intel/compute-runtime
1
0
Fork 0
mirror of https://github.com/intel/compute-runtime.git synced 2025-12-31 20:13:04 +08:00
Code Issues Packages Projects Releases Wiki Activity

Store ClDevice in DispatchInfo

Browse Source 
Related-To: NEO-5001
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
This commit is contained in:
Mateusz Jablonski
2020-11-16 12:08:30 +01:00
committed by Compute-Runtime-Automation
parent b848f20abc
commit 97119f5f3e
56 changed files with 335 additions and 317 deletions
Download Patch File Download Diff File Expand all files Collapse all files

64
opencl/test/unit_test/command_queue/dispatch_walker_tests.cpp
View File

@@ -169,7 +169,7 @@ HWTEST_F(DispatchWalkerTest, WhenDispatchingWalkerThenCommandStreamMemoryIsntCha
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);
DispatchInfo dispatchInfo(pClDevice, const_cast<MockKernel *>(&kernel), dimensions, workItems, nullptr, globalOffsets);
MultiDispatchInfo multiDispatchInfo;
multiDispatchInfo.push(dispatchInfo);
HardwareInterface<FamilyType>::dispatchWalker(
@@ -216,7 +216,7 @@ HWTEST_F(DispatchWalkerTest, GivenNoLocalIdsWhenDispatchingWalkerThenWalkerIsDis
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);
DispatchInfo dispatchInfo(pClDevice, const_cast<MockKernel *>(&kernel), dimensions, workItems, nullptr, globalOffsets);
MultiDispatchInfo multiDispatchInfo;
multiDispatchInfo.push(dispatchInfo);
HardwareInterface<FamilyType>::dispatchWalker(
@@ -245,7 +245,7 @@ HWTEST_F(DispatchWalkerTest, GivenDefaultLwsAlgorithmWhenDispatchingWalkerThenDi
for (uint32_t dimension = 1; dimension <= 3; ++dimension) {
workItems[dimension - 1] = 256;
DispatchInfo dispatchInfo(const_cast<MockKernel *>(&kernel), dimension, workItems, nullptr, globalOffsets);
DispatchInfo dispatchInfo(pClDevice, const_cast<MockKernel *>(&kernel), dimension, workItems, nullptr, globalOffsets);
MultiDispatchInfo multiDispatchInfo;
multiDispatchInfo.push(dispatchInfo);
HardwareInterface<FamilyType>::dispatchWalker(
@@ -275,7 +275,7 @@ HWTEST_F(DispatchWalkerTest, GivenSquaredLwsAlgorithmWhenDispatchingWalkerThenDi
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);
DispatchInfo dispatchInfo(pClDevice, const_cast<MockKernel *>(&kernel), dimension, workItems, nullptr, globalOffsets);
MultiDispatchInfo multiDispatchInfo;
multiDispatchInfo.push(dispatchInfo);
HardwareInterface<FamilyType>::dispatchWalker(
@@ -303,7 +303,7 @@ HWTEST_F(DispatchWalkerTest, GivenNdLwsAlgorithmWhenDispatchingWalkerThenDimensi
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);
DispatchInfo dispatchInfo(pClDevice, const_cast<MockKernel *>(&kernel), dimension, workItems, nullptr, globalOffsets);
MultiDispatchInfo multiDispatchInfo;
multiDispatchInfo.push(dispatchInfo);
HardwareInterface<FamilyType>::dispatchWalker(
@@ -332,7 +332,7 @@ HWTEST_F(DispatchWalkerTest, GivenOldLwsAlgorithmWhenDispatchingWalkerThenDimens
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);
DispatchInfo dispatchInfo(pClDevice, const_cast<MockKernel *>(&kernel), dimension, workItems, nullptr, globalOffsets);
MultiDispatchInfo multiDispatchInfo;
multiDispatchInfo.push(dispatchInfo);
HardwareInterface<FamilyType>::dispatchWalker(
@@ -361,7 +361,7 @@ HWTEST_F(DispatchWalkerTest, GivenNumWorkGroupsWhenDispatchingWalkerThenNumWorkG
size_t workGroupSize[3] = {1, 1, 1};
cl_uint dimensions = 3;
DispatchInfo dispatchInfo(const_cast<MockKernel *>(&kernel), dimensions, workItems, workGroupSize, globalOffsets);
DispatchInfo dispatchInfo(pClDevice, const_cast<MockKernel *>(&kernel), dimensions, workItems, workGroupSize, globalOffsets);
MultiDispatchInfo multiDispatchInfo;
multiDispatchInfo.push(dispatchInfo);
HardwareInterface<FamilyType>::dispatchWalker(
@@ -392,7 +392,7 @@ HWTEST_F(DispatchWalkerTest, GivenNoLocalWorkSizeAndDefaultAlgorithmWhenDispatch
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);
DispatchInfo dispatchInfo(pClDevice, const_cast<MockKernel *>(&kernel), dimensions, workItems, nullptr, globalOffsets);
MultiDispatchInfo multiDispatchInfo;
multiDispatchInfo.push(dispatchInfo);
HardwareInterface<FamilyType>::dispatchWalker(
@@ -422,7 +422,7 @@ HWTEST_F(DispatchWalkerTest, GivenNoLocalWorkSizeAndNdOnWhenDispatchingWalkerThe
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);
DispatchInfo dispatchInfo(pClDevice, const_cast<MockKernel *>(&kernel), dimensions, workItems, nullptr, globalOffsets);
MultiDispatchInfo multiDispatchInfo;
multiDispatchInfo.push(dispatchInfo);
HardwareInterface<FamilyType>::dispatchWalker(
@@ -453,7 +453,7 @@ HWTEST_F(DispatchWalkerTest, GivenNoLocalWorkSizeAndSquaredAlgorithmWhenDispatch
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);
DispatchInfo dispatchInfo(pClDevice, const_cast<MockKernel *>(&kernel), dimensions, workItems, nullptr, globalOffsets);
MultiDispatchInfo multiDispatchInfo;
multiDispatchInfo.push(dispatchInfo);
HardwareInterface<FamilyType>::dispatchWalker(
@@ -484,7 +484,7 @@ HWTEST_F(DispatchWalkerTest, GivenNoLocalWorkSizeAndSquaredAlgorithmOffAndNdOffW
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);
DispatchInfo dispatchInfo(pClDevice, const_cast<MockKernel *>(&kernel), dimensions, workItems, nullptr, globalOffsets);
MultiDispatchInfo multiDispatchInfo;
multiDispatchInfo.push(dispatchInfo);
HardwareInterface<FamilyType>::dispatchWalker(
@@ -513,7 +513,7 @@ HWTEST_F(DispatchWalkerTest, GivenNoLocalWorkSizeWhenDispatchingWalkerThenLwsIsC
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);
DispatchInfo dispatchInfo(pClDevice, const_cast<MockKernel *>(&kernel), dimensions, workItems, workGroupSize, globalOffsets);
MultiDispatchInfo multiDispatchInfo;
multiDispatchInfo.push(dispatchInfo);
HardwareInterface<FamilyType>::dispatchWalker(
@@ -545,7 +545,7 @@ HWTEST_F(DispatchWalkerTest, GivenTwoSetsOfLwsOffsetsWhenDispatchingWalkerThenLw
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);
DispatchInfo dispatchInfo(pClDevice, const_cast<MockKernel *>(&kernel), dimensions, workItems, workGroupSize, globalOffsets);
MultiDispatchInfo multiDispatchInfo;
multiDispatchInfo.push(dispatchInfo);
HardwareInterface<FamilyType>::dispatchWalker(
@@ -579,8 +579,8 @@ HWTEST_F(DispatchWalkerTest, GivenSplitKernelWhenDispatchingWalkerThenLwsIsCorre
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});
DispatchInfo di1(pClDevice, &kernel1, 3, {10, 10, 10}, {1, 2, 3}, {0, 0, 0});
DispatchInfo di2(pClDevice, &kernel2, 3, {10, 10, 10}, {4, 5, 6}, {0, 0, 0});
MockMultiDispatchInfo multiDispatchInfo(std::vector<DispatchInfo *>({&di1, &di2}));
@@ -627,8 +627,8 @@ HWTEST_F(DispatchWalkerTest, GivenSplitWalkerWhenDispatchingWalkerThenLwsIsCorre
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});
DispatchInfo di1(pClDevice, &kernel1, 3, {10, 10, 10}, {1, 2, 3}, {0, 0, 0});
DispatchInfo di2(pClDevice, &mainKernel, 3, {10, 10, 10}, {4, 5, 6}, {0, 0, 0});
MultiDispatchInfo multiDispatchInfo(&mainKernel);
multiDispatchInfo.push(di1);
@@ -682,7 +682,7 @@ HWTEST_F(DispatchWalkerTest, GivenBlockedQueueWhenDispatchingWalkerThenCommandSt
auto blockedCommandsData = createBlockedCommandsData(*pCmdQ);
DispatchInfo dispatchInfo(const_cast<MockKernel *>(&kernel), dimensions, workItems, workGroupSize, globalOffsets);
DispatchInfo dispatchInfo(pClDevice, const_cast<MockKernel *>(&kernel), dimensions, workItems, workGroupSize, globalOffsets);
MultiDispatchInfo multiDispatchInfo;
multiDispatchInfo.push(dispatchInfo);
HardwareInterface<FamilyType>::dispatchWalker(
@@ -715,7 +715,7 @@ HWTEST_F(DispatchWalkerTest, GivenBlockedQueueWhenDispatchingWalkerThenRequiredH
cl_uint dimensions = 1;
auto blockedCommandsData = createBlockedCommandsData(*pCmdQ);
DispatchInfo dispatchInfo(const_cast<MockKernel *>(&kernel), dimensions, workItems, workGroupSize, globalOffsets);
DispatchInfo dispatchInfo(pClDevice, const_cast<MockKernel *>(&kernel), dimensions, workItems, workGroupSize, globalOffsets);
MultiDispatchInfo multiDispatchInfo(&kernel);
multiDispatchInfo.push(dispatchInfo);
HardwareInterface<FamilyType>::dispatchWalker(
@@ -767,7 +767,7 @@ HWTEST_F(DispatchWalkerTest, GivenBlockedQueueWhenDispatchingWalkerThenRequiredH
MockKernel kernel(program.get(), kernelInfo);
ASSERT_EQ(CL_SUCCESS, kernel.initialize());
MockMultiDispatchInfo multiDispatchInfo(&kernel);
MockMultiDispatchInfo multiDispatchInfo(pClDevice, &kernel);
auto blockedCommandsData = createBlockedCommandsData(*pCmdQ);
@@ -794,7 +794,7 @@ HWTEST_F(DispatchWalkerTest, GivenBlockedQueueWhenDispatchingWalkerThenRequiredH
HWTEST_F(DispatchWalkerTest, givenBlockedQueueWhenDispatchWalkerIsCalledThenCommandStreamHasGpuAddress) {
MockKernel kernel(program.get(), kernelInfo);
ASSERT_EQ(CL_SUCCESS, kernel.initialize());
MockMultiDispatchInfo multiDispatchInfo(&kernel);
MockMultiDispatchInfo multiDispatchInfo(pClDevice, &kernel);
auto blockedCommandsData = createBlockedCommandsData(*pCmdQ);
HardwareInterface<FamilyType>::dispatchWalker(
@@ -815,7 +815,7 @@ HWTEST_F(DispatchWalkerTest, givenBlockedQueueWhenDispatchWalkerIsCalledThenComm
HWTEST_F(DispatchWalkerTest, givenThereAreAllocationsForReuseWhenDispatchWalkerIsCalledThenCommandStreamObtainsReusableAllocation) {
MockKernel kernel(program.get(), kernelInfo);
ASSERT_EQ(CL_SUCCESS, kernel.initialize());
MockMultiDispatchInfo multiDispatchInfo(&kernel);
MockMultiDispatchInfo multiDispatchInfo(pClDevice, &kernel);
auto &csr = pCmdQ->getGpgpuCommandStreamReceiver();
auto allocation = csr.getMemoryManager()->allocateGraphicsMemoryWithProperties({csr.getRootDeviceIndex(), MemoryConstants::pageSize64k + CSRequirements::csOverfetchSize,
@@ -845,7 +845,7 @@ HWTEST_F(DispatchWalkerTest, GivenMultipleKernelsWhenDispatchingWalkerThenWorkDi
MockKernel kernel2(program.get(), kernelInfo);
ASSERT_EQ(CL_SUCCESS, kernel2.initialize());
MockMultiDispatchInfo multiDispatchInfo(std::vector<Kernel *>({&kernel1, &kernel2}));
MockMultiDispatchInfo multiDispatchInfo(pClDevice, std::vector<Kernel *>({&kernel1, &kernel2}));
HardwareInterface<FamilyType>::dispatchWalker(
*pCmdQ,
@@ -880,7 +880,7 @@ HWCMDTEST_F(IGFX_GEN8_CORE, DispatchWalkerTest, GivenMultipleKernelsWhenDispatch
MockKernel kernel2(program.get(), kernelInfoWithSampler);
ASSERT_EQ(CL_SUCCESS, kernel2.initialize());
MockMultiDispatchInfo multiDispatchInfo(std::vector<Kernel *>({&kernel1, &kernel2}));
MockMultiDispatchInfo multiDispatchInfo(pClDevice, std::vector<Kernel *>({&kernel1, &kernel2}));
// create Indirect DSH heap
auto &indirectHeap = pCmdQ->getIndirectHeap(IndirectHeap::DYNAMIC_STATE, 8192);
@@ -967,7 +967,7 @@ HWCMDTEST_F(IGFX_GEN8_CORE, DispatchWalkerTest, GivenMultipleKernelsWhenDispatch
MockKernel kernel2(program.get(), kernelInfoWithSampler);
ASSERT_EQ(CL_SUCCESS, kernel2.initialize());
MockMultiDispatchInfo multiDispatchInfo(std::vector<Kernel *>({&kernel1, &kernel2}));
MockMultiDispatchInfo multiDispatchInfo(pClDevice, std::vector<Kernel *>({&kernel1, &kernel2}));
// create commandStream
auto &cmdStream = pCmdQ->getCS(0);
@@ -1012,7 +1012,7 @@ HWCMDTEST_F(IGFX_GEN8_CORE, DispatchWalkerTest, GivenMultipleKernelsWhenDispatch
MockKernel kernel2(program.get(), kernelInfoWithSampler);
ASSERT_EQ(CL_SUCCESS, kernel2.initialize());
MockMultiDispatchInfo multiDispatchInfo(std::vector<Kernel *>({&kernel1, &kernel2}));
MockMultiDispatchInfo multiDispatchInfo(pClDevice, std::vector<Kernel *>({&kernel1, &kernel2}));
// create commandStream
auto &cmdStream = pCmdQ->getCS(0);
@@ -1059,8 +1059,8 @@ HWCMDTEST_F(IGFX_GEN8_CORE, DispatchWalkerTest, GivenMultipleDispatchInfoAndSame
MockKernel kernel(program.get(), kernelInfo);
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});
DispatchInfo di1(pClDevice, &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(pClDevice, &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}));
@@ -1115,7 +1115,7 @@ HWTEST_F(DispatchWalkerTest, GivenCacheFlushAfterWalkerDisabledWhenAllocationReq
MockGraphicsAllocation cacheRequiringAllocation;
kernel1.kernelArgRequiresCacheFlush[0] = &cacheRequiringAllocation;
MockMultiDispatchInfo multiDispatchInfo(std::vector<Kernel *>({&kernel1}));
MockMultiDispatchInfo multiDispatchInfo(pClDevice, std::vector<Kernel *>({&kernel1}));
// create commandStream
auto &cmdStream = pCmdQ->getCS(0);
@@ -1153,7 +1153,7 @@ HWTEST_F(DispatchWalkerTest, GivenCacheFlushAfterWalkerEnabledWhenWalkerWithTwoK
kernel1.kernelArgRequiresCacheFlush[0] = &cacheRequiringAllocation;
kernel2.kernelArgRequiresCacheFlush[0] = &cacheRequiringAllocation;
MockMultiDispatchInfo multiDispatchInfo(std::vector<Kernel *>({&kernel1, &kernel2}));
MockMultiDispatchInfo multiDispatchInfo(pClDevice, std::vector<Kernel *>({&kernel1, &kernel2}));
// create commandStream
auto &cmdStream = pCmdQ->getCS(0);
@@ -1191,8 +1191,8 @@ HWTEST_F(DispatchWalkerTest, GivenCacheFlushAfterWalkerEnabledWhenTwoWalkersForQ
kernel1.kernelArgRequiresCacheFlush[0] = &cacheRequiringAllocation;
kernel2.kernelArgRequiresCacheFlush[0] = &cacheRequiringAllocation;
MockMultiDispatchInfo multiDispatchInfo1(std::vector<Kernel *>({&kernel1}));
MockMultiDispatchInfo multiDispatchInfo2(std::vector<Kernel *>({&kernel2}));
MockMultiDispatchInfo multiDispatchInfo1(pClDevice, std::vector<Kernel *>({&kernel1}));
MockMultiDispatchInfo multiDispatchInfo2(pClDevice, std::vector<Kernel *>({&kernel2}));
// create commandStream
auto &cmdStream = pCmdQ->getCS(0);