kernel1(program.get(), kernelInfo, *pClDevice);
ASSERT_EQ(CL_SUCCESS, kernel1.initialize());
MockKernel kernel2(program.get(), kernelInfo, *pClDevice);
ASSERT_EQ(CL_SUCCESS, kernel2.initialize());
MockMultiDispatchInfo multiDispatchInfo(std::vector<Kernel *>({&kernel1, &kernel2}));
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
multiDispatchInfo,
CsrDependencies(),
nullptr,
nullptr,
nullptr,
nullptr,
nullptr,
CL_COMMAND_NDRANGE_KERNEL);
for (auto &dispatchInfo : multiDispatchInfo) {
auto &kernel = *dispatchInfo.getKernel();
EXPECT_EQ(*kernel.workDim, dispatchInfo.getDim());
}
}
HWCMDTEST_F(IGFX_GEN8_CORE, DispatchWalkerTest, GivenMultipleKernelsWhenDispatchingWalkerThenInterfaceDescriptorsAreProgrammedCorrectly) {
using INTERFACE_DESCRIPTOR_DATA = typename FamilyType::INTERFACE_DESCRIPTOR_DATA;
auto memoryManager = this->pDevice->getMemoryManager();
auto kernelIsaAllocation = memoryManager->allocateGraphicsMemoryWithProperties({pDevice->getRootDeviceIndex(), MemoryConstants::pageSize, GraphicsAllocation::AllocationType::KERNEL_ISA});
auto kernelIsaWithSamplerAllocation = memoryManager->allocateGraphicsMemoryWithProperties({pDevice->getRootDeviceIndex(), MemoryConstants::pageSize, GraphicsAllocation::AllocationType::KERNEL_ISA});
kernelInfo.kernelAllocation = kernelIsaAllocation;
kernelInfoWithSampler.kernelAllocation = kernelIsaWithSamplerAllocation;
auto gpuAddress1 = kernelIsaAllocation->getGpuAddressToPatch();
auto gpuAddress2 = kernelIsaWithSamplerAllocation->getGpuAddressToPatch();
MockKernel kernel1(program.get(), kernelInfo, *pClDevice);
ASSERT_EQ(CL_SUCCESS, kernel1.initialize());
MockKernel kernel2(program.get(), kernelInfoWithSampler, *pClDevice);
ASSERT_EQ(CL_SUCCESS, kernel2.initialize());
MockMultiDispatchInfo multiDispatchInfo(std::vector<Kernel *>({&kernel1, &kernel2}));
// create Indirect DSH heap
auto &indirectHeap = pCmdQ->getIndirectHeap(IndirectHeap::DYNAMIC_STATE, 8192);
indirectHeap.align(HardwareCommandsHelper<FamilyType>::alignInterfaceDescriptorData);
auto dshBeforeMultiDisptach = indirectHeap.getUsed();
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
multiDispatchInfo,
CsrDependencies(),
nullptr,
nullptr,
nullptr,
nullptr,
nullptr,
CL_COMMAND_NDRANGE_KERNEL);
auto dshAfterMultiDisptach = indirectHeap.getUsed();
auto numberOfDispatches = multiDispatchInfo.size();
auto interfaceDesriptorTableSize = numberOfDispatches * sizeof(INTERFACE_DESCRIPTOR_DATA);
EXPECT_LE(dshBeforeMultiDisptach + interfaceDesriptorTableSize, dshAfterMultiDisptach);
INTERFACE_DESCRIPTOR_DATA *pID = reinterpret_cast<INTERFACE_DESCRIPTOR_DATA *>(ptrOffset(indirectHeap.getCpuBase(), dshBeforeMultiDisptach));
for (uint32_t index = 0; index < multiDispatchInfo.size(); index++) {
uint32_t addressLow = pID[index].getKernelStartPointer();
uint32_t addressHigh = pID[index].getKernelStartPointerHigh();
uint64_t fullAddress = ((uint64_t)addressHigh << 32) | addressLow;
if (index > 0) {
uint32_t addressLowOfPrevious = pID[index - 1].getKernelStartPointer();
uint32_t addressHighOfPrevious = pID[index - 1].getKernelStartPointerHigh();
uint64_t addressPrevious = ((uint64_t)addressHighOfPrevious << 32) | addressLowOfPrevious;
uint64_t address = ((uint64_t)addressHigh << 32) | addressLow;
EXPECT_NE(addressPrevious, address);
}
if (index == 0) {
auto samplerPointer = pID[index].getSamplerStatePointer();
auto samplerCount = pID[index].getSamplerCount();
EXPECT_EQ(0u, samplerPointer);
EXPECT_EQ(0u, samplerCount);
EXPECT_EQ(fullAddress, gpuAddress1);
}
if (index == 1) {
auto samplerPointer = pID[index].getSamplerStatePointer();
auto samplerCount = pID[index].getSamplerCount();
EXPECT_NE(0u, samplerPointer);
EXPECT_EQ(1u, samplerCount);
EXPECT_EQ(fullAddress, gpuAddress2);
}
}
HardwareParse hwParser;
auto &cmdStream = pCmdQ->getCS(0);
hwParser.parseCommands<FamilyType>(cmdStream, 0);
hwParser.findHardwareCommands<FamilyType>();
auto cmd = hwParser.getCommand<typename FamilyType::MEDIA_INTERFACE_DESCRIPTOR_LOAD>();
EXPECT_NE(nullptr, cmd);
auto IDStartAddress = cmd->getInterfaceDescriptorDataStartAddress();
auto IDSize = cmd->getInterfaceDescriptorTotalLength();
EXPECT_EQ(dshBeforeMultiDisptach, IDStartAddress);
EXPECT_EQ(interfaceDesriptorTableSize, IDSize);
memoryManager->freeGraphicsMemory(kernelIsaAllocation);
memoryManager->freeGraphicsMemory(kernelIsaWithSamplerAllocation);
}
HWCMDTEST_F(IGFX_GEN8_CORE, DispatchWalkerTest, GivenMultipleKernelsWhenDispatchingWalkerThenGpgpuWalkerIdOffsetIsProgrammedCorrectly) {
using GPGPU_WALKER = typename FamilyType::GPGPU_WALKER;
MockKernel kernel1(program.get(), kernelInfo, *pClDevice);
ASSERT_EQ(CL_SUCCESS, kernel1.initialize());
MockKernel kernel2(program.get(), kernelInfoWithSampler, *pClDevice);
ASSERT_EQ(CL_SUCCESS, kernel2.initialize());
MockMultiDispatchInfo multiDispatchInfo(std::vector<Kernel *>({&kernel1, &kernel2}));
// create commandStream
auto &cmdStream = pCmdQ->getCS(0);
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
multiDispatchInfo,
CsrDependencies(),
nullptr,
nullptr,
nullptr,
nullptr,
nullptr,
CL_COMMAND_NDRANGE_KERNEL);
HardwareParse hwParser;
hwParser.parseCommands<FamilyType>(cmdStream, 0);
hwParser.findHardwareCommands<FamilyType>();
auto walkerItor = hwParser.itorWalker;
ASSERT_NE(hwParser.cmdList.end(), walkerItor);
for (uint32_t index = 0; index < multiDispatchInfo.size(); index++) {
ASSERT_NE(hwParser.cmdList.end(), walkerItor);
auto *gpgpuWalker = (GPGPU_WALKER *)*walkerItor;
auto IDIndex = gpgpuWalker->getInterfaceDescriptorOffset();
EXPECT_EQ(index, IDIndex);
// move walker iterator
walkerItor++;
walkerItor = find<GPGPU_WALKER *>(walkerItor, hwParser.cmdList.end());
}
}
HWCMDTEST_F(IGFX_GEN8_CORE, DispatchWalkerTest, GivenMultipleKernelsWhenDispatchingWalkerThenThreadGroupIdStartingCoordinatesAreProgrammedCorrectly) {
using GPGPU_WALKER = typename FamilyType::GPGPU_WALKER;
MockKernel kernel1(program.get(), kernelInfo, *pClDevice);
ASSERT_EQ(CL_SUCCESS, kernel1.initialize());
MockKernel kernel2(program.get(), kernelInfoWithSampler, *pClDevice);
ASSERT_EQ(CL_SUCCESS, kernel2.initialize());
MockMultiDispatchInfo multiDispatchInfo(std::vector<Kernel *>({&kernel1, &kernel2}));
// create commandStream
auto &cmdStream = pCmdQ->getCS(0);
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
multiDispatchInfo,
CsrDependencies(),
nullptr,
nullptr,
nullptr,
nullptr,
nullptr,
CL_COMMAND_NDRANGE_KERNEL);
HardwareParse hwParser;
hwParser.parseCommands<FamilyType>(cmdStream, 0);
hwParser.findHardwareCommands<FamilyType>();
auto walkerItor = hwParser.itorWalker;
ASSERT_NE(hwParser.cmdList.end(), walkerItor);
for (uint32_t index = 0; index < multiDispatchInfo.size(); index++) {
ASSERT_NE(hwParser.cmdList.end(), walkerItor);
auto *gpgpuWalker = (GPGPU_WALKER *)*walkerItor;
auto coordinateX = gpgpuWalker->getThreadGroupIdStartingX();
EXPECT_EQ(coordinateX, 0u);
auto coordinateY = gpgpuWalker->getThreadGroupIdStartingY();
EXPECT_EQ(coordinateY, 0u);
auto coordinateZ = gpgpuWalker->getThreadGroupIdStartingResumeZ();
EXPECT_EQ(coordinateZ, 0u);
// move walker iterator
walkerItor++;
walkerItor = find<GPGPU_WALKER *>(walkerItor, hwParser.cmdList.end());
}
}
HWCMDTEST_F(IGFX_GEN8_CORE, DispatchWalkerTest, GivenMultipleDispatchInfoAndSameKernelWhenDispatchingWalkerThenGpgpuWalkerThreadGroupIdStartingCoordinatesAreCorrectlyProgrammed) {
using GPGPU_WALKER = typename FamilyType::GPGPU_WALKER;
MockKernel kernel(program.get(), kernelInfo, *pClDevice);
ASSERT_EQ(CL_SUCCESS, kernel.initialize());
DispatchInfo di1(&kernel, 1, {100, 1, 1}, {10, 1, 1}, {0, 0, 0}, {100, 1, 1}, {10, 1, 1}, {10, 1, 1}, {10, 1, 1}, {0, 0, 0});
DispatchInfo di2(&kernel, 1, {100, 1, 1}, {10, 1, 1}, {0, 0, 0}, {100, 1, 1}, {10, 1, 1}, {10, 1, 1}, {10, 1, 1}, {10, 0, 0});
MockMultiDispatchInfo multiDispatchInfo(std::vector<DispatchInfo *>({&di1, &di2}));
// create commandStream
auto &cmdStream = pCmdQ->getCS(0);
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
multiDispatchInfo,
CsrDependencies(),
nullptr,
nullptr,
nullptr,
nullptr,
nullptr,
CL_COMMAND_NDRANGE_KERNEL);
HardwareParse hwParser;
hwParser.parseCommands<FamilyType>(cmdStream, 0);
hwParser.findHardwareCommands<FamilyType>();
auto walkerItor = hwParser.itorWalker;
ASSERT_NE(hwParser.cmdList.end(), walkerItor);
for (uint32_t index = 0; index < multiDispatchInfo.size(); index++) {
ASSERT_NE(hwParser.cmdList.end(), walkerItor);
auto *gpgpuWalker = (GPGPU_WALKER *)*walkerItor;
auto coordinateX = gpgpuWalker->getThreadGroupIdStartingX();
EXPECT_EQ(coordinateX, index * 10u);
auto coordinateY = gpgpuWalker->getThreadGroupIdStartingY();
EXPECT_EQ(coordinateY, 0u);
auto coordinateZ = gpgpuWalker->getThreadGroupIdStartingResumeZ();
EXPECT_EQ(coordinateZ, 0u);
// move walker iterator
walkerItor++;
walkerItor = find<GPGPU_WALKER *>(walkerItor, hwParser.cmdList.end());
}
}
HWTEST_F(DispatchWalkerTest, GivenCacheFlushAfterWalkerDisabledWhenAllocationRequiresCacheFlushThenFlushCommandNotPresentAfterWalker) {
using PIPE_CONTROL = typename FamilyType::PIPE_CONTROL;
DebugManagerStateRestore dbgRestore;
DebugManager.flags.EnableCacheFlushAfterWalker.set(0);
MockKernel kernel1(program.get(), kernelInfo, *pClDevice);
ASSERT_EQ(CL_SUCCESS, kernel1.initialize());
kernel1.kernelArgRequiresCacheFlush.resize(1);
MockGraphicsAllocation cacheRequiringAllocation;
kernel1.kernelArgRequiresCacheFlush[0] = &cacheRequiringAllocation;
MockMultiDispatchInfo multiDispatchInfo(std::vector<Kernel *>({&kernel1}));
// create commandStream
auto &cmdStream = pCmdQ->getCS(0);
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
multiDispatchInfo,
CsrDependencies(),
nullptr,
nullptr,
nullptr,
nullptr,
nullptr,
CL_COMMAND_NDRANGE_KERNEL);
HardwareParse hwParse;
hwParse.parseCommands<FamilyType>(cmdStream);
PIPE_CONTROL *pipeControl = hwParse.getCommand<PIPE_CONTROL>();
EXPECT_EQ(nullptr, pipeControl);
}
HWTEST_F(DispatchWalkerTest, GivenCacheFlushAfterWalkerEnabledWhenWalkerWithTwoKernelsThenFlushCommandPresentOnce) {
using PIPE_CONTROL = typename FamilyType::PIPE_CONTROL;
DebugManagerStateRestore dbgRestore;
DebugManager.flags.EnableCacheFlushAfterWalker.set(1);
MockKernel kernel1(program.get(), kernelInfo, *pClDevice);
ASSERT_EQ(CL_SUCCESS, kernel1.initialize());
MockKernel kernel2(program.get(), kernelInfoWithSampler, *pClDevice);
ASSERT_EQ(CL_SUCCESS, kernel2.initialize());
kernel1.kernelArgRequiresCacheFlush.resize(1);
kernel2.kernelArgRequiresCacheFlush.resize(1);
MockGraphicsAllocation cacheRequiringAllocation;
kernel1.kernelArgRequiresCacheFlush[0] = &cacheRequiringAllocation;
kernel2.kernelArgRequiresCacheFlush[0] = &cacheRequiringAllocation;
MockMultiDispatchInfo multiDispatchInfo(std::vector<Kernel *>({&kernel1, &kernel2}));
// create commandStream
auto &cmdStream = pCmdQ->getCS(0);
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
multiDispatchInfo,
CsrDependencies(),
nullptr,
nullptr,
nullptr,
nullptr,
nullptr,
CL_COMMAND_NDRANGE_KERNEL);
HardwareParse hwParse;
hwParse.parseCommands<FamilyType>(cmdStream);
uint32_t pipeControlCount = hwParse.getCommandCount<PIPE_CONTROL>();
EXPECT_EQ(pipeControlCount, 1u);
}
HWTEST_F(DispatchWalkerTest, GivenCacheFlushAfterWalkerEnabledWhenTwoWalkersForQueueThenFlushCommandPresentTwice) {
using PIPE_CONTROL = typename FamilyType::PIPE_CONTROL;
DebugManagerStateRestore dbgRestore;
DebugManager.flags.EnableCacheFlushAfterWalker.set(1);
MockKernel kernel1(program.get(), kernelInfo, *pClDevice);
ASSERT_EQ(CL_SUCCESS, kernel1.initialize());
MockKernel kernel2(program.get(), kernelInfoWithSampler, *pClDevice);
ASSERT_EQ(CL_SUCCESS, kernel2.initialize());
kernel1.kernelArgRequiresCacheFlush.resize(1);
kernel2.kernelArgRequiresCacheFlush.resize(1);
MockGraphicsAllocation cacheRequiringAllocation;
kernel1.kernelArgRequiresCacheFlush[0] = &cacheRequiringAllocation;
kernel2.kernelArgRequiresCacheFlush[0] = &cacheRequiringAllocation;
MockMultiDispatchInfo multiDispatchInfo1(std::vector<Kernel *>({&kernel1}));
MockMultiDispatchInfo multiDispatchInfo2(std::vector<Kernel *>({&kernel2}));
// create commandStream
auto &cmdStream = pCmdQ->getCS(0);
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
multiDispatchInfo1,
CsrDependencies(),
nullptr,
nullptr,
nullptr,
nullptr,
nullptr,
CL_COMMAND_NDRANGE_KERNEL);
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
multiDispatchInfo2,
CsrDependencies(),
nullptr,
nullptr,
nullptr,
nullptr,
nullptr,
CL_COMMAND_NDRANGE_KERNEL);
HardwareParse hwParse;
hwParse.parseCommands<FamilyType>(cmdStream);
uint32_t pipeControlCount = hwParse.getCommandCount<PIPE_CONTROL>();
EXPECT_EQ(pipeControlCount, 2u);
}
TEST(DispatchWalker, WhenCalculatingDispatchDimensionsThenCorrectValuesAreReturned) {
Vec3<size_t> dim0{0, 0, 0};
Vec3<size_t> dim1{2, 1, 1};
Vec3<size_t> dim2{2, 2, 1};
Vec3<size_t> dim3{2, 2, 2};
Vec3<size_t> dispatches[] = {dim0, dim1, dim2, dim3};
uint32_t testDims[] = {0, 1, 2, 3};
for (const auto &lhs : testDims) {
for (const auto &rhs : testDims) {
uint32_t dimTest = calculateDispatchDim(dispatches[lhs], dispatches[rhs]);
uint32_t dimRef = std::max(1U, std::max(lhs, rhs));
EXPECT_EQ(dimRef, dimTest);
}
}
}
HWTEST_F(DispatchWalkerTest, givenKernelWhenAuxToNonAuxWhenTranslationRequiredThenPipeControlWithStallAndDCFlushAdded) {
BuiltinDispatchInfoBuilder &baseBuilder = BuiltInDispatchBuilderOp::getBuiltinDispatchInfoBuilder(EBuiltInOps::AuxTranslation, *pDevice);
auto &builder = static_cast<BuiltInOp<EBuiltInOps::AuxTranslation> &>(baseBuilder);
MockKernel kernel(program.get(), kernelInfo, *pClDevice);
kernelInfo.workloadInfo.workDimOffset = 0;
ASSERT_EQ(CL_SUCCESS, kernel.initialize());
auto &cmdStream = pCmdQ->getCS(0);
void *buffer = cmdStream.getCpuBase();
kernel.auxTranslationRequired = true;
MockBuffer mockBuffer[2];
MultiDispatchInfo multiDispatchInfo;
MemObjsForAuxTranslation memObjsForAuxTranslation;
multiDispatchInfo.setMemObjsForAuxTranslation(memObjsForAuxTranslation);
memObjsForAuxTranslation.insert(&mockBuffer[0]);
memObjsForAuxTranslation.insert(&mockBuffer[1]);
BuiltinOpParams builtinOpsParams;
builtinOpsParams.auxTranslationDirection = AuxTranslationDirection::AuxToNonAux;
builder.buildDispatchInfosForAuxTranslation<FamilyType>(multiDispatchInfo, builtinOpsParams);
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
multiDispatchInfo,
CsrDependencies(),
nullptr,
nullptr,
nullptr,
nullptr,
nullptr,
CL_COMMAND_NDRANGE_KERNEL);
auto sizeUsed = cmdStream.getUsed();
GenCmdList cmdList;
ASSERT_TRUE(FamilyType::PARSE::parseCommandBuffer(cmdList, buffer, sizeUsed));
auto pipeControls = findAll<typename FamilyType::PIPE_CONTROL *>(cmdList.begin(), cmdList.end());
ASSERT_EQ(2u, pipeControls.size());
auto beginPipeControl = genCmdCast<typename FamilyType::PIPE_CONTROL *>(*(pipeControls[0]));
EXPECT_TRUE(beginPipeControl->getDcFlushEnable());
EXPECT_TRUE(beginPipeControl->getCommandStreamerStallEnable());
auto endPipeControl = genCmdCast<typename FamilyType::PIPE_CONTROL *>(*(pipeControls[1]));
bool dcFlushRequired = (pClDevice->getHardwareInfo().platform.eRenderCoreFamily == IGFX_GEN8_CORE);
EXPECT_EQ(dcFlushRequired, endPipeControl->getDcFlushEnable());
EXPECT_TRUE(endPipeControl->getCommandStreamerStallEnable());
}
HWTEST_F(DispatchWalkerTest, givenKernelWhenNonAuxToAuxWhenTranslationRequiredThenPipeControlWithStallAdded) {
BuiltinDispatchInfoBuilder &baseBuilder = BuiltInDispatchBuilderOp::getBuiltinDispatchInfoBuilder(EBuiltInOps::AuxTranslation, *pDevice);
auto &builder = static_cast<BuiltInOp<EBuiltInOps::AuxTranslation> &>(baseBuilder);
MockKernel kernel(program.get(), kernelInfo, *pClDevice);
kernelInfo.workloadInfo.workDimOffset = 0;
ASSERT_EQ(CL_SUCCESS, kernel.initialize());
auto &cmdStream = pCmdQ->getCS(0);
void *buffer = cmdStream.getCpuBase();
kernel.auxTranslationRequired = true;
MockBuffer mockBuffer[2];
MultiDispatchInfo multiDispatchInfo;
MemObjsForAuxTranslation memObjsForAuxTranslation;
multiDispatchInfo.setMemObjsForAuxTranslation(memObjsForAuxTranslation);
memObjsForAuxTranslation.insert(&mockBuffer[0]);
memObjsForAuxTranslation.insert(&mockBuffer[1]);
BuiltinOpParams builtinOpsParams;
builtinOpsParams.auxTranslationDirection = AuxTranslationDirection::NonAuxToAux;
builder.buildDispatchInfosForAuxTranslation<FamilyType>(multiDispatchInfo, builtinOpsParams);
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
multiDispatchInfo,
CsrDependencies(),
nullptr,
nullptr,
nullptr,
nullptr,
nullptr,
CL_COMMAND_NDRANGE_KERNEL);
auto sizeUsed = cmdStream.getUsed();
GenCmdList cmdList;
ASSERT_TRUE(FamilyType::PARSE::parseCommandBuffer(cmdList, buffer, sizeUsed));
auto pipeControls = findAll<typename FamilyType::PIPE_CONTROL *>(cmdList.begin(), cmdList.end());
ASSERT_EQ(2u, pipeControls.size());
bool dcFlushRequired = (pClDevice->getHardwareInfo().platform.eRenderCoreFamily == IGFX_GEN8_CORE);
auto beginPipeControl = genCmdCast<typename FamilyType::PIPE_CONTROL *>(*(pipeControls[0]));
EXPECT_TRUE(beginPipeControl->getDcFlushEnable());
EXPECT_TRUE(beginPipeControl->getCommandStreamerStallEnable());
auto endPipeControl = genCmdCast<typename FamilyType::PIPE_CONTROL *>(*(pipeControls[1]));
EXPECT_EQ(dcFlushRequired, endPipeControl->getDcFlushEnable());
EXPECT_TRUE(endPipeControl->getCommandStreamerStallEnable());
}
struct ProfilingCommandsTest : public DispatchWalkerTest, ::testing::WithParamInterface<bool> {
void SetUp() override {
DispatchWalkerTest::SetUp();
}
void TearDown() override {
DispatchWalkerTest::TearDown();
}
};
HWTEST_P(ProfilingCommandsTest, givenKernelWhenProfilingCommandStartIsTakenThenTimeStampAddressIsProgrammedCorrectly) {
using MI_STORE_REGISTER_MEM = typename FamilyType::MI_STORE_REGISTER_MEM;
using PIPE_CONTROL = typename FamilyType::PIPE_CONTROL;
bool checkForStart = GetParam();
auto &cmdStream = pCmdQ->getCS(0);
TagAllocator<HwTimeStamps> timeStampAllocator(pDevice->getRootDeviceIndex(), this->pDevice->getMemoryManager(), 10,
MemoryConstants::cacheLineSize, sizeof(HwTimeStamps), false, {});
auto hwTimeStamp1 = timeStampAllocator.getTag();
ASSERT_NE(nullptr, hwTimeStamp1);
if (checkForStart) {
GpgpuWalkerHelper<FamilyType>::dispatchProfilingCommandsStart(*hwTimeStamp1, &cmdStream, pDevice->getHardwareInfo());
} else {
GpgpuWalkerHelper<FamilyType>::dispatchProfilingCommandsEnd(*hwTimeStamp1, &cmdStream);
}
auto hwTimeStamp2 = timeStampAllocator.getTag();
ASSERT_NE(nullptr, hwTimeStamp2);
if (checkForStart) {
GpgpuWalkerHelper<FamilyType>::dispatchProfilingCommandsStart(*hwTimeStamp2, &cmdStream, pDevice->getHardwareInfo());
} else {
GpgpuWalkerHelper<FamilyType>::dispatchProfilingCommandsEnd(*hwTimeStamp2, &cmdStream);
}
GenCmdList cmdList;
ASSERT_TRUE(FamilyType::PARSE::parseCommandBuffer(cmdList, cmdStream.getCpuBase(), cmdStream.getUsed()));
auto itorStoreReg = find<typename FamilyType::MI_STORE_REGISTER_MEM *>(cmdList.begin(), cmdList.end());
ASSERT_NE(cmdList.end(), itorStoreReg);
auto storeReg = genCmdCast<MI_STORE_REGISTER_MEM *>(*itorStoreReg);
ASSERT_NE(nullptr, storeReg);
uint64_t gpuAddress = storeReg->getMemoryAddress();
auto contextTimestampFieldOffset = checkForStart ? offsetof(HwTimeStamps, ContextStartTS) : offsetof(HwTimeStamps, ContextEndTS);
uint64_t expectedAddress = hwTimeStamp1->getGpuAddress() + contextTimestampFieldOffset;
EXPECT_EQ(expectedAddress, gpuAddress);
itorStoreReg++;
itorStoreReg = find<typename FamilyType::MI_STORE_REGISTER_MEM *>(itorStoreReg, cmdList.end());
ASSERT_NE(cmdList.end(), itorStoreReg);
storeReg = genCmdCast<MI_STORE_REGISTER_MEM *>(*itorStoreReg);
ASSERT_NE(nullptr, storeReg);
gpuAddress = storeReg->getMemoryAddress();
expectedAddress = hwTimeStamp2->getGpuAddress() + contextTimestampFieldOffset;
EXPECT_EQ(expectedAddress, gpuAddress);
if (checkForStart) {
auto itorPipeCtrl = find<typename FamilyType::PIPE_CONTROL *>(cmdList.begin(), cmdList.end());
ASSERT_NE(cmdList.end(), itorPipeCtrl);
if (HardwareCommandsHelper<FamilyType>::isPipeControlWArequired(pDevice->getHardwareInfo())) {
itorPipeCtrl++;
}
if (UnitTestHelper<FamilyType>::isAdditionalSynchronizationRequired(pDevice->getHardwareInfo())) {
itorPipeCtrl++;
}
auto pipeControl = genCmdCast<PIPE_CONTROL *>(*itorPipeCtrl);
ASSERT_NE(nullptr, pipeControl);
gpuAddress = static_cast<uint64_t>(pipeControl->getAddress()) | (static_cast<uint64_t>(pipeControl->getAddressHigh()) << 32);
expectedAddress = hwTimeStamp1->getGpuAddress() + offsetof(HwTimeStamps, GlobalStartTS);
EXPECT_EQ(expectedAddress, gpuAddress);
itorPipeCtrl++;
itorPipeCtrl = find<typename FamilyType::PIPE_CONTROL *>(itorPipeCtrl, cmdList.end());
if (HardwareCommandsHelper<FamilyType>::isPipeControlWArequired(pDevice->getHardwareInfo())) {
itorPipeCtrl++;
}
if (UnitTestHelper<FamilyType>::isAdditionalSynchronizationRequired(pDevice->getHardwareInfo())) {
itorPipeCtrl++;
}
ASSERT_NE(cmdList.end(), itorPipeCtrl);
pipeControl = genCmdCast<PIPE_CONTROL *>(*itorPipeCtrl);
ASSERT_NE(nullptr, pipeControl);
gpuAddress = static_cast<uint64_t>(pipeControl->getAddress()) | static_cast<uint64_t>(pipeControl->getAddressHigh()) << 32;
expectedAddress = hwTimeStamp2->getGpuAddress() + offsetof(HwTimeStamps, GlobalStartTS);
EXPECT_EQ(expectedAddress, gpuAddress);
}
}
INSTANTIATE_TEST_CASE_P(StartEndFlag,
ProfilingCommandsTest, ::testing::Bool());
Reference in New Issue View Git Blame Copy Permalink