mirror of
https://github.com/intel/compute-runtime.git
synced 2025-09-10 12:53:42 +08:00
Introduce MultiDeviceKernel class
Related-To: NEO-5001 Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
This commit is contained in:
Mateusz Jablonski
committed by
Compute-Runtime-Automation
Compute-Runtime-Automation
parent
71940061b8
commit
04eca48ee0
@ -761,7 +761,8 @@ TEST_F(GTPinTests, givenInitializedGTPinInterfaceWhenKernelIsExecutedThenGTPinCa
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
EXPECT_EQ(prevCount11 + 1, KernelCreateCallbackCount);
|
||||
|
||||
Kernel *pKernel1 = (Kernel *)kernel1;
|
||||
MultiDeviceKernel *pMultiDeviceKernel1 = static_cast<MultiDeviceKernel *>(kernel1);
|
||||
Kernel *pKernel1 = pMultiDeviceKernel1->getKernel(rootDeviceIndex);
|
||||
const KernelInfo &kInfo1 = pKernel1->getKernelInfo(rootDeviceIndex);
|
||||
uint64_t gtpinKernelId1 = pKernel1->getKernelId(rootDeviceIndex);
|
||||
EXPECT_EQ(kInfo1.shaderHashCode, gtpinKernelId1);
|
||||
@ -770,9 +771,9 @@ TEST_F(GTPinTests, givenInitializedGTPinInterfaceWhenKernelIsExecutedThenGTPinCa
|
||||
auto buff10 = clCreateBuffer(context, 0, n * sizeof(unsigned int), nullptr, nullptr);
|
||||
auto buff11 = clCreateBuffer(context, 0, n * sizeof(unsigned int), nullptr, nullptr);
|
||||
|
||||
retVal = clSetKernelArg(pKernel1, 0, sizeof(cl_mem), &buff10);
|
||||
retVal = clSetKernelArg(pMultiDeviceKernel1, 0, sizeof(cl_mem), &buff10);
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
retVal = clSetKernelArg(pKernel1, 1, sizeof(cl_mem), &buff11);
|
||||
retVal = clSetKernelArg(pMultiDeviceKernel1, 1, sizeof(cl_mem), &buff11);
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
|
||||
int prevCount12 = KernelSubmitCallbackCount;
|
||||
@ -782,7 +783,7 @@ TEST_F(GTPinTests, givenInitializedGTPinInterfaceWhenKernelIsExecutedThenGTPinCa
|
||||
size_t globalWorkOffset[3] = {0, 0, 0};
|
||||
size_t globalWorkSize[3] = {n, 1, 1};
|
||||
size_t localWorkSize[3] = {1, 1, 1};
|
||||
retVal = clEnqueueNDRangeKernel(cmdQ, pKernel1, workDim, globalWorkOffset, globalWorkSize, localWorkSize, 0, nullptr, nullptr);
|
||||
retVal = clEnqueueNDRangeKernel(cmdQ, pMultiDeviceKernel1, workDim, globalWorkOffset, globalWorkSize, localWorkSize, 0, nullptr, nullptr);
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
EXPECT_EQ(prevCount12 + 1, KernelSubmitCallbackCount);
|
||||
EXPECT_EQ(prevCount13 + 1, CommandBufferCreateCallbackCount);
|
||||
@ -795,7 +796,8 @@ TEST_F(GTPinTests, givenInitializedGTPinInterfaceWhenKernelIsExecutedThenGTPinCa
|
||||
// Verify that GT-Pin Kernel Create callback is not called multiple times for the same kernel
|
||||
EXPECT_EQ(prevCount21, KernelCreateCallbackCount);
|
||||
|
||||
Kernel *pKernel2 = (Kernel *)kernel2;
|
||||
MultiDeviceKernel *pMultiDeviceKernel2 = static_cast<MultiDeviceKernel *>(kernel2);
|
||||
Kernel *pKernel2 = pMultiDeviceKernel2->getKernel(rootDeviceIndex);
|
||||
const KernelInfo &kInfo2 = pKernel2->getKernelInfo(rootDeviceIndex);
|
||||
uint64_t gtpinKernelId2 = pKernel2->getKernelId(rootDeviceIndex);
|
||||
EXPECT_EQ(kInfo2.shaderHashCode, gtpinKernelId2);
|
||||
@ -803,15 +805,15 @@ TEST_F(GTPinTests, givenInitializedGTPinInterfaceWhenKernelIsExecutedThenGTPinCa
|
||||
auto buff20 = clCreateBuffer(context, 0, n * sizeof(unsigned int), nullptr, nullptr);
|
||||
auto buff21 = clCreateBuffer(context, 0, n * sizeof(unsigned int), nullptr, nullptr);
|
||||
|
||||
retVal = clSetKernelArg(pKernel2, 0, sizeof(cl_mem), &buff20);
|
||||
retVal = clSetKernelArg(pMultiDeviceKernel2, 0, sizeof(cl_mem), &buff20);
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
retVal = clSetKernelArg(pKernel2, 1, sizeof(cl_mem), &buff21);
|
||||
retVal = clSetKernelArg(pMultiDeviceKernel2, 1, sizeof(cl_mem), &buff21);
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
|
||||
int prevCount22 = KernelSubmitCallbackCount;
|
||||
int prevCount23 = CommandBufferCreateCallbackCount;
|
||||
int prevCount24 = CommandBufferCompleteCallbackCount;
|
||||
retVal = clEnqueueNDRangeKernel(cmdQ, pKernel2, workDim, globalWorkOffset, globalWorkSize, localWorkSize, 0, nullptr, nullptr);
|
||||
retVal = clEnqueueNDRangeKernel(cmdQ, pMultiDeviceKernel2, workDim, globalWorkOffset, globalWorkSize, localWorkSize, 0, nullptr, nullptr);
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
EXPECT_EQ(prevCount22 + 1, KernelSubmitCallbackCount);
|
||||
EXPECT_EQ(prevCount23 + 1, CommandBufferCreateCallbackCount);
|
||||
@ -909,7 +911,8 @@ TEST_F(GTPinTests, givenInitializedGTPinInterfaceWhenKernelINTELIsExecutedThenGT
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
EXPECT_EQ(prevCount11 + 1, KernelCreateCallbackCount);
|
||||
|
||||
Kernel *pKernel1 = (Kernel *)kernel1;
|
||||
MultiDeviceKernel *pMultiDeviceKernel1 = static_cast<MultiDeviceKernel *>(kernel1);
|
||||
Kernel *pKernel1 = pMultiDeviceKernel1->getKernel(rootDeviceIndex);
|
||||
const KernelInfo &kInfo1 = pKernel1->getKernelInfo(rootDeviceIndex);
|
||||
uint64_t gtpinKernelId1 = pKernel1->getKernelId(rootDeviceIndex);
|
||||
EXPECT_EQ(kInfo1.shaderHashCode, gtpinKernelId1);
|
||||
@ -926,9 +929,9 @@ TEST_F(GTPinTests, givenInitializedGTPinInterfaceWhenKernelINTELIsExecutedThenGT
|
||||
auto buff10 = clCreateBuffer(context, 0, n * sizeof(unsigned int), nullptr, nullptr);
|
||||
auto buff11 = clCreateBuffer(context, 0, n * sizeof(unsigned int), nullptr, nullptr);
|
||||
|
||||
retVal = clSetKernelArg(pKernel1, 0, sizeof(cl_mem), &buff10);
|
||||
retVal = clSetKernelArg(pMultiDeviceKernel1, 0, sizeof(cl_mem), &buff10);
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
retVal = clSetKernelArg(pKernel1, 1, sizeof(cl_mem), &buff11);
|
||||
retVal = clSetKernelArg(pMultiDeviceKernel1, 1, sizeof(cl_mem), &buff11);
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
|
||||
int prevCount12 = KernelSubmitCallbackCount;
|
||||
@ -936,7 +939,7 @@ TEST_F(GTPinTests, givenInitializedGTPinInterfaceWhenKernelINTELIsExecutedThenGT
|
||||
int prevCount14 = CommandBufferCompleteCallbackCount;
|
||||
size_t globalWorkOffset[3] = {0, 0, 0};
|
||||
size_t workgroupCount[3] = {n, 1, 1};
|
||||
retVal = clEnqueueNDCountKernelINTEL(cmdQ, pKernel1, workDim, globalWorkOffset, workgroupCount, localWorkSize, 0, nullptr, nullptr);
|
||||
retVal = clEnqueueNDCountKernelINTEL(cmdQ, pMultiDeviceKernel1, workDim, globalWorkOffset, workgroupCount, localWorkSize, 0, nullptr, nullptr);
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
EXPECT_EQ(prevCount12 + 1, KernelSubmitCallbackCount);
|
||||
EXPECT_EQ(prevCount13 + 1, CommandBufferCreateCallbackCount);
|
||||
@ -949,7 +952,8 @@ TEST_F(GTPinTests, givenInitializedGTPinInterfaceWhenKernelINTELIsExecutedThenGT
|
||||
// Verify that GT-Pin Kernel Create callback is not called multiple times for the same kernel
|
||||
EXPECT_EQ(prevCount21, KernelCreateCallbackCount);
|
||||
|
||||
Kernel *pKernel2 = (Kernel *)kernel2;
|
||||
MultiDeviceKernel *pMultiDeviceKernel2 = static_cast<MultiDeviceKernel *>(kernel2);
|
||||
Kernel *pKernel2 = pMultiDeviceKernel2->getKernel(rootDeviceIndex);
|
||||
const KernelInfo &kInfo2 = pKernel2->getKernelInfo(rootDeviceIndex);
|
||||
uint64_t gtpinKernelId2 = pKernel2->getKernelId(rootDeviceIndex);
|
||||
EXPECT_EQ(kInfo2.shaderHashCode, gtpinKernelId2);
|
||||
@ -957,15 +961,15 @@ TEST_F(GTPinTests, givenInitializedGTPinInterfaceWhenKernelINTELIsExecutedThenGT
|
||||
auto buff20 = clCreateBuffer(context, 0, n * sizeof(unsigned int), nullptr, nullptr);
|
||||
auto buff21 = clCreateBuffer(context, 0, n * sizeof(unsigned int), nullptr, nullptr);
|
||||
|
||||
retVal = clSetKernelArg(pKernel2, 0, sizeof(cl_mem), &buff20);
|
||||
retVal = clSetKernelArg(pMultiDeviceKernel2, 0, sizeof(cl_mem), &buff20);
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
retVal = clSetKernelArg(pKernel2, 1, sizeof(cl_mem), &buff21);
|
||||
retVal = clSetKernelArg(pMultiDeviceKernel2, 1, sizeof(cl_mem), &buff21);
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
|
||||
int prevCount22 = KernelSubmitCallbackCount;
|
||||
int prevCount23 = CommandBufferCreateCallbackCount;
|
||||
int prevCount24 = CommandBufferCompleteCallbackCount;
|
||||
retVal = clEnqueueNDCountKernelINTEL(cmdQ, pKernel2, workDim, globalWorkOffset, workgroupCount, localWorkSize, 0, nullptr, nullptr);
|
||||
retVal = clEnqueueNDCountKernelINTEL(cmdQ, pMultiDeviceKernel2, workDim, globalWorkOffset, workgroupCount, localWorkSize, 0, nullptr, nullptr);
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
EXPECT_EQ(prevCount22 + 1, KernelSubmitCallbackCount);
|
||||
EXPECT_EQ(prevCount23 + 1, CommandBufferCreateCallbackCount);
|
||||
@ -1183,13 +1187,12 @@ HWCMDTEST_F(IGFX_GEN8_CORE, GTPinTests, givenInitializedGTPinInterfaceWhenKernel
|
||||
size_t localWorkSize[3] = {1, 1, 1};
|
||||
|
||||
MockParentKernel *parentKernel = MockParentKernel::create(*pContext);
|
||||
auto pMultiDeviceKernel = std::make_unique<MultiDeviceKernel>(parentKernel);
|
||||
|
||||
retVal = clEnqueueNDRangeKernel(cmdQ, parentKernel, workDim, globalWorkOffset, globalWorkSize, localWorkSize, 0, nullptr, nullptr);
|
||||
retVal = clEnqueueNDRangeKernel(cmdQ, pMultiDeviceKernel.get(), workDim, globalWorkOffset, globalWorkSize, localWorkSize, 0, nullptr, nullptr);
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
EXPECT_EQ(prevCount2, KernelSubmitCallbackCount);
|
||||
|
||||
delete parentKernel;
|
||||
|
||||
// Cleanup
|
||||
retVal = clReleaseKernel(kernel);
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
@ -1262,7 +1265,8 @@ TEST_F(GTPinTests, givenInitializedGTPinInterfaceWhenKernelWithoutSSHIsUsedThenG
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
EXPECT_EQ(prevCount1 + 1, KernelCreateCallbackCount);
|
||||
|
||||
Kernel *pKernel = (Kernel *)kernel;
|
||||
MultiDeviceKernel *pMultiDeviceKernel = static_cast<MultiDeviceKernel *>(kernel);
|
||||
Kernel *pKernel = pMultiDeviceKernel->getKernel(rootDeviceIndex);
|
||||
const KernelInfo &kInfo = pKernel->getKernelInfo(rootDeviceIndex);
|
||||
uint64_t gtpinKernelId = pKernel->getKernelId(rootDeviceIndex);
|
||||
EXPECT_EQ(kInfo.shaderHashCode, gtpinKernelId);
|
||||
@ -1271,9 +1275,9 @@ TEST_F(GTPinTests, givenInitializedGTPinInterfaceWhenKernelWithoutSSHIsUsedThenG
|
||||
auto buff0 = clCreateBuffer(context, 0, n * sizeof(unsigned int), nullptr, nullptr);
|
||||
auto buff1 = clCreateBuffer(context, 0, n * sizeof(unsigned int), nullptr, nullptr);
|
||||
|
||||
retVal = clSetKernelArg(pKernel, 0, sizeof(cl_mem), &buff0);
|
||||
retVal = clSetKernelArg(pMultiDeviceKernel, 0, sizeof(cl_mem), &buff0);
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
retVal = clSetKernelArg(pKernel, 1, sizeof(cl_mem), &buff1);
|
||||
retVal = clSetKernelArg(pMultiDeviceKernel, 1, sizeof(cl_mem), &buff1);
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
|
||||
// Verify that when SSH is removed then during kernel execution
|
||||
@ -1287,7 +1291,7 @@ TEST_F(GTPinTests, givenInitializedGTPinInterfaceWhenKernelWithoutSSHIsUsedThenG
|
||||
size_t globalWorkOffset[3] = {0, 0, 0};
|
||||
size_t globalWorkSize[3] = {n, 1, 1};
|
||||
size_t localWorkSize[3] = {1, 1, 1};
|
||||
retVal = clEnqueueNDRangeKernel(cmdQ, pKernel, workDim, globalWorkOffset, globalWorkSize, localWorkSize, 0, nullptr, nullptr);
|
||||
retVal = clEnqueueNDRangeKernel(cmdQ, pMultiDeviceKernel, workDim, globalWorkOffset, globalWorkSize, localWorkSize, 0, nullptr, nullptr);
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
EXPECT_EQ(prevCount2, KernelSubmitCallbackCount);
|
||||
EXPECT_EQ(prevCount3, CommandBufferCreateCallbackCount);
|
||||
@ -1375,7 +1379,8 @@ TEST_F(GTPinTests, givenInitializedGTPinInterfaceWhenBlockedKernelWithoutSSHIsUs
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
EXPECT_EQ(prevCount1 + 1, KernelCreateCallbackCount);
|
||||
|
||||
Kernel *pKernel = (Kernel *)kernel;
|
||||
MultiDeviceKernel *pMultiDeviceKernel = static_cast<MultiDeviceKernel *>(kernel);
|
||||
Kernel *pKernel = pMultiDeviceKernel->getKernel(rootDeviceIndex);
|
||||
const KernelInfo &kInfo = pKernel->getKernelInfo(rootDeviceIndex);
|
||||
uint64_t gtpinKernelId = pKernel->getKernelId(rootDeviceIndex);
|
||||
EXPECT_EQ(kInfo.shaderHashCode, gtpinKernelId);
|
||||
@ -1384,9 +1389,9 @@ TEST_F(GTPinTests, givenInitializedGTPinInterfaceWhenBlockedKernelWithoutSSHIsUs
|
||||
auto buff0 = clCreateBuffer(context, 0, n * sizeof(unsigned int), nullptr, nullptr);
|
||||
auto buff1 = clCreateBuffer(context, 0, n * sizeof(unsigned int), nullptr, nullptr);
|
||||
|
||||
retVal = clSetKernelArg(pKernel, 0, sizeof(cl_mem), &buff0);
|
||||
retVal = clSetKernelArg(pMultiDeviceKernel, 0, sizeof(cl_mem), &buff0);
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
retVal = clSetKernelArg(pKernel, 1, sizeof(cl_mem), &buff1);
|
||||
retVal = clSetKernelArg(pMultiDeviceKernel, 1, sizeof(cl_mem), &buff1);
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
|
||||
// Verify that when SSH is removed then during kernel execution
|
||||
@ -1403,7 +1408,7 @@ TEST_F(GTPinTests, givenInitializedGTPinInterfaceWhenBlockedKernelWithoutSSHIsUs
|
||||
size_t globalWorkOffset[3] = {0, 0, 0};
|
||||
size_t globalWorkSize[3] = {n, 1, 1};
|
||||
size_t localWorkSize[3] = {1, 1, 1};
|
||||
retVal = clEnqueueNDRangeKernel(cmdQ, pKernel, workDim, globalWorkOffset, globalWorkSize, localWorkSize, 1, &userEvent, nullptr);
|
||||
retVal = clEnqueueNDRangeKernel(cmdQ, pMultiDeviceKernel, workDim, globalWorkOffset, globalWorkSize, localWorkSize, 1, &userEvent, nullptr);
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
EXPECT_EQ(prevCount2, KernelSubmitCallbackCount);
|
||||
EXPECT_EQ(prevCount3, CommandBufferCreateCallbackCount);
|
||||
@ -1499,7 +1504,8 @@ TEST_F(GTPinTests, givenInitializedGTPinInterfaceWhenTheSameKerneIsExecutedTwice
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
EXPECT_EQ(prevCount11 + 1, KernelCreateCallbackCount);
|
||||
|
||||
Kernel *pKernel1 = (Kernel *)kernel1;
|
||||
MultiDeviceKernel *pMultiDeviceKernel1 = static_cast<MultiDeviceKernel *>(kernel1);
|
||||
Kernel *pKernel1 = pMultiDeviceKernel1->getKernel(rootDeviceIndex);
|
||||
const KernelInfo &kInfo1 = pKernel1->getKernelInfo(rootDeviceIndex);
|
||||
uint64_t gtpinKernelId1 = pKernel1->getKernelId(rootDeviceIndex);
|
||||
EXPECT_EQ(kInfo1.shaderHashCode, gtpinKernelId1);
|
||||
@ -1508,9 +1514,9 @@ TEST_F(GTPinTests, givenInitializedGTPinInterfaceWhenTheSameKerneIsExecutedTwice
|
||||
auto buff10 = clCreateBuffer(context, 0, n * sizeof(unsigned int), nullptr, nullptr);
|
||||
auto buff11 = clCreateBuffer(context, 0, n * sizeof(unsigned int), nullptr, nullptr);
|
||||
|
||||
retVal = clSetKernelArg(pKernel1, 0, sizeof(cl_mem), &buff10);
|
||||
retVal = clSetKernelArg(pMultiDeviceKernel1, 0, sizeof(cl_mem), &buff10);
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
retVal = clSetKernelArg(pKernel1, 1, sizeof(cl_mem), &buff11);
|
||||
retVal = clSetKernelArg(pMultiDeviceKernel1, 1, sizeof(cl_mem), &buff11);
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
|
||||
cl_event userEvent = clCreateUserEvent(context, &retVal);
|
||||
@ -1523,7 +1529,7 @@ TEST_F(GTPinTests, givenInitializedGTPinInterfaceWhenTheSameKerneIsExecutedTwice
|
||||
size_t globalWorkOffset[3] = {0, 0, 0};
|
||||
size_t globalWorkSize[3] = {n, 1, 1};
|
||||
size_t localWorkSize[3] = {1, 1, 1};
|
||||
retVal = clEnqueueNDRangeKernel(cmdQ, pKernel1, workDim, globalWorkOffset, globalWorkSize, localWorkSize, 1, &userEvent, nullptr);
|
||||
retVal = clEnqueueNDRangeKernel(cmdQ, pMultiDeviceKernel1, workDim, globalWorkOffset, globalWorkSize, localWorkSize, 1, &userEvent, nullptr);
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
EXPECT_EQ(prevCount12 + 1, KernelSubmitCallbackCount);
|
||||
EXPECT_EQ(prevCount13 + 1, CommandBufferCreateCallbackCount);
|
||||
@ -1537,7 +1543,8 @@ TEST_F(GTPinTests, givenInitializedGTPinInterfaceWhenTheSameKerneIsExecutedTwice
|
||||
// Verify that Kernel Create callback was not called now
|
||||
EXPECT_EQ(prevCount21, KernelCreateCallbackCount);
|
||||
|
||||
Kernel *pKernel2 = (Kernel *)kernel2;
|
||||
MultiDeviceKernel *pMultiDeviceKernel2 = static_cast<MultiDeviceKernel *>(kernel2);
|
||||
Kernel *pKernel2 = pMultiDeviceKernel2->getKernel(rootDeviceIndex);
|
||||
const KernelInfo &kInfo2 = pKernel2->getKernelInfo(rootDeviceIndex);
|
||||
uint64_t gtpinKernelId2 = pKernel2->getKernelId(rootDeviceIndex);
|
||||
EXPECT_EQ(kInfo2.shaderHashCode, gtpinKernelId2);
|
||||
@ -1545,16 +1552,16 @@ TEST_F(GTPinTests, givenInitializedGTPinInterfaceWhenTheSameKerneIsExecutedTwice
|
||||
auto buff20 = clCreateBuffer(context, 0, n * sizeof(unsigned int), nullptr, nullptr);
|
||||
auto buff21 = clCreateBuffer(context, 0, n * sizeof(unsigned int), nullptr, nullptr);
|
||||
|
||||
retVal = clSetKernelArg(pKernel2, 0, sizeof(cl_mem), &buff20);
|
||||
retVal = clSetKernelArg(pMultiDeviceKernel2, 0, sizeof(cl_mem), &buff20);
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
retVal = clSetKernelArg(pKernel2, 1, sizeof(cl_mem), &buff21);
|
||||
retVal = clSetKernelArg(pMultiDeviceKernel2, 1, sizeof(cl_mem), &buff21);
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
|
||||
int prevCount22 = KernelSubmitCallbackCount;
|
||||
int prevCount23 = CommandBufferCreateCallbackCount;
|
||||
int prevCount24 = CommandBufferCompleteCallbackCount;
|
||||
EXPECT_EQ(prevCount14, prevCount24);
|
||||
retVal = clEnqueueNDRangeKernel(cmdQ, pKernel2, workDim, globalWorkOffset, globalWorkSize, localWorkSize, 0, nullptr, nullptr);
|
||||
retVal = clEnqueueNDRangeKernel(cmdQ, pMultiDeviceKernel2, workDim, globalWorkOffset, globalWorkSize, localWorkSize, 0, nullptr, nullptr);
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
EXPECT_EQ(prevCount22 + 1, KernelSubmitCallbackCount);
|
||||
EXPECT_EQ(prevCount23 + 1, CommandBufferCreateCallbackCount);
|
||||
@ -1656,12 +1663,13 @@ TEST_F(GTPinTests, givenMultipleKernelSubmissionsWhenOneOfGtpinSurfacesIsNullThe
|
||||
EXPECT_NE(nullptr, kernel1);
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
|
||||
Kernel *pKernel1 = (Kernel *)kernel1;
|
||||
MultiDeviceKernel *pMultiDeviceKernel1 = static_cast<MultiDeviceKernel *>(kernel1);
|
||||
Kernel *pKernel1 = pMultiDeviceKernel1->getKernel(rootDeviceIndex);
|
||||
returnNullResource = true;
|
||||
|
||||
auto pCmdQueue = castToObject<CommandQueue>(cmdQ);
|
||||
|
||||
gtpinNotifyKernelSubmit(pKernel1, pCmdQueue);
|
||||
gtpinNotifyKernelSubmit(pMultiDeviceKernel1, pCmdQueue);
|
||||
EXPECT_EQ(nullptr, kernelExecQueue[0].gtpinResource);
|
||||
|
||||
CommandStreamReceiver &csr = pCmdQueue->getGpgpuCommandStreamReceiver();
|
||||
@ -1674,13 +1682,13 @@ TEST_F(GTPinTests, givenMultipleKernelSubmissionsWhenOneOfGtpinSurfacesIsNullThe
|
||||
|
||||
returnNullResource = false;
|
||||
|
||||
gtpinNotifyKernelSubmit(pKernel1, pCmdQueue);
|
||||
gtpinNotifyKernelSubmit(pMultiDeviceKernel1, pCmdQueue);
|
||||
EXPECT_NE(nullptr, kernelExecQueue[1].gtpinResource);
|
||||
gtpinNotifyMakeResident(pKernel1, &csr);
|
||||
EXPECT_TRUE(kernelExecQueue[1].isResourceResident);
|
||||
cl_mem gtpinBuffer1 = kernelExecQueue[1].gtpinResource;
|
||||
|
||||
gtpinNotifyKernelSubmit(pKernel1, pCmdQueue);
|
||||
gtpinNotifyKernelSubmit(pMultiDeviceKernel1, pCmdQueue);
|
||||
EXPECT_NE(nullptr, kernelExecQueue[2].gtpinResource);
|
||||
gtpinNotifyUpdateResidencyList(pKernel1, &residencyVector);
|
||||
EXPECT_EQ(1u, residencyVector.size());
|
||||
@ -1782,7 +1790,8 @@ TEST_F(GTPinTests, givenInitializedGTPinInterfaceWhenKernelIsCreatedThenAllKerne
|
||||
EXPECT_EQ(prevCount1 + 1, KernelCreateCallbackCount);
|
||||
|
||||
// Simulate that created kernel was sent for execution
|
||||
auto pKernel = castToObject<Kernel>(kernel);
|
||||
auto pMultiDeviceKernel = castToObject<MultiDeviceKernel>(kernel);
|
||||
auto pKernel = pMultiDeviceKernel->getKernel(rootDeviceIndex);
|
||||
auto pCmdQueue = castToObject<CommandQueue>(cmdQ);
|
||||
ASSERT_NE(nullptr, pKernel);
|
||||
EXPECT_EQ(0u, kernelExecQueue.size());
|
||||
@ -1824,7 +1833,8 @@ TEST_F(GTPinTests, givenInitializedGTPinInterfaceWhenKernelIsCreatedThenAllKerne
|
||||
ASSERT_NE(nullptr, kernel2);
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
// ... and simulate that it was sent for execution
|
||||
auto pKernel2 = castToObject<Kernel>(kernel2);
|
||||
auto pMultiDeviceKernel2 = castToObject<MultiDeviceKernel>(kernel2);
|
||||
auto pKernel2 = pMultiDeviceKernel2->getKernel(rootDeviceIndex);
|
||||
ASSERT_NE(nullptr, pKernel2);
|
||||
EXPECT_EQ(1u, kernelExecQueue.size());
|
||||
EXPECT_EQ(1u, kernelResources.size());
|
||||
@ -1975,7 +1985,8 @@ TEST_F(GTPinTests, givenInitializedGTPinInterfaceWhenOneKernelIsSubmittedSeveral
|
||||
EXPECT_EQ(prevCount1 + 1, KernelCreateCallbackCount);
|
||||
|
||||
// Simulate that created kernel was sent for execution two times in a row
|
||||
auto pKernel = castToObject<Kernel>(kernel);
|
||||
auto pMultiDeviceKernel = castToObject<MultiDeviceKernel>(kernel);
|
||||
auto pKernel = pMultiDeviceKernel->getKernel(rootDeviceIndex);
|
||||
auto pCmdQueue = castToObject<CommandQueue>(cmdQ);
|
||||
ASSERT_NE(nullptr, pKernel);
|
||||
EXPECT_EQ(0u, kernelExecQueue.size());
|
||||
@ -2219,7 +2230,8 @@ TEST_F(GTPinTests, givenKernelWithSSHThenVerifyThatSSHResizeWorksWell) {
|
||||
kernel = clCreateKernel(pProgram, "CopyBuffer", &retVal);
|
||||
ASSERT_NE(nullptr, kernel);
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
Kernel *pKernel = castToObject<Kernel>(kernel);
|
||||
auto pMultiDeviceKernel = castToObject<MultiDeviceKernel>(kernel);
|
||||
auto pKernel = pMultiDeviceKernel->getKernel(rootDeviceIndex);
|
||||
ASSERT_NE(nullptr, pKernel);
|
||||
|
||||
size_t numBTS1 = pKernel->getNumberOfBindingTableStates(rootDeviceIndex);
|
||||
@ -2319,7 +2331,8 @@ TEST_F(GTPinTests, givenKernelThenVerifyThatKernelCodeSubstitutionWorksWell) {
|
||||
kernel = clCreateKernel(pProgram, "CopyBuffer", &retVal);
|
||||
ASSERT_NE(nullptr, kernel);
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
Kernel *pKernel = castToObject<Kernel>(kernel);
|
||||
auto pMultiDeviceKernel = castToObject<MultiDeviceKernel>(kernel);
|
||||
auto pKernel = pMultiDeviceKernel->getKernel(rootDeviceIndex);
|
||||
ASSERT_NE(nullptr, pKernel);
|
||||
|
||||
bool isKernelCodeSubstituted = pKernel->isKernelHeapSubstituted(rootDeviceIndex);
|
||||
@ -2393,16 +2406,17 @@ TEST_F(GTPinTests, givenInitializedGTPinInterfaceWhenOnKernelSubitIsCalledThenCo
|
||||
|
||||
auto pProgramm = std::make_unique<MockProgram>(context.get(), false, toClDeviceVector(*pDevice));
|
||||
std::unique_ptr<MockCommandQueue> cmdQ(new MockCommandQueue(context.get(), pDevice, nullptr));
|
||||
std::unique_ptr<MockKernel> pKernel(new MockKernel(pProgramm.get(), MockKernel::toKernelInfoContainer(*pKernelInfo, rootDeviceIndex)));
|
||||
auto pKernel = new MockKernel(pProgramm.get(), MockKernel::toKernelInfoContainer(*pKernelInfo, rootDeviceIndex));
|
||||
auto pMultiDeviceKernel = std::make_unique<MultiDeviceKernel>(pKernel);
|
||||
|
||||
pKernel->setSshLocal(nullptr, sizeof(surfaceStateHeap), rootDeviceIndex);
|
||||
|
||||
kernelOffset = 0x1234;
|
||||
EXPECT_NE(pKernel->getStartOffset(), kernelOffset);
|
||||
returnNullResource = true;
|
||||
cl_context ctxt = (cl_context)((Context *)context.get());
|
||||
cl_context ctxt = context.get();
|
||||
currContext = (gtpin::context_handle_t)ctxt;
|
||||
gtpinNotifyKernelSubmit(pKernel.get(), cmdQ.get());
|
||||
gtpinNotifyKernelSubmit(pMultiDeviceKernel.get(), cmdQ.get());
|
||||
EXPECT_EQ(pKernel->getStartOffset(), kernelOffset);
|
||||
|
||||
EXPECT_EQ(CL_SUCCESS, retVal);
|
||||
@ -2467,7 +2481,7 @@ HWTEST_F(GTPinTests, givenGtPinInitializedWhenSubmittingKernelCommandThenFlushed
|
||||
bool slmUsed = false;
|
||||
bool ndRangeKernel = false;
|
||||
|
||||
gtpinNotifyKernelSubmit(kernel, mockCmdQ.get());
|
||||
gtpinNotifyKernelSubmit(kernel.mockMultiDeviceKernel, mockCmdQ.get());
|
||||
|
||||
std::unique_ptr<Command> command(new CommandComputeKernel(*mockCmdQ, kernelOperation, surfaces, flushDC, slmUsed, ndRangeKernel, nullptr, preemptionMode, kernel, 1));
|
||||
CompletionStamp stamp = command->submit(20, false);
|
||||
|
||||
Reference in New Issue
Block a user