From 738bf16008f140b18b5b1189b3671b6dd92b4523 Mon Sep 17 00:00:00 2001 From: amilanov-amd Date: Fri, 14 Nov 2025 13:00:30 +0100 Subject: [PATCH] [hip-tests] Tag multigpu tests with Catch2 tags (#1315) --- .../catch/unit/atomics/atomicAdd_system.cc | 15 +++++--- .../catch/unit/atomics/atomicAnd_system.cc | 15 +++++--- .../catch/unit/atomics/atomicCAS_system.cc | 15 +++++--- .../catch/unit/atomics/atomicExch_system.cc | 29 +++++++------- .../catch/unit/atomics/atomicMax_system.cc | 31 +++++++++------ .../catch/unit/atomics/atomicMin_system.cc | 31 +++++++++------ .../catch/unit/atomics/atomicOr_system.cc | 15 +++++--- .../catch/unit/atomics/atomicSub_system.cc | 15 +++++--- .../catch/unit/atomics/atomicXor_system.cc | 15 +++++--- .../unit/callback/hipGetStreamDeviceId.cc | 3 +- .../hipCGMultiGridGroupType_old.cc | 4 +- ...pLaunchCooperativeKernelMultiDevice_old.cc | 2 +- .../unit/cooperativeGrps/multi_grid_group.cc | 9 +++-- .../hipDeviceEnableDisablePeerAccess.cc | 6 +-- .../catch/unit/device/hipDeviceGetName.cc | 3 +- .../catch/unit/device/hipDeviceGetUuid.cc | 5 ++- .../catch/unit/device/hipDeviceSetGetLimit.cc | 2 +- .../unit/device/hipDeviceSetLimit_old.cc | 2 +- .../catch/unit/device/hipDeviceTotalMem.cc | 2 +- .../unit/device/hipGetDeviceProperties.cc | 2 +- .../unit/device/hipGetProcAddressDevMgmt.cc | 4 +- .../catch/unit/device/hipSetGetDevice.cc | 6 +-- .../catch/unit/device/hipSetValidDevices.cc | 2 +- .../unit/deviceLib/threadfence_system.cc | 2 +- .../unit/errorHandling/hipExtGetLastError.cc | 2 +- .../unit/errorHandling/hipGetLastError.cc | 2 +- .../errorHandling/hipGetLastErrorOnAbort.cc | 2 +- .../unit/event/Unit_hipEventElapsedTime.cc | 2 +- .../unit/event/Unit_hipEventMGpuMThreads.cc | 4 +- .../catch/unit/event/Unit_hipEventQuery.cc | 2 +- .../catch/unit/event/Unit_hipEventRecord.cc | 2 +- .../hipExtLaunchMultiKernelMultiDevice.cc | 6 ++- .../hipLaunchCooperativeKernelMultiDevice.cc | 6 ++- .../graph/hipDeviceGetGraphMemAttribute.cc | 3 +- .../unit/graph/hipDrvGraphAddMemcpyNode.cc | 2 +- .../unit/graph/hipGraphAddChildGraphNode.cc | 2 +- .../unit/graph/hipGraphAddMemAllocNode.cc | 6 +-- .../unit/graph/hipGraphAddMemcpyNode1D_old.cc | 2 +- .../hipGraphAddMemcpyNodeFromSymbol_old.cc | 6 ++- .../hipGraphAddMemcpyNodeToSymbol_old.cc | 6 ++- .../unit/graph/hipGraphAddMemcpyNode_old.cc | 2 +- .../catch/unit/graph/hipGraphClone.cc | 2 +- .../catch/unit/graph/hipGraphCloneComplx.cc | 6 +-- .../hipGraphExecEventRecordNodeSetEvent.cc | 3 +- ...hipGraphExecMemcpyNodeSetParamsToSymbol.cc | 3 +- .../graph/hipGraphExecMemsetNodeSetParams.cc | 3 +- .../catch/unit/graph/hipGraphExecUpdate.cc | 3 +- .../graph/hipGraphInstantiateWithFlags.cc | 6 ++- .../catch/unit/graph/hipGraphLaunch_old.cc | 2 +- .../graph/hipGraphMemAllocNodeGetParams.cc | 3 +- .../catch/unit/graph/hipGraphPerf.cc | 38 ++++++++++++------- .../catch/unit/graph/hipGraphUpload.cc | 2 +- .../catch/unit/graph/hipStreamBeginCapture.cc | 2 +- .../unit/graph/hipStreamBeginCapture_old.cc | 2 +- .../catch/unit/kernel/hipPrintfKernel.cc | 2 +- .../catch/unit/memory/hipArrayCreate.cc | 2 +- .../unit/memory/hipArrayGetDescriptor.cc | 7 ++-- .../catch/unit/memory/hipDeviceGetMemPool.cc | 2 +- .../catch/unit/memory/hipDeviceSetMemPool.cc | 2 +- .../unit/memory/hipDrvMemcpy2DUnaligned.cc | 2 +- .../unit/memory/hipDrvMemcpy3DAsync_old.cc | 8 ++-- .../catch/unit/memory/hipDrvMemcpy3D_old.cc | 6 +-- .../memory/hipGetProcAddressMemoryApis.cc | 4 +- .../catch/unit/memory/hipHostRegister.cc | 6 ++- .../catch/unit/memory/hipMalloc3D.cc | 2 +- .../catch/unit/memory/hipMalloc3DArray.cc | 2 +- .../catch/unit/memory/hipMallocArray.cc | 2 +- .../catch/unit/memory/hipMallocAsync.cc | 6 +-- .../catch/unit/memory/hipMallocConcurrency.cc | 2 +- .../unit/memory/hipMallocFromPoolAsync.cc | 6 +-- .../unit/memory/hipMallocManagedFlagsTst.cc | 4 +- .../memory/hipMallocManaged_MultiScenario.cc | 9 +++-- .../unit/memory/hipMallocMipmappedArray.cc | 2 +- .../unit/memory/hipMallocMngdMultiThread.cc | 4 +- .../catch/unit/memory/hipMallocPitch.cc | 2 +- .../catch/unit/memory/hipMemAdvise.cc | 2 +- .../catch/unit/memory/hipMemAdvise_old.cc | 8 ++-- .../catch/unit/memory/hipMemAdvise_v2.cc | 2 +- .../catch/unit/memory/hipMemAllocHost.cc | 2 +- .../catch/unit/memory/hipMemPoolCreate.cc | 2 +- .../unit/memory/hipMemPoolSetGetAccess.cc | 6 +-- .../catch/unit/memory/hipMemPrefetchAsync.cc | 2 +- .../unit/memory/hipMemPrefetchAsync_v2.cc | 2 +- .../catch/unit/memory/hipMemcpy2D.cc | 2 +- .../catch/unit/memory/hipMemcpy2DAsync.cc | 2 +- .../catch/unit/memory/hipMemcpy2DAsync_old.cc | 8 ++-- .../catch/unit/memory/hipMemcpy2DFromArray.cc | 3 +- .../unit/memory/hipMemcpy2DFromArrayAsync.cc | 2 +- .../memory/hipMemcpy2DFromArrayAsync_old.cc | 6 ++- .../unit/memory/hipMemcpy2DFromArray_old.cc | 5 ++- .../catch/unit/memory/hipMemcpy2DToArray.cc | 3 +- .../unit/memory/hipMemcpy2DToArrayAsync.cc | 3 +- .../memory/hipMemcpy2DToArrayAsync_old.cc | 6 ++- .../unit/memory/hipMemcpy2DToArray_old.cc | 5 ++- .../catch/unit/memory/hipMemcpy2D_old.cc | 5 ++- .../catch/unit/memory/hipMemcpy3DAsync_old.cc | 4 +- .../catch/unit/memory/hipMemcpy3D_old.cc | 4 +- .../catch/unit/memory/hipMemcpyAsync_old.cc | 6 ++- .../catch/unit/memory/hipMemcpyAtoH_old.cc | 4 +- .../catch/unit/memory/hipMemcpyDtoD.cc | 3 +- .../catch/unit/memory/hipMemcpyDtoDAsync.cc | 3 +- .../catch/unit/memory/hipMemcpyHtoAAsync.cc | 2 +- .../catch/unit/memory/hipMemcpyHtoA_old.cc | 4 +- .../catch/unit/memory/hipMemcpyParam2D.cc | 2 +- .../unit/memory/hipMemcpyParam2DAsync.cc | 2 +- .../unit/memory/hipMemcpyParam2DAsync_old.cc | 13 ++++--- .../catch/unit/memory/hipMemcpyParam2D_old.cc | 5 ++- .../catch/unit/memory/hipMemcpyPeer.cc | 9 +++-- .../catch/unit/memory/hipMemcpyPeerAsync.cc | 9 +++-- .../unit/memory/hipMemcpyPeerAsync_old.cc | 6 +-- .../catch/unit/memory/hipMemcpyPeer_old.cc | 4 +- .../memory/hipMemcpyWithStreamMultiThread.cc | 2 +- .../unit/memory/hipMemcpyWithStream_old.cc | 13 +++++-- .../catch/unit/memory/hipMemcpy_EdgeCases.cc | 6 ++- .../unit/memory/hipMemcpy_MultiThread.cc | 2 +- .../memory/hipMemset3DRegressMultiThread.cc | 4 +- .../unit/memory/hipPointerGetAttribute.cc | 2 +- .../unit/memory/hipPointerGetAttributes.cc | 14 ++++--- .../catch/unit/memory/hipPtrGetAttribute.cc | 2 +- .../unit/memory/hipSVMTestByteGranularity.cpp | 2 +- .../hipSVMTestFineGrainMemoryConsistency.cpp | 2 +- .../hipSVMTestSharedAddressSpaceFineGrain.cpp | 4 +- .../unit/memory/hipStreamAttachMemAsync.cc | 2 +- .../hipExtLaunchMultiKernelMultiDevice.cc | 2 +- .../unit/module/hipExtModuleLaunchKernel.cc | 2 +- .../catch/unit/module/hipGetFuncBySymbol.cc | 4 +- .../module/hipGetProcAddressModuleApis.cc | 2 +- .../catch/unit/module/hipManagedKeyword.cc | 2 +- .../catch/unit/module/hipModuleGetFunction.cc | 2 +- .../catch/unit/module/hipModuleGetGlobal.cc | 2 +- ...oduleLaunchCooperativeKernelMultiDevice.cc | 11 ++++-- .../unit/module/hipModuleLaunchKernel.cc | 2 +- .../hipModuleLoadMultProcessOnMultGPU.cc | 2 +- ...PotentialBlockSizeVariableSMemWithFlags.cc | 3 +- .../catch/unit/printf/hipPrintfManyDevices.cc | 2 +- .../catch/unit/stream/hipLaunchHostFunc.cc | 2 +- .../catch/unit/stream/hipStreamGetDevice.cc | 6 +-- .../catch/unit/stream/hipStreamGetId.cc | 2 +- .../catch/unit/stream/hipStreamLegacy_Ext.cc | 9 +++-- .../cache_coherency_gpu_gpu.cc | 2 +- .../catch/unit/threadfence/__threadfence.cc | 2 +- .../unit/threadfence/__threadfence_block.cc | 2 +- .../unit/threadfence/__threadfence_system.cc | 2 +- .../hipMemGetHandleForAddressRange.cc | 6 ++- .../unit/virtualMemoryManagement/hipMemMap.cc | 4 +- .../hipMemSetGetAccess.cc | 8 ++-- 146 files changed, 425 insertions(+), 316 deletions(-) diff --git a/projects/hip-tests/catch/unit/atomics/atomicAdd_system.cc b/projects/hip-tests/catch/unit/atomics/atomicAdd_system.cc index e42a0edbaf..0e2a023b3b 100644 --- a/projects/hip-tests/catch/unit/atomics/atomicAdd_system.cc +++ b/projects/hip-tests/catch/unit/atomics/atomicAdd_system.cc @@ -54,8 +54,9 @@ THE SOFTWARE. * ------------------------ * - HIP_VERSION >= 5.2 */ -TEMPLATE_TEST_CASE("Unit_atomicAdd_system_Positive_Peer_GPUs", "", int, unsigned int, unsigned long, - unsigned long long, float, double) { +TEMPLATE_TEST_CASE("Unit_atomicAdd_system_Positive_Peer_GPUs", "[multigpu]", + int, unsigned int, unsigned long, unsigned long long, float, + double) { int warp_size = 0; HIP_CHECK(hipDeviceGetAttribute(&warp_size, hipDeviceAttributeWarpSize, 0)); const auto cache_line_size = 128u; @@ -103,8 +104,9 @@ TEMPLATE_TEST_CASE("Unit_atomicAdd_system_Positive_Peer_GPUs", "", int, unsigned * ------------------------ * - HIP_VERSION >= 5.2 */ -TEMPLATE_TEST_CASE("Unit_atomicAdd_system_Positive_Host_And_GPU", "", int, unsigned int, - unsigned long, unsigned long long, float, double) { +TEMPLATE_TEST_CASE("Unit_atomicAdd_system_Positive_Host_And_GPU", "[multigpu]", + int, unsigned int, unsigned long, unsigned long long, float, + double) { int warp_size = 0; HIP_CHECK(hipDeviceGetAttribute(&warp_size, hipDeviceAttributeWarpSize, 0)); const auto cache_line_size = 128u; @@ -152,8 +154,9 @@ TEMPLATE_TEST_CASE("Unit_atomicAdd_system_Positive_Host_And_GPU", "", int, unsig * ------------------------ * - HIP_VERSION >= 5.2 */ -TEMPLATE_TEST_CASE("Unit_atomicAdd_system_Positive_Host_And_Peer_GPUs", "", int, unsigned int, - unsigned long, unsigned long long, float, double) { +TEMPLATE_TEST_CASE("Unit_atomicAdd_system_Positive_Host_And_Peer_GPUs", + "[multigpu]", int, unsigned int, unsigned long, + unsigned long long, float, double) { int warp_size = 0; HIP_CHECK(hipDeviceGetAttribute(&warp_size, hipDeviceAttributeWarpSize, 0)); const auto cache_line_size = 128u; diff --git a/projects/hip-tests/catch/unit/atomics/atomicAnd_system.cc b/projects/hip-tests/catch/unit/atomics/atomicAnd_system.cc index a42eb974a8..6b70ca050e 100644 --- a/projects/hip-tests/catch/unit/atomics/atomicAnd_system.cc +++ b/projects/hip-tests/catch/unit/atomics/atomicAnd_system.cc @@ -45,8 +45,9 @@ THE SOFTWARE. * - Multi-device * - HIP_VERSION >= 5.2 */ -TEMPLATE_TEST_CASE("Unit_atomicAnd_system_Positive_Peer_GPUs_Same_Address", "", int, unsigned int, - unsigned long, unsigned long long) { +TEMPLATE_TEST_CASE("Unit_atomicAnd_system_Positive_Peer_GPUs_Same_Address", + "[multigpu]", int, unsigned int, unsigned long, + unsigned long long) { for (auto current = 0; current < 1; ++current) { DYNAMIC_SECTION("Same address " << current) { Bitwise::MultipleDeviceMultipleKernelTest( @@ -68,8 +69,9 @@ TEMPLATE_TEST_CASE("Unit_atomicAnd_system_Positive_Peer_GPUs_Same_Address", "", * - Multi-device * - HIP_VERSION >= 5.2 */ -TEMPLATE_TEST_CASE("Unit_atomicAnd_system_Positive_Peer_GPUs_Adjacent_Addresses", "", int, - unsigned int, unsigned long, unsigned long long) { +TEMPLATE_TEST_CASE( + "Unit_atomicAnd_system_Positive_Peer_GPUs_Adjacent_Addresses", "[multigpu]", + int, unsigned int, unsigned long, unsigned long long) { int warp_size = 0; HIP_CHECK(hipDeviceGetAttribute(&warp_size, hipDeviceAttributeWarpSize, 0)); @@ -94,8 +96,9 @@ TEMPLATE_TEST_CASE("Unit_atomicAnd_system_Positive_Peer_GPUs_Adjacent_Addresses" * - Multi-device * - HIP_VERSION >= 5.2 */ -TEMPLATE_TEST_CASE("Unit_atomicAnd_system_Positive_Peer_GPUs_Scattered_Addresses", "", int, - unsigned int, unsigned long, unsigned long long) { +TEMPLATE_TEST_CASE( + "Unit_atomicAnd_system_Positive_Peer_GPUs_Scattered_Addresses", + "[multigpu]", int, unsigned int, unsigned long, unsigned long long) { int warp_size = 0; HIP_CHECK(hipDeviceGetAttribute(&warp_size, hipDeviceAttributeWarpSize, 0)); const auto cache_line_size = 128u; diff --git a/projects/hip-tests/catch/unit/atomics/atomicCAS_system.cc b/projects/hip-tests/catch/unit/atomics/atomicCAS_system.cc index 9bb84f672d..ea628b01c3 100644 --- a/projects/hip-tests/catch/unit/atomics/atomicCAS_system.cc +++ b/projects/hip-tests/catch/unit/atomics/atomicCAS_system.cc @@ -60,8 +60,9 @@ THE SOFTWARE. * ------------------------ * - HIP_VERSION >= 5.2 */ -TEMPLATE_TEST_CASE("Unit_atomicCAS_system_Positive_Peer_GPUs", "", int, unsigned int, - unsigned long long, unsigned short int TYPES) { +TEMPLATE_TEST_CASE("Unit_atomicCAS_system_Positive_Peer_GPUs", "[multigpu]", + int, unsigned int, unsigned long long, + unsigned short int TYPES) { int warp_size = 0; HIP_CHECK(hipDeviceGetAttribute(&warp_size, hipDeviceAttributeWarpSize, 0)); const auto cache_line_size = 128u; @@ -105,8 +106,9 @@ TEMPLATE_TEST_CASE("Unit_atomicCAS_system_Positive_Peer_GPUs", "", int, unsigned * ------------------------ * - HIP_VERSION >= 5.2 */ -TEMPLATE_TEST_CASE("Unit_atomicCAS_system_Positive_Host_And_GPU", "", int, unsigned int, - unsigned long long, unsigned short int TYPES) { +TEMPLATE_TEST_CASE("Unit_atomicCAS_system_Positive_Host_And_GPU", "[multigpu]", + int, unsigned int, unsigned long long, + unsigned short int TYPES) { int warp_size = 0; HIP_CHECK(hipDeviceGetAttribute(&warp_size, hipDeviceAttributeWarpSize, 0)); const auto cache_line_size = 128u; @@ -155,8 +157,9 @@ TEMPLATE_TEST_CASE("Unit_atomicCAS_system_Positive_Host_And_GPU", "", int, unsig * ------------------------ * - HIP_VERSION >= 5.2 */ -TEMPLATE_TEST_CASE("Unit_atomicCAS_system_Positive_Host_And_Peer_GPUs", "", int, unsigned int, - unsigned long long, unsigned short int TYPES) { +TEMPLATE_TEST_CASE("Unit_atomicCAS_system_Positive_Host_And_Peer_GPUs", + "[multigpu]", int, unsigned int, unsigned long long, + unsigned short int TYPES) { int warp_size = 0; HIP_CHECK(hipDeviceGetAttribute(&warp_size, hipDeviceAttributeWarpSize, 0)); const auto cache_line_size = 128u; diff --git a/projects/hip-tests/catch/unit/atomics/atomicExch_system.cc b/projects/hip-tests/catch/unit/atomics/atomicExch_system.cc index eeb19c3661..1bda154dfe 100644 --- a/projects/hip-tests/catch/unit/atomics/atomicExch_system.cc +++ b/projects/hip-tests/catch/unit/atomics/atomicExch_system.cc @@ -55,11 +55,12 @@ THE SOFTWARE. * - HIP_VERSION >= 5.2 */ #if HT_NVIDIA -TEMPLATE_TEST_CASE("Unit_atomicExch_system_Positive_Peer_GPUs", "", int, unsigned int, - unsigned long long, float) { +TEMPLATE_TEST_CASE("Unit_atomicExch_system_Positive_Peer_GPUs", "[multigpu]", + int, unsigned int, unsigned long long, float) { #else -TEMPLATE_TEST_CASE("Unit_atomicExch_system_Positive_Peer_GPUs", "", int, unsigned int, - unsigned long, unsigned long long, float, double) { +TEMPLATE_TEST_CASE("Unit_atomicExch_system_Positive_Peer_GPUs", "[multigpu]", + int, unsigned int, unsigned long, unsigned long long, float, + double) { #endif // HT_NVIDIA int warp_size = 0; HIP_CHECK(hipDeviceGetAttribute(&warp_size, hipDeviceAttributeWarpSize, 0)); @@ -109,12 +110,13 @@ TEMPLATE_TEST_CASE("Unit_atomicExch_system_Positive_Peer_GPUs", "", int, unsigne * - HIP_VERSION >= 5.2 */ #if HT_NVIDIA -TEMPLATE_TEST_CASE("Unit_atomicExch_system_Positive_Host_And_GPU", "", int, unsigned int, - unsigned long long, float) { +TEMPLATE_TEST_CASE("Unit_atomicExch_system_Positive_Host_And_GPU", "[multigpu]", + int, unsigned int, unsigned long long, float) { #else -TEMPLATE_TEST_CASE("Unit_atomicExch_system_Positive_Host_And_GPU", "", int, unsigned int, - unsigned long, unsigned long long, float, double) { -#endif // HT_NVIDIA +TEMPLATE_TEST_CASE("Unit_atomicExch_system_Positive_Host_And_GPU", "[multigpu]", + int, unsigned int, unsigned long, unsigned long long, float, + double) { +#endif // HT_NVIDIA int warp_size = 0; HIP_CHECK(hipDeviceGetAttribute(&warp_size, hipDeviceAttributeWarpSize, 0)); const auto cache_line_size = 128u; @@ -164,11 +166,12 @@ TEMPLATE_TEST_CASE("Unit_atomicExch_system_Positive_Host_And_GPU", "", int, unsi * - HIP_VERSION >= 5.2 */ #if HT_NVIDIA -TEMPLATE_TEST_CASE("Unit_atomicExch_system_Positive_Host_And_Peer_GPUs", "", int, unsigned int, - unsigned long long, float) { +TEMPLATE_TEST_CASE("Unit_atomicExch_system_Positive_Host_And_Peer_GPUs", + "[multigpu]", int, unsigned int, unsigned long long, float) { #else -TEMPLATE_TEST_CASE("Unit_atomicExch_system_Positive_Host_And_Peer_GPUs", "", int, unsigned int, - unsigned long, unsigned long long, float, double) { +TEMPLATE_TEST_CASE("Unit_atomicExch_system_Positive_Host_And_Peer_GPUs", + "[multigpu]", int, unsigned int, unsigned long, + unsigned long long, float, double) { #endif // HT_NVIDIA int warp_size = 0; HIP_CHECK(hipDeviceGetAttribute(&warp_size, hipDeviceAttributeWarpSize, 0)); diff --git a/projects/hip-tests/catch/unit/atomics/atomicMax_system.cc b/projects/hip-tests/catch/unit/atomics/atomicMax_system.cc index 8484a3ebd0..ae221c95cb 100644 --- a/projects/hip-tests/catch/unit/atomics/atomicMax_system.cc +++ b/projects/hip-tests/catch/unit/atomics/atomicMax_system.cc @@ -46,11 +46,13 @@ THE SOFTWARE. * - HIP_VERSION >= 5.2 */ #if HT_AMD -TEMPLATE_TEST_CASE("Unit_atomicMax_system_Positive_Peer_GPUs_Same_Address", "", int, unsigned int, - unsigned long, unsigned long long, float, double) { +TEMPLATE_TEST_CASE("Unit_atomicMax_system_Positive_Peer_GPUs_Same_Address", + "[multigpu]", int, unsigned int, unsigned long, + unsigned long long, float, double) { #else -TEMPLATE_TEST_CASE("Unit_atomicMax_system_Positive_Peer_GPUs_Same_Address", "", int, unsigned int, - unsigned long, unsigned long long) { +TEMPLATE_TEST_CASE("Unit_atomicMax_system_Positive_Peer_GPUs_Same_Address", + "[multigpu]", int, unsigned int, unsigned long, + unsigned long long) { #endif for (auto current = 0; current < 1; ++current) { DYNAMIC_SECTION("Same address " << current) { @@ -74,11 +76,13 @@ TEMPLATE_TEST_CASE("Unit_atomicMax_system_Positive_Peer_GPUs_Same_Address", "", * - HIP_VERSION >= 5.2 */ #if HT_AMD -TEMPLATE_TEST_CASE("Unit_atomicMax_system_Positive_Peer_GPUs_Adjacent_Addresses", "", int, - unsigned int, unsigned long, unsigned long long, float, double) { +TEMPLATE_TEST_CASE( + "Unit_atomicMax_system_Positive_Peer_GPUs_Adjacent_Addresses", "[multigpu]", + int, unsigned int, unsigned long, unsigned long long, float, double) { #else -TEMPLATE_TEST_CASE("Unit_atomicMax_system_Positive_Peer_GPUs_Adjacent_Addresses", "", int, - unsigned int, unsigned long, unsigned long long) { +TEMPLATE_TEST_CASE( + "Unit_atomicMax_system_Positive_Peer_GPUs_Adjacent_Addresses", "[multigpu]", + int, unsigned int, unsigned long, unsigned long long) { #endif int warp_size = 0; HIP_CHECK(hipDeviceGetAttribute(&warp_size, hipDeviceAttributeWarpSize, 0)); @@ -105,11 +109,14 @@ TEMPLATE_TEST_CASE("Unit_atomicMax_system_Positive_Peer_GPUs_Adjacent_Addresses" * - HIP_VERSION >= 5.2 */ #if HT_AMD -TEMPLATE_TEST_CASE("Unit_atomicMax_system_Positive_Peer_GPUs_Scattered_Addresses", "", int, - unsigned int, unsigned long, unsigned long long, float, double) { +TEMPLATE_TEST_CASE( + "Unit_atomicMax_system_Positive_Peer_GPUs_Scattered_Addresses", + "[multigpu]", int, unsigned int, unsigned long, unsigned long long, float, + double) { #else -TEMPLATE_TEST_CASE("Unit_atomicMax_system_Positive_Peer_GPUs_Scattered_Addresses", "", int, - unsigned int, unsigned long, unsigned long long) { +TEMPLATE_TEST_CASE( + "Unit_atomicMax_system_Positive_Peer_GPUs_Scattered_Addresses", + "[multigpu]", int, unsigned int, unsigned long, unsigned long long) { #endif int warp_size = 0; HIP_CHECK(hipDeviceGetAttribute(&warp_size, hipDeviceAttributeWarpSize, 0)); diff --git a/projects/hip-tests/catch/unit/atomics/atomicMin_system.cc b/projects/hip-tests/catch/unit/atomics/atomicMin_system.cc index 9d7f646b36..1a0428ee14 100644 --- a/projects/hip-tests/catch/unit/atomics/atomicMin_system.cc +++ b/projects/hip-tests/catch/unit/atomics/atomicMin_system.cc @@ -46,11 +46,13 @@ THE SOFTWARE. * - HIP_VERSION >= 5.2 */ #if HT_AMD -TEMPLATE_TEST_CASE("Unit_atomicMin_system_Positive_Peer_GPUs_Same_Address", "", int, unsigned int, - unsigned long, unsigned long long, float, double) { +TEMPLATE_TEST_CASE("Unit_atomicMin_system_Positive_Peer_GPUs_Same_Address", + "[multigpu]", int, unsigned int, unsigned long, + unsigned long long, float, double) { #else -TEMPLATE_TEST_CASE("Unit_atomicMin_system_Positive_Peer_GPUs_Same_Address", "", int, unsigned int, - unsigned long, unsigned long long) { +TEMPLATE_TEST_CASE("Unit_atomicMin_system_Positive_Peer_GPUs_Same_Address", + "[multigpu]", int, unsigned int, unsigned long, + unsigned long long) { #endif for (auto current = 0; current < 1; ++current) { DYNAMIC_SECTION("Same address " << current) { @@ -74,11 +76,13 @@ TEMPLATE_TEST_CASE("Unit_atomicMin_system_Positive_Peer_GPUs_Same_Address", "", * - HIP_VERSION >= 5.2 */ #if HT_AMD -TEMPLATE_TEST_CASE("Unit_atomicMin_system_Positive_Peer_GPUs_Adjacent_Addresses", "", int, - unsigned int, unsigned long, unsigned long long, float, double) { +TEMPLATE_TEST_CASE( + "Unit_atomicMin_system_Positive_Peer_GPUs_Adjacent_Addresses", "[multigpu]", + int, unsigned int, unsigned long, unsigned long long, float, double) { #else -TEMPLATE_TEST_CASE("Unit_atomicMin_system_Positive_Peer_GPUs_Adjacent_Addresses", "", int, - unsigned int, unsigned long, unsigned long long) { +TEMPLATE_TEST_CASE( + "Unit_atomicMin_system_Positive_Peer_GPUs_Adjacent_Addresses", "[multigpu]", + int, unsigned int, unsigned long, unsigned long long) { #endif int warp_size = 0; HIP_CHECK(hipDeviceGetAttribute(&warp_size, hipDeviceAttributeWarpSize, 0)); @@ -105,11 +109,14 @@ TEMPLATE_TEST_CASE("Unit_atomicMin_system_Positive_Peer_GPUs_Adjacent_Addresses" * - HIP_VERSION >= 5.2 */ #if HT_AMD -TEMPLATE_TEST_CASE("Unit_atomicMin_system_Positive_Peer_GPUs_Scattered_Addresses", "", int, - unsigned int, unsigned long, unsigned long long, float, double) { +TEMPLATE_TEST_CASE( + "Unit_atomicMin_system_Positive_Peer_GPUs_Scattered_Addresses", + "[multigpu]", int, unsigned int, unsigned long, unsigned long long, float, + double) { #else -TEMPLATE_TEST_CASE("Unit_atomicMin_system_Positive_Peer_GPUs_Scattered_Addresses", "", int, - unsigned int, unsigned long, unsigned long long) { +TEMPLATE_TEST_CASE( + "Unit_atomicMin_system_Positive_Peer_GPUs_Scattered_Addresses", + "[multigpu]", int, unsigned int, unsigned long, unsigned long long) { #endif int warp_size = 0; HIP_CHECK(hipDeviceGetAttribute(&warp_size, hipDeviceAttributeWarpSize, 0)); diff --git a/projects/hip-tests/catch/unit/atomics/atomicOr_system.cc b/projects/hip-tests/catch/unit/atomics/atomicOr_system.cc index ba0906560c..009091a1a3 100644 --- a/projects/hip-tests/catch/unit/atomics/atomicOr_system.cc +++ b/projects/hip-tests/catch/unit/atomics/atomicOr_system.cc @@ -45,8 +45,9 @@ THE SOFTWARE. * - Multi-device * - HIP_VERSION >= 5.2 */ -TEMPLATE_TEST_CASE("Unit_atomicOr_system_Positive_Peer_GPUs_Same_Address", "", int, unsigned int, - unsigned long, unsigned long long) { +TEMPLATE_TEST_CASE("Unit_atomicOr_system_Positive_Peer_GPUs_Same_Address", + "[multigpu]", int, unsigned int, unsigned long, + unsigned long long) { for (auto current = 0; current < 1; ++current) { DYNAMIC_SECTION("Same address " << current) { Bitwise::MultipleDeviceMultipleKernelTest( @@ -68,8 +69,9 @@ TEMPLATE_TEST_CASE("Unit_atomicOr_system_Positive_Peer_GPUs_Same_Address", "", i * - Multi-device * - HIP_VERSION >= 5.2 */ -TEMPLATE_TEST_CASE("Unit_atomicOr_system_Positive_Peer_GPUs_Adjacent_Addresses", "", int, - unsigned int, unsigned long, unsigned long long) { +TEMPLATE_TEST_CASE("Unit_atomicOr_system_Positive_Peer_GPUs_Adjacent_Addresses", + "[multigpu]", int, unsigned int, unsigned long, + unsigned long long) { int warp_size = 0; HIP_CHECK(hipDeviceGetAttribute(&warp_size, hipDeviceAttributeWarpSize, 0)); @@ -94,8 +96,9 @@ TEMPLATE_TEST_CASE("Unit_atomicOr_system_Positive_Peer_GPUs_Adjacent_Addresses", * - Multi-device * - HIP_VERSION >= 5.2 */ -TEMPLATE_TEST_CASE("Unit_atomicOr_system_Positive_Peer_GPUs_Scattered_Addresses", "", int, - unsigned int, unsigned long, unsigned long long) { +TEMPLATE_TEST_CASE( + "Unit_atomicOr_system_Positive_Peer_GPUs_Scattered_Addresses", "[multigpu]", + int, unsigned int, unsigned long, unsigned long long) { int warp_size = 0; HIP_CHECK(hipDeviceGetAttribute(&warp_size, hipDeviceAttributeWarpSize, 0)); const auto cache_line_size = 128u; diff --git a/projects/hip-tests/catch/unit/atomics/atomicSub_system.cc b/projects/hip-tests/catch/unit/atomics/atomicSub_system.cc index a5a1111b52..df5133493e 100644 --- a/projects/hip-tests/catch/unit/atomics/atomicSub_system.cc +++ b/projects/hip-tests/catch/unit/atomics/atomicSub_system.cc @@ -54,8 +54,9 @@ THE SOFTWARE. * ------------------------ * - HIP_VERSION >= 5.2 */ -TEMPLATE_TEST_CASE("Unit_atomicSub_system_Positive_Peer_GPUs", "", int, unsigned int, unsigned long, - unsigned long long, float, double) { +TEMPLATE_TEST_CASE("Unit_atomicSub_system_Positive_Peer_GPUs", "[multigpu]", + int, unsigned int, unsigned long, unsigned long long, float, + double) { int warp_size = 0; HIP_CHECK(hipDeviceGetAttribute(&warp_size, hipDeviceAttributeWarpSize, 0)); const auto cache_line_size = 128u; @@ -103,8 +104,9 @@ TEMPLATE_TEST_CASE("Unit_atomicSub_system_Positive_Peer_GPUs", "", int, unsigned * ------------------------ * - HIP_VERSION >= 5.2 */ -TEMPLATE_TEST_CASE("Unit_atomicSub_system_Positive_Host_And_GPU", "", int, unsigned int, - unsigned long, unsigned long long, float, double) { +TEMPLATE_TEST_CASE("Unit_atomicSub_system_Positive_Host_And_GPU", "[multigpu]", + int, unsigned int, unsigned long, unsigned long long, float, + double) { int warp_size = 0; HIP_CHECK(hipDeviceGetAttribute(&warp_size, hipDeviceAttributeWarpSize, 0)); const auto cache_line_size = 128u; @@ -152,8 +154,9 @@ TEMPLATE_TEST_CASE("Unit_atomicSub_system_Positive_Host_And_GPU", "", int, unsig * ------------------------ * - HIP_VERSION >= 5.2 */ -TEMPLATE_TEST_CASE("Unit_atomicSub_system_Positive_Host_And_Peer_GPUs", "", int, unsigned int, - unsigned long, unsigned long long, float, double) { +TEMPLATE_TEST_CASE("Unit_atomicSub_system_Positive_Host_And_Peer_GPUs", + "[multigpu]", int, unsigned int, unsigned long, + unsigned long long, float, double) { int warp_size = 0; HIP_CHECK(hipDeviceGetAttribute(&warp_size, hipDeviceAttributeWarpSize, 0)); const auto cache_line_size = 128u; diff --git a/projects/hip-tests/catch/unit/atomics/atomicXor_system.cc b/projects/hip-tests/catch/unit/atomics/atomicXor_system.cc index 0d3523b382..71346d8904 100644 --- a/projects/hip-tests/catch/unit/atomics/atomicXor_system.cc +++ b/projects/hip-tests/catch/unit/atomics/atomicXor_system.cc @@ -45,8 +45,9 @@ THE SOFTWARE. * - Multi-device * - HIP_VERSION >= 5.2 */ -TEMPLATE_TEST_CASE("Unit_atomicXor_system_Positive_Peer_GPUs_Same_Address", "", int, unsigned int, - unsigned long, unsigned long long) { +TEMPLATE_TEST_CASE("Unit_atomicXor_system_Positive_Peer_GPUs_Same_Address", + "[multigpu]", int, unsigned int, unsigned long, + unsigned long long) { for (auto current = 0; current < 1; ++current) { DYNAMIC_SECTION("Same address " << current) { Bitwise::MultipleDeviceMultipleKernelTest( @@ -68,8 +69,9 @@ TEMPLATE_TEST_CASE("Unit_atomicXor_system_Positive_Peer_GPUs_Same_Address", "", * - Multi-device * - HIP_VERSION >= 5.2 */ -TEMPLATE_TEST_CASE("Unit_atomicXor_system_Positive_Peer_GPUs_Adjacent_Addresses", "", int, - unsigned int, unsigned long, unsigned long long) { +TEMPLATE_TEST_CASE( + "Unit_atomicXor_system_Positive_Peer_GPUs_Adjacent_Addresses", "[multigpu]", + int, unsigned int, unsigned long, unsigned long long) { int warp_size = 0; HIP_CHECK(hipDeviceGetAttribute(&warp_size, hipDeviceAttributeWarpSize, 0)); @@ -94,8 +96,9 @@ TEMPLATE_TEST_CASE("Unit_atomicXor_system_Positive_Peer_GPUs_Adjacent_Addresses" * - Multi-device * - HIP_VERSION >= 5.2 */ -TEMPLATE_TEST_CASE("Unit_atomicXor_system_Positive_Peer_GPUs_Scattered_Addresses", "", int, - unsigned int, unsigned long, unsigned long long) { +TEMPLATE_TEST_CASE( + "Unit_atomicXor_system_Positive_Peer_GPUs_Scattered_Addresses", + "[multigpu]", int, unsigned int, unsigned long, unsigned long long) { int warp_size = 0; HIP_CHECK(hipDeviceGetAttribute(&warp_size, hipDeviceAttributeWarpSize, 0)); const auto cache_line_size = 128u; diff --git a/projects/hip-tests/catch/unit/callback/hipGetStreamDeviceId.cc b/projects/hip-tests/catch/unit/callback/hipGetStreamDeviceId.cc index 1387d58ae9..c014336797 100644 --- a/projects/hip-tests/catch/unit/callback/hipGetStreamDeviceId.cc +++ b/projects/hip-tests/catch/unit/callback/hipGetStreamDeviceId.cc @@ -65,7 +65,8 @@ TEST_CASE("Unit_hipGetStreamDeviceId_Positive_Threaded_Basic") { * - Platform specific (AMD) * - Multithreaded GPU */ -TEST_CASE("Unit_hipGetStreamDeviceId_Positive_Multithreaded_Basic") { +TEST_CASE("Unit_hipGetStreamDeviceId_Positive_Multithreaded_Basic", + "[multigpu]") { const unsigned int max_threads = std::thread::hardware_concurrency(); const int device_count = HipTest::getDeviceCount(); diff --git a/projects/hip-tests/catch/unit/cooperativeGrps/hipCGMultiGridGroupType_old.cc b/projects/hip-tests/catch/unit/cooperativeGrps/hipCGMultiGridGroupType_old.cc index 410a3e49f8..050402689e 100644 --- a/projects/hip-tests/catch/unit/cooperativeGrps/hipCGMultiGridGroupType_old.cc +++ b/projects/hip-tests/catch/unit/cooperativeGrps/hipCGMultiGridGroupType_old.cc @@ -378,7 +378,7 @@ template static void test_cg_multi_grid_group_type(F kernel_func, i } } -TEST_CASE("Unit_hipCGMultiGridGroupType_Basic") { +TEST_CASE("Unit_hipCGMultiGridGroupType_Basic", "[multigpu]") { int num_devices = 0; HIP_CHECK(hipGetDeviceCount(&num_devices)); num_devices = min(num_devices, MaxGPUs); @@ -425,7 +425,7 @@ TEST_CASE("Unit_hipCGMultiGridGroupType_Basic") { } } -TEST_CASE("Unit_hipCGMultiGridGroupType_Barrier") { +TEST_CASE("Unit_hipCGMultiGridGroupType_Barrier", "[multigpu]") { int num_devices = 0; uint32_t loops = GENERATE(1, 2, 3, 4); uint32_t warps = GENERATE(4, 8, 16, 32); diff --git a/projects/hip-tests/catch/unit/cooperativeGrps/hipLaunchCooperativeKernelMultiDevice_old.cc b/projects/hip-tests/catch/unit/cooperativeGrps/hipLaunchCooperativeKernelMultiDevice_old.cc index 6a6e4d41b4..6be0cab225 100644 --- a/projects/hip-tests/catch/unit/cooperativeGrps/hipLaunchCooperativeKernelMultiDevice_old.cc +++ b/projects/hip-tests/catch/unit/cooperativeGrps/hipLaunchCooperativeKernelMultiDevice_old.cc @@ -130,7 +130,7 @@ __global__ void test_gws(uint* buf, uint buf_size, long* tmp_buf, long* result) } } -TEST_CASE("Unit_hipLaunchCooperativeKernelMultiDevice_Basic") { +TEST_CASE("Unit_hipLaunchCooperativeKernelMultiDevice_Basic", "[multigpu]") { constexpr uint num_kernel_args = 4; int device_num = 0; diff --git a/projects/hip-tests/catch/unit/cooperativeGrps/multi_grid_group.cc b/projects/hip-tests/catch/unit/cooperativeGrps/multi_grid_group.cc index f3ddc839a0..1c337db864 100644 --- a/projects/hip-tests/catch/unit/cooperativeGrps/multi_grid_group.cc +++ b/projects/hip-tests/catch/unit/cooperativeGrps/multi_grid_group.cc @@ -154,7 +154,7 @@ static void get_multi_grid_dims(dim3& grid_dim, dim3& block_dim, unsigned int de * - HIP_VERSION >= 5.2 * - Devices support cooperative multi device launch */ -TEST_CASE("Unit_Multi_Grid_Group_Getters_Positive_Basic") { +TEST_CASE("Unit_Multi_Grid_Group_Getters_Positive_Basic", "[multigpu]") { int num_devices = 0; HIP_CHECK(hipGetDeviceCount(&num_devices)); num_devices = min(num_devices, kMaxGPUs); @@ -302,7 +302,7 @@ TEST_CASE("Unit_Multi_Grid_Group_Getters_Positive_Basic") { * - HIP_VERSION >= 5.2 * - Devices support cooperative multi device launch */ -TEST_CASE("Unit_Multi_Grid_Group_Getters_Positive_Base_Type") { +TEST_CASE("Unit_Multi_Grid_Group_Getters_Positive_Base_Type", "[multigpu]") { int num_devices = 0; HIP_CHECK(hipGetDeviceCount(&num_devices)); num_devices = min(num_devices, kMaxGPUs); @@ -423,7 +423,8 @@ TEST_CASE("Unit_Multi_Grid_Group_Getters_Positive_Base_Type") { * - HIP_VERSION >= 5.2 * - Devices support cooperative multi device launch */ -TEST_CASE("Unit_Multi_Grid_Group_Getters_Positive_Non_Member_Functions") { +TEST_CASE("Unit_Multi_Grid_Group_Getters_Positive_Non_Member_Functions", + "[multigpu]") { int num_devices = 0; HIP_CHECK(hipGetDeviceCount(&num_devices)); num_devices = min(num_devices, kMaxGPUs); @@ -535,7 +536,7 @@ TEST_CASE("Unit_Multi_Grid_Group_Getters_Positive_Non_Member_Functions") { * - HIP_VERSION >= 5.2 * - Devices support cooperative multi device launch */ -TEST_CASE("Unit_Multi_Grid_Group_Positive_Sync") { +TEST_CASE("Unit_Multi_Grid_Group_Positive_Sync", "[multigpu]") { CHECK_IMAGE_SUPPORT int num_devices = 0; HIP_CHECK(hipGetDeviceCount(&num_devices)); diff --git a/projects/hip-tests/catch/unit/device/hipDeviceEnableDisablePeerAccess.cc b/projects/hip-tests/catch/unit/device/hipDeviceEnableDisablePeerAccess.cc index c97318d894..63e30cba91 100644 --- a/projects/hip-tests/catch/unit/device/hipDeviceEnableDisablePeerAccess.cc +++ b/projects/hip-tests/catch/unit/device/hipDeviceEnableDisablePeerAccess.cc @@ -46,7 +46,7 @@ THE SOFTWARE. * - Multi-device * - HIP_VERSION >= 5.2 */ -TEST_CASE("Unit_hipDeviceEnableDisablePeerAccess_positive") { +TEST_CASE("Unit_hipDeviceEnableDisablePeerAccess_positive", "[multigpu]") { int canAccessPeer = 0; int deviceCount = HipTest::getGeviceCount(); if (deviceCount < 2) { @@ -95,7 +95,7 @@ TEST_CASE("Unit_hipDeviceEnableDisablePeerAccess_positive") { * - Multi-device * - HIP_VERSION >= 5.2 */ -TEST_CASE("Unit_hipDeviceEnablePeerAccess_negative") { +TEST_CASE("Unit_hipDeviceEnablePeerAccess_negative", "[multigpu]") { int deviceCount = HipTest::getGeviceCount(); if (deviceCount < 2) { HipTest::HIP_SKIP_TEST("Skipping because devices < 2"); @@ -159,7 +159,7 @@ TEST_CASE("Unit_hipDeviceEnablePeerAccess_negative") { * - Multi-device * - HIP_VERSION >= 5.2 */ -TEST_CASE("Unit_hipDeviceDisablePeerAccess_negative") { +TEST_CASE("Unit_hipDeviceDisablePeerAccess_negative", "[multigpu]") { int deviceCount = HipTest::getGeviceCount(); if (deviceCount < 2) { HipTest::HIP_SKIP_TEST("Skipping because devices < 2"); diff --git a/projects/hip-tests/catch/unit/device/hipDeviceGetName.cc b/projects/hip-tests/catch/unit/device/hipDeviceGetName.cc index 028358bf66..a294ba0182 100644 --- a/projects/hip-tests/catch/unit/device/hipDeviceGetName.cc +++ b/projects/hip-tests/catch/unit/device/hipDeviceGetName.cc @@ -210,7 +210,8 @@ static inline std::vector parseVisibleDevices() { * ------------------------ * - HIP_VERSION >= 5.7 */ -TEST_CASE("Unit_hipDeviceName_gcnArchName_And_rocm_agent_enumerator") { +TEST_CASE("Unit_hipDeviceName_gcnArchName_And_rocm_agent_enumerator", + "[multigpu]") { int deviceCount = 0; HIP_CHECK(hipGetDeviceCount(&deviceCount)); if (deviceCount <= 0) { diff --git a/projects/hip-tests/catch/unit/device/hipDeviceGetUuid.cc b/projects/hip-tests/catch/unit/device/hipDeviceGetUuid.cc index 551cc595ca..9f5874c950 100644 --- a/projects/hip-tests/catch/unit/device/hipDeviceGetUuid.cc +++ b/projects/hip-tests/catch/unit/device/hipDeviceGetUuid.cc @@ -145,7 +145,7 @@ static inline std::vector parseVisibleDevices() { * ------------------------ * - HIP_VERSION >= 5.7 */ -TEST_CASE("Unit_hipDeviceGetUuid_From_RocmInfo") { +TEST_CASE("Unit_hipDeviceGetUuid_From_RocmInfo", "[multigpu]") { int deviceCount = 0; HIP_CHECK(hipGetDeviceCount(&deviceCount)); assert(deviceCount > 0); @@ -219,7 +219,8 @@ TEST_CASE("Unit_hipDeviceGetUuid_From_RocmInfo") { */ // Guarding it against NVIDIA as this test is faling on it. #if HT_AMD -TEST_CASE("Unit_hipDeviceGetUuid_VerifyUuidFrm_hipGetDeviceProperties") { +TEST_CASE("Unit_hipDeviceGetUuid_VerifyUuidFrm_hipGetDeviceProperties", + "[multigpu]") { int deviceCount = 0; hipDevice_t device; hipDeviceProp_t prop; diff --git a/projects/hip-tests/catch/unit/device/hipDeviceSetGetLimit.cc b/projects/hip-tests/catch/unit/device/hipDeviceSetGetLimit.cc index 5f70c18666..a7ba453746 100644 --- a/projects/hip-tests/catch/unit/device/hipDeviceSetGetLimit.cc +++ b/projects/hip-tests/catch/unit/device/hipDeviceSetGetLimit.cc @@ -462,7 +462,7 @@ void getMinMaxCurrentAndSetCurrent() { * ------------------------ * - HIP_VERSION >= 6.5 */ -TEST_CASE("Unit_hipDeviceGetSetLimit_Scratch_MultiDevice") { +TEST_CASE("Unit_hipDeviceGetSetLimit_Scratch_MultiDevice", "[multigpu]") { int deviceCount = 0; HIP_CHECK(hipGetDeviceCount(&deviceCount)); if (deviceCount < 2) { diff --git a/projects/hip-tests/catch/unit/device/hipDeviceSetLimit_old.cc b/projects/hip-tests/catch/unit/device/hipDeviceSetLimit_old.cc index 657a7b6bfc..9ccf71bdc4 100644 --- a/projects/hip-tests/catch/unit/device/hipDeviceSetLimit_old.cc +++ b/projects/hip-tests/catch/unit/device/hipDeviceSetLimit_old.cc @@ -65,7 +65,7 @@ static bool testSetLimitFunc(hipLimit_t limit_to_test) { * ------------------------ * - HIP_VERSION >= 5.2 */ -TEST_CASE("Unit_hipDeviceSetLimit_SetGet") { +TEST_CASE("Unit_hipDeviceSetLimit_SetGet", "[multigpu]") { size_t value = 0; // Scenario1 SECTION("Set Get Test hipLimitStackSize") { diff --git a/projects/hip-tests/catch/unit/device/hipDeviceTotalMem.cc b/projects/hip-tests/catch/unit/device/hipDeviceTotalMem.cc index 8097219298..67addeddaa 100644 --- a/projects/hip-tests/catch/unit/device/hipDeviceTotalMem.cc +++ b/projects/hip-tests/catch/unit/device/hipDeviceTotalMem.cc @@ -115,7 +115,7 @@ TEST_CASE("Unit_hipDeviceTotalMem_ValidateTotalMem") { * - Multi-device test * - HIP_VERSION >= 5.2 */ -TEST_CASE("Unit_hipDeviceTotalMem_NonSelectedDevice") { +TEST_CASE("Unit_hipDeviceTotalMem_NonSelectedDevice", "[multigpu]") { auto deviceCount = HipTest::getDeviceCount(); if (deviceCount < 2) { HipTest::HIP_SKIP_TEST("Multi Device Test, will not run on single gpu systems. Skipping."); diff --git a/projects/hip-tests/catch/unit/device/hipGetDeviceProperties.cc b/projects/hip-tests/catch/unit/device/hipGetDeviceProperties.cc index 690da82ea5..dc01b7167e 100644 --- a/projects/hip-tests/catch/unit/device/hipGetDeviceProperties.cc +++ b/projects/hip-tests/catch/unit/device/hipGetDeviceProperties.cc @@ -148,7 +148,7 @@ static void validateDeviceMacro(int* archProp_h, hipDeviceProp_t* prop) { * - Platform specific (AMD) * - HIP_VERSION >= 5.2 */ -TEST_CASE("Unit_hipGetDeviceProperties_ArchPropertiesTst") { +TEST_CASE("Unit_hipGetDeviceProperties_ArchPropertiesTst", "[multigpu]") { int *archProp_h, *archProp_d; archProp_h = new int[NUM_OF_ARCHPROP]; hipDeviceProp_t prop; diff --git a/projects/hip-tests/catch/unit/device/hipGetProcAddressDevMgmt.cc b/projects/hip-tests/catch/unit/device/hipGetProcAddressDevMgmt.cc index f056a99345..cad02d3158 100644 --- a/projects/hip-tests/catch/unit/device/hipGetProcAddressDevMgmt.cc +++ b/projects/hip-tests/catch/unit/device/hipGetProcAddressDevMgmt.cc @@ -372,7 +372,7 @@ TEST_CASE("Unit_hipGetProcAddress_ValidateDeviceApis") { * - HIP_VERSION >= 6.2 */ -TEST_CASE("Unit_hipGetProcAddress_PeerDeviceAccessAPIs") { +TEST_CASE("Unit_hipGetProcAddress_PeerDeviceAccessAPIs", "[multigpu]") { void* hipDeviceCanAccessPeer_ptr = nullptr; void* hipSetDevice_ptr = nullptr; void* hipGetDevice_ptr = nullptr; @@ -453,7 +453,7 @@ bool CheckMemPoolSupport(const int device) { return true; } -TEST_CASE("Unit_hipGetProcAddress_SetGetMemPoolAPIs") { +TEST_CASE("Unit_hipGetProcAddress_SetGetMemPoolAPIs", "[multigpu]") { void* hipDeviceSetMemPool_ptr = nullptr; void* hipDeviceGetMemPool_ptr = nullptr; int currentHipVersion = 0; diff --git a/projects/hip-tests/catch/unit/device/hipSetGetDevice.cc b/projects/hip-tests/catch/unit/device/hipSetGetDevice.cc index 561bf6d6a7..db740cc65e 100644 --- a/projects/hip-tests/catch/unit/device/hipSetGetDevice.cc +++ b/projects/hip-tests/catch/unit/device/hipSetGetDevice.cc @@ -44,7 +44,7 @@ THE SOFTWARE. * ------------------------ * - HIP_VERSION >= 5.2 */ -TEST_CASE("Unit_hipSetDevice_BasicSetGet") { +TEST_CASE("Unit_hipSetDevice_BasicSetGet", "[multigpu]") { int numDevices = 0; int device{}; HIP_CHECK(hipGetDeviceCount(&numDevices)); @@ -73,7 +73,7 @@ TEST_CASE("Unit_hipSetDevice_BasicSetGet") { * ------------------------ * - HIP_VERSION >= 5.2 */ -TEST_CASE("Unit_hipGetSetDevice_MultiThreaded") { +TEST_CASE("Unit_hipGetSetDevice_MultiThreaded", "[multigpu]") { auto maxThreads = std::thread::hardware_concurrency(); auto deviceCount = HipTest::getDeviceCount(); @@ -126,7 +126,7 @@ TEST_CASE("Unit_hipGetSetDevice_MultiThreaded") { * ------------------------ * - HIP_VERSION >= 5.2 */ -TEST_CASE("Unit_hipSetGetDevice_Positive_Threaded_Basic") { +TEST_CASE("Unit_hipSetGetDevice_Positive_Threaded_Basic", "[multigpu]") { class HipSetGetDeviceThreadedTest : public ThreadedZigZagTest { public: void TestPart1() { HIP_CHECK(hipSetDevice(0)); } diff --git a/projects/hip-tests/catch/unit/device/hipSetValidDevices.cc b/projects/hip-tests/catch/unit/device/hipSetValidDevices.cc index 3f4643be77..25eb63b648 100644 --- a/projects/hip-tests/catch/unit/device/hipSetValidDevices.cc +++ b/projects/hip-tests/catch/unit/device/hipSetValidDevices.cc @@ -158,7 +158,7 @@ TEST_CASE("Unit_hipSetValidDevices_Negative_Length_Lessthan_DeviceArrSize") { * ------------------------ * - HIP_VERSION >= 7.1 */ -TEST_CASE("Unit_hipSetValidDevices_Positive_Basic") { +TEST_CASE("Unit_hipSetValidDevices_Positive_Basic", "[multigpu]") { int totalDevices = HipTest::getDeviceCount(); if (totalDevices < 2) { HipTest::HIP_SKIP_TEST("This test requires 2 or more GPUs. Skipping."); diff --git a/projects/hip-tests/catch/unit/deviceLib/threadfence_system.cc b/projects/hip-tests/catch/unit/deviceLib/threadfence_system.cc index 5d65366708..d2b28ee1e3 100644 --- a/projects/hip-tests/catch/unit/deviceLib/threadfence_system.cc +++ b/projects/hip-tests/catch/unit/deviceLib/threadfence_system.cc @@ -51,7 +51,7 @@ __global__ void gpu_round_robin(const int id, const int num_dev, const int num_i round_robin(id, num_dev, num_iter, data, flag); } -TEST_CASE("Unit_threadfence_system") { +TEST_CASE("Unit_threadfence_system", "[multigpu]") { int num_gpus = 0; HIP_CHECK(hipGetDeviceCount(&num_gpus)); REQUIRE(num_gpus > 0); diff --git a/projects/hip-tests/catch/unit/errorHandling/hipExtGetLastError.cc b/projects/hip-tests/catch/unit/errorHandling/hipExtGetLastError.cc index 8e0acd21a4..1f5b503801 100644 --- a/projects/hip-tests/catch/unit/errorHandling/hipExtGetLastError.cc +++ b/projects/hip-tests/catch/unit/errorHandling/hipExtGetLastError.cc @@ -93,7 +93,7 @@ TEST_CASE("Unit_hipExtGetLastError_Positive_Threaded") { *  - HIP_VERSION >= 6.4 */ -TEST_CASE("Unit_hipExtGetLastError_with_hipMemcpyPeerAsync") { +TEST_CASE("Unit_hipExtGetLastError_with_hipMemcpyPeerAsync", "[multigpu]") { const auto device_count = HipTest::getDeviceCount(); if (device_count < 2) { HipTest::HIP_SKIP_TEST("Skipping because devices < 2"); diff --git a/projects/hip-tests/catch/unit/errorHandling/hipGetLastError.cc b/projects/hip-tests/catch/unit/errorHandling/hipGetLastError.cc index 0b7dfa6d95..624548dc20 100644 --- a/projects/hip-tests/catch/unit/errorHandling/hipGetLastError.cc +++ b/projects/hip-tests/catch/unit/errorHandling/hipGetLastError.cc @@ -94,7 +94,7 @@ TEST_CASE("Unit_hipGetLastError_Positive_Threaded") { *  - HIP_VERSION >= 6.0 */ -TEST_CASE("Unit_hipGetLastError_with_hipMemcpyPeerAsync") { +TEST_CASE("Unit_hipGetLastError_with_hipMemcpyPeerAsync", "[multigpu]") { const auto device_count = HipTest::getDeviceCount(); if (device_count < 2) { HipTest::HIP_SKIP_TEST("Skipping because devices < 2"); diff --git a/projects/hip-tests/catch/unit/errorHandling/hipGetLastErrorOnAbort.cc b/projects/hip-tests/catch/unit/errorHandling/hipGetLastErrorOnAbort.cc index 1570e55ed6..1d56ae4a9a 100644 --- a/projects/hip-tests/catch/unit/errorHandling/hipGetLastErrorOnAbort.cc +++ b/projects/hip-tests/catch/unit/errorHandling/hipGetLastErrorOnAbort.cc @@ -102,7 +102,7 @@ TEST_CASE("Unit_hipGetLastError_KernelFailure_ValidAndInvalidOperations") { * ------------------------ * - HIP_VERSION >= 7.0 */ -TEST_CASE("Unit_hipGetLastError_KernelFailure_TwoDevices") { +TEST_CASE("Unit_hipGetLastError_KernelFailure_TwoDevices", "[multigpu]") { int deviceCount = 0; HIP_CHECK(hipGetDeviceCount(&deviceCount)); if (deviceCount < 2) { diff --git a/projects/hip-tests/catch/unit/event/Unit_hipEventElapsedTime.cc b/projects/hip-tests/catch/unit/event/Unit_hipEventElapsedTime.cc index 7fe8b9a5be..b8aec0ff23 100644 --- a/projects/hip-tests/catch/unit/event/Unit_hipEventElapsedTime.cc +++ b/projects/hip-tests/catch/unit/event/Unit_hipEventElapsedTime.cc @@ -108,7 +108,7 @@ TEST_CASE("Unit_hipEventElapsedTime_DisableTiming") { * ------------------------ * - HIP_VERSION >= 5.2 */ -TEST_CASE("Unit_hipEventElapsedTime_DifferentDevices") { +TEST_CASE("Unit_hipEventElapsedTime_DifferentDevices", "[multigpu]") { int devCount = 0; HIP_CHECK(hipGetDeviceCount(&devCount)); if (devCount > 1) { diff --git a/projects/hip-tests/catch/unit/event/Unit_hipEventMGpuMThreads.cc b/projects/hip-tests/catch/unit/event/Unit_hipEventMGpuMThreads.cc index 67d0c099b2..1c855dc73b 100644 --- a/projects/hip-tests/catch/unit/event/Unit_hipEventMGpuMThreads.cc +++ b/projects/hip-tests/catch/unit/event/Unit_hipEventMGpuMThreads.cc @@ -217,7 +217,7 @@ TEST_CASE("Unit_hipEventMGpuMThreads_1") { testEventMGpuMThreads(1); } * ------------------------ * - HIP_VERSION >= 5.2 */ -TEST_CASE("Unit_hipEventMGpuMThreads_2") { +TEST_CASE("Unit_hipEventMGpuMThreads_2", "[multigpu]") { int numDevices = 0; HIP_CHECK(hipGetDeviceCount(&numDevices)); if (numDevices > 1) { @@ -238,7 +238,7 @@ TEST_CASE("Unit_hipEventMGpuMThreads_2") { * ------------------------ * - HIP_VERSION >= 5.2 */ -TEST_CASE("Unit_hipEventMGpuMThreads_3") { +TEST_CASE("Unit_hipEventMGpuMThreads_3", "[multigpu]") { int numDevices = 0; HIP_CHECK(hipGetDeviceCount(&numDevices)); if (numDevices > 1) { diff --git a/projects/hip-tests/catch/unit/event/Unit_hipEventQuery.cc b/projects/hip-tests/catch/unit/event/Unit_hipEventQuery.cc index 955447c38d..a0ddb5104c 100644 --- a/projects/hip-tests/catch/unit/event/Unit_hipEventQuery.cc +++ b/projects/hip-tests/catch/unit/event/Unit_hipEventQuery.cc @@ -41,7 +41,7 @@ THE SOFTWARE. * ------------------------ * - HIP_VERSION >= 5.2 */ -TEST_CASE("Unit_hipEventQuery_DifferentDevice") { +TEST_CASE("Unit_hipEventQuery_DifferentDevice", "[multigpu]") { hipEvent_t event1{}, event2{}; HIP_CHECK(hipEventCreate(&event1)); HIP_CHECK(hipEventCreate(&event2)); diff --git a/projects/hip-tests/catch/unit/event/Unit_hipEventRecord.cc b/projects/hip-tests/catch/unit/event/Unit_hipEventRecord.cc index dd3f99fdf6..96734042ab 100644 --- a/projects/hip-tests/catch/unit/event/Unit_hipEventRecord.cc +++ b/projects/hip-tests/catch/unit/event/Unit_hipEventRecord.cc @@ -158,7 +158,7 @@ TEST_CASE("Unit_hipEventRecord") { * ------------------------ * - HIP_VERSION >= 5.2 */ -TEST_CASE("Unit_hipEventRecord_Negative") { +TEST_CASE("Unit_hipEventRecord_Negative", "[multigpu]") { SECTION("Nullptr event") { HIP_CHECK_ERROR(hipEventRecord(nullptr, nullptr), hipErrorInvalidResourceHandle); } diff --git a/projects/hip-tests/catch/unit/executionControl/hipExtLaunchMultiKernelMultiDevice.cc b/projects/hip-tests/catch/unit/executionControl/hipExtLaunchMultiKernelMultiDevice.cc index 2c19a33d15..6f76830f94 100644 --- a/projects/hip-tests/catch/unit/executionControl/hipExtLaunchMultiKernelMultiDevice.cc +++ b/projects/hip-tests/catch/unit/executionControl/hipExtLaunchMultiKernelMultiDevice.cc @@ -27,7 +27,8 @@ THE SOFTWARE. #include #include -TEST_CASE("Unit_hipExtLaunchMultiKernelMultiDevice_Positive_Basic") { +TEST_CASE("Unit_hipExtLaunchMultiKernelMultiDevice_Positive_Basic", + "[multigpu]") { const auto device_count = HipTest::getDeviceCount(); std::vector params_list(device_count); @@ -54,7 +55,8 @@ TEST_CASE("Unit_hipExtLaunchMultiKernelMultiDevice_Positive_Basic") { } } -TEST_CASE("Unit_hipExtLaunchMultiKernelMultiDevice_Negative_Parameters") { +TEST_CASE("Unit_hipExtLaunchMultiKernelMultiDevice_Negative_Parameters", + "[multigpu]") { const auto device_count = HipTest::getDeviceCount(); std::vector params_list(device_count); diff --git a/projects/hip-tests/catch/unit/executionControl/hipLaunchCooperativeKernelMultiDevice.cc b/projects/hip-tests/catch/unit/executionControl/hipLaunchCooperativeKernelMultiDevice.cc index 686adeccd7..f8a0b9b1da 100644 --- a/projects/hip-tests/catch/unit/executionControl/hipLaunchCooperativeKernelMultiDevice.cc +++ b/projects/hip-tests/catch/unit/executionControl/hipLaunchCooperativeKernelMultiDevice.cc @@ -27,7 +27,8 @@ THE SOFTWARE. #include #include -TEST_CASE("Unit_hipLaunchCooperativeKernelMultiDevice_Positive_Basic") { +TEST_CASE("Unit_hipLaunchCooperativeKernelMultiDevice_Positive_Basic", + "[multigpu]") { if (!DeviceAttributesSupport(0, hipDeviceAttributeCooperativeLaunch)) { HipTest::HIP_SKIP_TEST("CooperativeLaunch not supported"); return; @@ -59,7 +60,8 @@ TEST_CASE("Unit_hipLaunchCooperativeKernelMultiDevice_Positive_Basic") { } } -TEST_CASE("Unit_hipLaunchCooperativeKernelMultiDevice_Negative_Parameters") { +TEST_CASE("Unit_hipLaunchCooperativeKernelMultiDevice_Negative_Parameters", + "[multigpu]") { if (!DeviceAttributesSupport(0, hipDeviceAttributeCooperativeLaunch)) { HipTest::HIP_SKIP_TEST("CooperativeLaunch not supported"); return; diff --git a/projects/hip-tests/catch/unit/graph/hipDeviceGetGraphMemAttribute.cc b/projects/hip-tests/catch/unit/graph/hipDeviceGetGraphMemAttribute.cc index 6c15e5665d..59cddc0c79 100644 --- a/projects/hip-tests/catch/unit/graph/hipDeviceGetGraphMemAttribute.cc +++ b/projects/hip-tests/catch/unit/graph/hipDeviceGetGraphMemAttribute.cc @@ -309,7 +309,8 @@ TEST_CASE("Unit_hipDeviceGetGraphMemAttribute_Functional") { Unit_hipDeviceGetGraphMemAttribute_Functional(); } -TEST_CASE("Unit_hipDeviceGetGraphMemAttribute_Functional_Multi_Device") { +TEST_CASE("Unit_hipDeviceGetGraphMemAttribute_Functional_Multi_Device", + "[multigpu]") { int numDevices = 0; HIP_CHECK(hipGetDeviceCount(&numDevices)); diff --git a/projects/hip-tests/catch/unit/graph/hipDrvGraphAddMemcpyNode.cc b/projects/hip-tests/catch/unit/graph/hipDrvGraphAddMemcpyNode.cc index 03a899e132..9edabb5449 100644 --- a/projects/hip-tests/catch/unit/graph/hipDrvGraphAddMemcpyNode.cc +++ b/projects/hip-tests/catch/unit/graph/hipDrvGraphAddMemcpyNode.cc @@ -368,7 +368,7 @@ TEST_CASE("Unit_hipDrvGraphAddMemcpyNode_test") { * - HIP_VERSION >= 6.1 */ -TEST_CASE("Unit_hipDrvGraphAddMemcpyNode_MulitDevice") { +TEST_CASE("Unit_hipDrvGraphAddMemcpyNode_MulitDevice", "[multigpu]") { CHECK_IMAGE_SUPPORT int numDevices = 0; diff --git a/projects/hip-tests/catch/unit/graph/hipGraphAddChildGraphNode.cc b/projects/hip-tests/catch/unit/graph/hipGraphAddChildGraphNode.cc index 75382d073b..117cbd8a55 100644 --- a/projects/hip-tests/catch/unit/graph/hipGraphAddChildGraphNode.cc +++ b/projects/hip-tests/catch/unit/graph/hipGraphAddChildGraphNode.cc @@ -1083,7 +1083,7 @@ TEST_CASE("Unit_hipGraphAddChildGraphNode_MultGraphsAsSingleGraph") { in multi GPU environment. Create one nested graph per GPU context. Execute all the created graphs in their respective GPUs and validate the output. */ -TEST_CASE("Unit_hipGraphAddChildGraphNode_CmplxNstGrph_MultGPU") { +TEST_CASE("Unit_hipGraphAddChildGraphNode_CmplxNstGrph_MultGPU", "[multigpu]") { int devcount = 0; HIP_CHECK(hipGetDeviceCount(&devcount)); // If only single GPU is detected then return diff --git a/projects/hip-tests/catch/unit/graph/hipGraphAddMemAllocNode.cc b/projects/hip-tests/catch/unit/graph/hipGraphAddMemAllocNode.cc index 60a2797031..0a47091469 100644 --- a/projects/hip-tests/catch/unit/graph/hipGraphAddMemAllocNode.cc +++ b/projects/hip-tests/catch/unit/graph/hipGraphAddMemAllocNode.cc @@ -584,7 +584,7 @@ TEST_CASE("Unit_hipGraphAddMemAllocNode_Functional_1") { * ------------------------ *  - HIP_VERSION >= 6.1 */ -TEST_CASE("Unit_hipGraphAddMemAllocNode_Functional_2") { +TEST_CASE("Unit_hipGraphAddMemAllocNode_Functional_2", "[multigpu]") { int mem_pool_support = 0; HIP_CHECK(hipDeviceGetAttribute(&mem_pool_support, hipDeviceAttributeMemoryPoolsSupported, 0)); if (!mem_pool_support) { @@ -653,7 +653,7 @@ TEST_CASE("Unit_hipGraphAddMemAllocNode_Functional_2") { * ------------------------ * - HIP_VERSION >= 6.1 */ -TEST_CASE("Unit_hipGraphAddMemAllocNode_Functional_3") { +TEST_CASE("Unit_hipGraphAddMemAllocNode_Functional_3", "[multigpu]") { int mem_pool_support = 0; HIP_CHECK(hipDeviceGetAttribute(&mem_pool_support, hipDeviceAttributeMemoryPoolsSupported, 0)); if (!mem_pool_support) { @@ -727,7 +727,7 @@ TEST_CASE("Unit_hipGraphAddMemAllocNode_Functional_3") { * ------------------------ * - HIP_VERSION >= 6.1 */ -TEST_CASE("Unit_hipGraphAddMemAllocNode_Functional_4") { +TEST_CASE("Unit_hipGraphAddMemAllocNode_Functional_4", "[multigpu]") { int mem_pool_support = 0; HIP_CHECK(hipDeviceGetAttribute(&mem_pool_support, hipDeviceAttributeMemoryPoolsSupported, 0)); if (!mem_pool_support) { diff --git a/projects/hip-tests/catch/unit/graph/hipGraphAddMemcpyNode1D_old.cc b/projects/hip-tests/catch/unit/graph/hipGraphAddMemcpyNode1D_old.cc index 1a9bb608f7..1730f63eb4 100644 --- a/projects/hip-tests/catch/unit/graph/hipGraphAddMemcpyNode1D_old.cc +++ b/projects/hip-tests/catch/unit/graph/hipGraphAddMemcpyNode1D_old.cc @@ -115,7 +115,7 @@ static void validateMemcpyNode1DArray(bool peerAccess, * For Peer device test: Memory allocations happen on device(0) and memcpy operations * are performed from device(1). */ -TEST_CASE("Unit_hipGraphAddMemcpyNode1D_Functional") { +TEST_CASE("Unit_hipGraphAddMemcpyNode1D_Functional", "[multigpu]") { SECTION("Memcpy with 1D array on default device") { validateMemcpyNode1DArray(false); } SECTION("Memcpy with 1D array using DeviceToDeviceNoCU") { validateMemcpyNode1DArray(false, hipMemcpyDeviceToDeviceNoCU); diff --git a/projects/hip-tests/catch/unit/graph/hipGraphAddMemcpyNodeFromSymbol_old.cc b/projects/hip-tests/catch/unit/graph/hipGraphAddMemcpyNodeFromSymbol_old.cc index d0c55f9282..64caeb4f94 100644 --- a/projects/hip-tests/catch/unit/graph/hipGraphAddMemcpyNodeFromSymbol_old.cc +++ b/projects/hip-tests/catch/unit/graph/hipGraphAddMemcpyNodeFromSymbol_old.cc @@ -255,7 +255,8 @@ in GPU-0 and add the MemcpyNodeFromSymbol node to the graph and verifying the result in GPU-1 */ #if HT_NVIDIA -TEST_CASE("Unit_hipGraphAddMemcpyNodeFromSymbol_GlobalMemoryPeerDevice") { +TEST_CASE("Unit_hipGraphAddMemcpyNodeFromSymbol_GlobalMemoryPeerDevice", + "[multigpu]") { int numDevices = 0; int canAccessPeer = 0; if (numDevices > 1) { @@ -276,7 +277,8 @@ in GPU-0 and add the MemcpyNodeFromSymbol node to the graph and verifying the result in GPU-1 */ -TEST_CASE("Unit_hipGraphAddMemcpyNodeFromSymbol_GlobalConstMemoryPeerDevice") { +TEST_CASE("Unit_hipGraphAddMemcpyNodeFromSymbol_GlobalConstMemoryPeerDevice", + "[multigpu]") { int numDevices = 0; int canAccessPeer = 0; if (numDevices > 1) { diff --git a/projects/hip-tests/catch/unit/graph/hipGraphAddMemcpyNodeToSymbol_old.cc b/projects/hip-tests/catch/unit/graph/hipGraphAddMemcpyNodeToSymbol_old.cc index 60edcc02aa..582a93d805 100644 --- a/projects/hip-tests/catch/unit/graph/hipGraphAddMemcpyNodeToSymbol_old.cc +++ b/projects/hip-tests/catch/unit/graph/hipGraphAddMemcpyNodeToSymbol_old.cc @@ -257,7 +257,8 @@ This testcase verifies allocating global const symbol memory and device variable in GPU-0 and add the MemcpyNodeToSymbol node to the graph and verifying the result in GPU-1 */ -TEST_CASE("Unit_hipGraphAddMemcpyNodeToSymbol_GlobalConstMemoryPeerDevice") { +TEST_CASE("Unit_hipGraphAddMemcpyNodeToSymbol_GlobalConstMemoryPeerDevice", + "[multigpu]") { int numDevices = 0; int canAccessPeer = 0; HIP_CHECK(hipGetDeviceCount(&numDevices)); @@ -278,7 +279,8 @@ This testcaser verifies allocating global memory, Add MemcpyToSymbolNode,KernelNode and memcpynode and validating the behaviour */ -TEST_CASE("Unit_hipGraphAddMemcpyNodeToSymbol_MemcpyToSymbolNodeWithKernel") { +TEST_CASE("Unit_hipGraphAddMemcpyNodeToSymbol_MemcpyToSymbolNodeWithKernel", + "[multigpu]") { constexpr size_t Nbytes = SIZE * sizeof(int); constexpr auto blocksPerCU = 6; // to hide latency constexpr auto threadsPerBlock = 256; diff --git a/projects/hip-tests/catch/unit/graph/hipGraphAddMemcpyNode_old.cc b/projects/hip-tests/catch/unit/graph/hipGraphAddMemcpyNode_old.cc index 4ac85de821..b04cff4185 100644 --- a/projects/hip-tests/catch/unit/graph/hipGraphAddMemcpyNode_old.cc +++ b/projects/hip-tests/catch/unit/graph/hipGraphAddMemcpyNode_old.cc @@ -484,7 +484,7 @@ TEST_CASE("Unit_hipGraphAddMemcpyNode_BasicFunctional") { * are performed from device(1). * Tests also verify memcpy node addition with 1D, 2D and 3D objects. */ -TEST_CASE("Unit_hipGraphAddMemcpyNode_PeerAccessFunctional") { +TEST_CASE("Unit_hipGraphAddMemcpyNode_PeerAccessFunctional", "[multigpu]") { CHECK_IMAGE_SUPPORT int numDevices{}, peerAccess{}; diff --git a/projects/hip-tests/catch/unit/graph/hipGraphClone.cc b/projects/hip-tests/catch/unit/graph/hipGraphClone.cc index cb7c7e5e42..4b37f4a642 100644 --- a/projects/hip-tests/catch/unit/graph/hipGraphClone.cc +++ b/projects/hip-tests/catch/unit/graph/hipGraphClone.cc @@ -229,7 +229,7 @@ This testcase verifies following scenarios validate the result of the cloned graph 3. Device context change for cloned graph */ -TEST_CASE("Unit_hipGraphClone_Functional") { +TEST_CASE("Unit_hipGraphClone_Functional", "[multigpu]") { SECTION("hipGraphClone Basic Functionality") { hipGraphClone_Func(); } SECTION("hipGraphClone Modify Original graph") { hipGraphClone_Func(true); } diff --git a/projects/hip-tests/catch/unit/graph/hipGraphCloneComplx.cc b/projects/hip-tests/catch/unit/graph/hipGraphCloneComplx.cc index ada6801727..56bc65d081 100644 --- a/projects/hip-tests/catch/unit/graph/hipGraphCloneComplx.cc +++ b/projects/hip-tests/catch/unit/graph/hipGraphCloneComplx.cc @@ -1490,7 +1490,7 @@ TEST_CASE("Unit_hipGraphClone_Test_hipGraphEventWaitNodeSetEvent_and_Exec") { Execute both original graph and cloned graph in loop: with multiple device. Loop: Update input data -> Launch Graph -> Validate output data -> Goto Loop */ -TEST_CASE("Unit_hipGraphClone_address_change_in_loop") { +TEST_CASE("Unit_hipGraphClone_address_change_in_loop", "[multigpu]") { constexpr size_t Nbytes = N * sizeof(int); constexpr auto blocksPerCU = 6; // to hide latency constexpr auto threadsPerBlock = 256; @@ -1644,7 +1644,7 @@ static void hipGraphClone_address_change_in_thread(hipGraph_t* graph, hipGraphNo memory addresses in each Node and create executable graphs. Launch the graphs in their respective GPUs. Validate the outputs. */ -TEST_CASE("Unit_hipGraphClone_address_change_in_thread") { +TEST_CASE("Unit_hipGraphClone_address_change_in_thread", "[multigpu]") { constexpr size_t Nbytes = N * sizeof(int); constexpr auto blocksPerCU = 6; // to hide latency constexpr auto threadsPerBlock = 256; @@ -1735,7 +1735,7 @@ static void hipGraphClone_Test_All_API(int dev) { Create a graph with Memcpy and Kernel nodes. and its cloned graph. Run all the above writen test cases for multiple GPU scenarios */ -TEST_CASE("Unit_hipGraphClone_multi_GPU_test") { +TEST_CASE("Unit_hipGraphClone_multi_GPU_test", "[multigpu]") { // FIXME: This test tests 3D as well, decouple it CHECK_IMAGE_SUPPORT diff --git a/projects/hip-tests/catch/unit/graph/hipGraphExecEventRecordNodeSetEvent.cc b/projects/hip-tests/catch/unit/graph/hipGraphExecEventRecordNodeSetEvent.cc index e820dc1cdd..fa27fc6370 100644 --- a/projects/hip-tests/catch/unit/graph/hipGraphExecEventRecordNodeSetEvent.cc +++ b/projects/hip-tests/catch/unit/graph/hipGraphExecEventRecordNodeSetEvent.cc @@ -187,7 +187,8 @@ TEST_CASE("Unit_hipGraphExecEventRecordNodeSetEvent_VerifyEventNotChanged") { * Scenario 3: This test verifies event in node of the executable graph can be changed to event on * different device */ -TEST_CASE("Unit_hipGraphExecEventRecordNodeSetEvent_Positive_DifferentDevices") { +TEST_CASE("Unit_hipGraphExecEventRecordNodeSetEvent_Positive_DifferentDevices", + "[multigpu]") { const auto device_count = HipTest::getDeviceCount(); if (device_count < 2) { HipTest::HIP_SKIP_TEST("Skipping because devices < 2"); diff --git a/projects/hip-tests/catch/unit/graph/hipGraphExecMemcpyNodeSetParamsToSymbol.cc b/projects/hip-tests/catch/unit/graph/hipGraphExecMemcpyNodeSetParamsToSymbol.cc index 0d4821c968..d895b2fd8b 100644 --- a/projects/hip-tests/catch/unit/graph/hipGraphExecMemcpyNodeSetParamsToSymbol.cc +++ b/projects/hip-tests/catch/unit/graph/hipGraphExecMemcpyNodeSetParamsToSymbol.cc @@ -142,7 +142,8 @@ TEST_CASE("Unit_hipGraphExecMemcpyNodeSetParamsToSymbol_Positive_Basic") { * ------------------------ * - HIP_VERSION >= 5.2 */ -TEST_CASE("Unit_hipGraphExecMemcpyNodeSetParamsToSymbol_Negative_Parameters") { +TEST_CASE("Unit_hipGraphExecMemcpyNodeSetParamsToSymbol_Negative_Parameters", + "[multigpu]") { using namespace std::placeholders; hipGraph_t graph = nullptr; HIP_CHECK(hipGraphCreate(&graph, 0)); diff --git a/projects/hip-tests/catch/unit/graph/hipGraphExecMemsetNodeSetParams.cc b/projects/hip-tests/catch/unit/graph/hipGraphExecMemsetNodeSetParams.cc index 95226bf944..d8a7594b29 100644 --- a/projects/hip-tests/catch/unit/graph/hipGraphExecMemsetNodeSetParams.cc +++ b/projects/hip-tests/catch/unit/graph/hipGraphExecMemsetNodeSetParams.cc @@ -129,7 +129,8 @@ TEMPLATE_TEST_CASE("Unit_hipGraphExecMemsetNodeSetParams_Positive_Basic", "", ui * ------------------------ * - HIP_VERSION >= 5.2 */ -TEST_CASE("Unit_hipGraphExecMemsetNodeSetParams_Negative_Parameters") { +TEST_CASE("Unit_hipGraphExecMemsetNodeSetParams_Negative_Parameters", + "[multigpu]") { // FIXME: this test tests 1D/2D/3D stuff in one single go, need to decouple it so that it can run // on devices with no image support CHECK_IMAGE_SUPPORT diff --git a/projects/hip-tests/catch/unit/graph/hipGraphExecUpdate.cc b/projects/hip-tests/catch/unit/graph/hipGraphExecUpdate.cc index 3616bd36aa..fb053e177f 100644 --- a/projects/hip-tests/catch/unit/graph/hipGraphExecUpdate.cc +++ b/projects/hip-tests/catch/unit/graph/hipGraphExecUpdate.cc @@ -639,7 +639,8 @@ TEST_CASE("Unit_hipGraphExecUpdate_Negative_NodeType_Changed") { * - HIP_VERSION >= 6.0 */ -TEST_CASE("Unit_hipGraphExecUpdate_Negative_MultiDevice_Context_Changed") { +TEST_CASE("Unit_hipGraphExecUpdate_Negative_MultiDevice_Context_Changed", + "[multigpu]") { constexpr size_t N = 1024; constexpr size_t Nbytes = N * sizeof(int); constexpr auto blocksPerCU = 6; // to hide latency diff --git a/projects/hip-tests/catch/unit/graph/hipGraphInstantiateWithFlags.cc b/projects/hip-tests/catch/unit/graph/hipGraphInstantiateWithFlags.cc index f44e712c52..177e10d5b0 100644 --- a/projects/hip-tests/catch/unit/graph/hipGraphInstantiateWithFlags.cc +++ b/projects/hip-tests/catch/unit/graph/hipGraphInstantiateWithFlags.cc @@ -269,7 +269,8 @@ This testcase verifies hipGraphInstantiateWithFlags API by creating dependency graph on GPU-0 and instantiate, launching and verifying the result on GPU-1 */ -TEST_CASE("Unit_hipGraphInstantiateWithFlags_DependencyGraphDeviceCtxtChg") { +TEST_CASE("Unit_hipGraphInstantiateWithFlags_DependencyGraphDeviceCtxtChg", + "[multigpu]") { int numDevices = 0; int canAccessPeer = 0; HIP_CHECK(hipGetDeviceCount(&numDevices)); @@ -311,7 +312,8 @@ This testcase verifies hipGraphInstantiateWithFlags API by creating capture graph on GPU-0 and instantiate, launching and verifying the result on GPU-1 */ -TEST_CASE("Unit_hipGraphInstantiateWithFlags_StreamCaptureDeviceContextChg") { +TEST_CASE("Unit_hipGraphInstantiateWithFlags_StreamCaptureDeviceContextChg", + "[multigpu]") { int numDevices = 0; int canAccessPeer = 0; HIP_CHECK(hipGetDeviceCount(&numDevices)); diff --git a/projects/hip-tests/catch/unit/graph/hipGraphLaunch_old.cc b/projects/hip-tests/catch/unit/graph/hipGraphLaunch_old.cc index c2cf5708fe..a0b70b9245 100644 --- a/projects/hip-tests/catch/unit/graph/hipGraphLaunch_old.cc +++ b/projects/hip-tests/catch/unit/graph/hipGraphLaunch_old.cc @@ -254,7 +254,7 @@ static void hipGraphLaunch_test() { HIP_CHECK(hipStreamDestroy(streamForGraph)); } -TEST_CASE("Unit_hipGraphLaunch_Functional_multidevice_test") { +TEST_CASE("Unit_hipGraphLaunch_Functional_multidevice_test", "[multigpu]") { int numDevices = 0; HIP_CHECK(hipGetDeviceCount(&numDevices)); diff --git a/projects/hip-tests/catch/unit/graph/hipGraphMemAllocNodeGetParams.cc b/projects/hip-tests/catch/unit/graph/hipGraphMemAllocNodeGetParams.cc index ef0a596b6e..37fd88746e 100644 --- a/projects/hip-tests/catch/unit/graph/hipGraphMemAllocNodeGetParams.cc +++ b/projects/hip-tests/catch/unit/graph/hipGraphMemAllocNodeGetParams.cc @@ -135,7 +135,8 @@ TEST_CASE("Unit_hipGraphMem_Alloc_Free_NodeGetParams_Functional") { hipGraphMemAllocNodeGetParams_Functional(); } -TEST_CASE("Unit_hipGraphMem_Alloc_Free_NodeGetParams_Functional_MultiDevice") { +TEST_CASE("Unit_hipGraphMem_Alloc_Free_NodeGetParams_Functional_MultiDevice", + "[multigpu]") { int numDevices = 0; HIP_CHECK(hipGetDeviceCount(&numDevices)); diff --git a/projects/hip-tests/catch/unit/graph/hipGraphPerf.cc b/projects/hip-tests/catch/unit/graph/hipGraphPerf.cc index 990c0c8216..55c292d285 100644 --- a/projects/hip-tests/catch/unit/graph/hipGraphPerf.cc +++ b/projects/hip-tests/catch/unit/graph/hipGraphPerf.cc @@ -614,7 +614,8 @@ static void hipGraph_PerfCheck_hipGraphExecKernelNodeSetParams(const hipStream_t * ------------------------ * - HIP_VERSION >= 6.1 */ -TEST_CASE("Unit_hipGraph_PerfCheck_hipGraphExecKernelNodeSetParams") { +TEST_CASE("Unit_hipGraph_PerfCheck_hipGraphExecKernelNodeSetParams", + "[multigpu]") { if ((setenv("DEBUG_CLR_GRAPH_PACKET_CAPTURE", "true", 1)) != 0) { HipTest::HIP_SKIP_TEST( "Unable to turn on " @@ -732,7 +733,8 @@ static void hipGraph_PerfCheck_hipGraphExecKernelNodeSetParams_inLoop(const hipS * ------------------------ * - HIP_VERSION >= 6.1 */ -TEST_CASE("Unit_hipGraph_PerfCheck_hipGraphExecKernelNodeSetParams_inLoop") { +TEST_CASE("Unit_hipGraph_PerfCheck_hipGraphExecKernelNodeSetParams_inLoop", + "[multigpu]") { if ((setenv("DEBUG_CLR_GRAPH_PACKET_CAPTURE", "true", 1)) != 0) { HipTest::HIP_SKIP_TEST( "Unable to turn on " @@ -954,7 +956,8 @@ static void hipGraph_PerfCheck_hipGraphExecMemcpyNodeSetParams_inLoop(const hipS * - HIP_VERSION >= 6.1 */ -TEST_CASE("Unit_hipGraph_PerfCheck_hipGraphExecMemcpyNodeSetParams_inLoop") { +TEST_CASE("Unit_hipGraph_PerfCheck_hipGraphExecMemcpyNodeSetParams_inLoop", + "[multigpu]") { if ((setenv("DEBUG_CLR_GRAPH_PACKET_CAPTURE", "true", 1)) != 0) { HipTest::HIP_SKIP_TEST( "Unable to turn on " @@ -1070,7 +1073,8 @@ static void hipGraph_PerfCheck_hipGraphExecMemcpyNodeSetParams1D_inLoop(const hi * - HIP_VERSION >= 6.1 */ -TEST_CASE("Unit_hipGraph_PerfCheck_hipGraphExecMemcpyNodeSetParams1D_inLoop") { +TEST_CASE("Unit_hipGraph_PerfCheck_hipGraphExecMemcpyNodeSetParams1D_inLoop", + "[multigpu]") { if ((setenv("DEBUG_CLR_GRAPH_PACKET_CAPTURE", "true", 1)) != 0) { HipTest::HIP_SKIP_TEST( "Unable to turn on " @@ -1178,7 +1182,8 @@ static void hipGraph_PerfCheck_hipGraphExecMemcpyNodeSetParamsFrmSymbol(const hi * - HIP_VERSION >= 6.1 */ -TEST_CASE("Unit_hipGraph_PerfCheck_hipGraphExecMemcpyNodeSetParamsFrmSymbol") { +TEST_CASE("Unit_hipGraph_PerfCheck_hipGraphExecMemcpyNodeSetParamsFrmSymbol", + "[multigpu]") { if ((setenv("DEBUG_CLR_GRAPH_PACKET_CAPTURE", "true", 1)) != 0) { HipTest::HIP_SKIP_TEST( "Unable to turn on " @@ -1285,7 +1290,8 @@ static void hipGraph_PerfCheck_hipGraphExecMemcpyNodeSetParamsToSymbol(const hip * ------------------------ * - HIP_VERSION >= 6.1 */ -TEST_CASE("Unit_hipGraph_PerfCheck_hipGraphExecMemcpyNodeSetParamsToSymbol") { +TEST_CASE("Unit_hipGraph_PerfCheck_hipGraphExecMemcpyNodeSetParamsToSymbol", + "[multigpu]") { if ((setenv("DEBUG_CLR_GRAPH_PACKET_CAPTURE", "true", 1)) != 0) { HipTest::HIP_SKIP_TEST( "Unable to turn on " @@ -1438,7 +1444,8 @@ static void hipGraph_PerfCheck_hipGraphExecMemsetNodeSetParams(const hipStream_t * ------------------------ * - HIP_VERSION >= 6.1 */ -TEST_CASE("Unit_hipGraph_PerfCheck_hipGraphExecMemsetNodeSetParams") { +TEST_CASE("Unit_hipGraph_PerfCheck_hipGraphExecMemsetNodeSetParams", + "[multigpu]") { if ((setenv("DEBUG_CLR_GRAPH_PACKET_CAPTURE", "true", 1)) != 0) { HipTest::HIP_SKIP_TEST( "Unable to turn on " @@ -1871,7 +1878,8 @@ static void hipGraph_PerfCheck_hipGraphExecChildGraphNodeSetParams_mKernel( * ------------------------ * - HIP_VERSION >= 6.1 */ -TEST_CASE("Unit_hipGraph_PerfCheck_hipGraphExecChildGraphNodeSetParams") { +TEST_CASE("Unit_hipGraph_PerfCheck_hipGraphExecChildGraphNodeSetParams", + "[multigpu]") { if ((setenv("DEBUG_CLR_GRAPH_PACKET_CAPTURE", "true", 1)) != 0) { HipTest::HIP_SKIP_TEST( "Unable to turn on " @@ -2018,7 +2026,8 @@ static void hipGraph_PerfCheck_hipGraphExecEventRecordNodeSetEvent(const hipStre * ------------------------ * - HIP_VERSION >= 6.1 */ -TEST_CASE("Unit_hipGraph_PerfCheck_hipGraphExecEventRecordNodeSetEvent") { +TEST_CASE("Unit_hipGraph_PerfCheck_hipGraphExecEventRecordNodeSetEvent", + "[multigpu]") { if ((setenv("DEBUG_CLR_GRAPH_PACKET_CAPTURE", "true", 1)) != 0) { HipTest::HIP_SKIP_TEST( "Unable to turn on " @@ -2205,7 +2214,8 @@ static void hipGraph_PerfCheck_hipGraphExecEventWaitNodeSetEvent(const hipStream * ------------------------ * - HIP_VERSION >= 6.1 */ -TEST_CASE("Unit_hipGraph_PerfCheck_hipGraphExecEventWaitNodeSetEvent") { +TEST_CASE("Unit_hipGraph_PerfCheck_hipGraphExecEventWaitNodeSetEvent", + "[multigpu]") { if ((setenv("DEBUG_CLR_GRAPH_PACKET_CAPTURE", "true", 1)) != 0) { HipTest::HIP_SKIP_TEST( "Unable to turn on " @@ -2359,7 +2369,8 @@ static void hipGraph_PerfCheck_hipGraphExecHostNodeSetParams(const hipStream_t& * ------------------------ * - HIP_VERSION >= 6.1 */ -TEST_CASE("Unit_hipGraph_PerfCheck_hipGraphExecHostNodeSetParams") { +TEST_CASE("Unit_hipGraph_PerfCheck_hipGraphExecHostNodeSetParams", + "[multigpu]") { if ((setenv("DEBUG_CLR_GRAPH_PACKET_CAPTURE", "true", 1)) != 0) { HipTest::HIP_SKIP_TEST( "Unable to turn on " @@ -2482,7 +2493,7 @@ static void hipGraph_PerfCheck_hipGraphExecUpdate(const hipStream_t& stream) { * ------------------------ * - HIP_VERSION >= 6.1 */ -TEST_CASE("Unit_hipGraph_PerfCheck_hipGraphExecUpdate") { +TEST_CASE("Unit_hipGraph_PerfCheck_hipGraphExecUpdate", "[multigpu]") { if ((setenv("DEBUG_CLR_GRAPH_PACKET_CAPTURE", "true", 1)) != 0) { HipTest::HIP_SKIP_TEST( "Unable to turn on " @@ -2626,7 +2637,8 @@ static void hipGraph_PerfCheck_hipGraphExecUpdate_kernel_inLoop(const hipStream_ * ------------------------ * - HIP_VERSION >= 6.1 */ -TEST_CASE("Unit_hipGraph_PerfCheck_hipGraphExecUpdate_kernel_inLoop") { +TEST_CASE("Unit_hipGraph_PerfCheck_hipGraphExecUpdate_kernel_inLoop", + "[multigpu]") { if ((setenv("DEBUG_CLR_GRAPH_PACKET_CAPTURE", "true", 1)) != 0) { HipTest::HIP_SKIP_TEST( "Unable to turn on " diff --git a/projects/hip-tests/catch/unit/graph/hipGraphUpload.cc b/projects/hip-tests/catch/unit/graph/hipGraphUpload.cc index b093dfa237..343412936b 100644 --- a/projects/hip-tests/catch/unit/graph/hipGraphUpload.cc +++ b/projects/hip-tests/catch/unit/graph/hipGraphUpload.cc @@ -152,7 +152,7 @@ TEST_CASE("Unit_hipGraphUpload_Functional") { } } -TEST_CASE("Unit_hipGraphUpload_Functional_multidevice_test") { +TEST_CASE("Unit_hipGraphUpload_Functional_multidevice_test", "[multigpu]") { int numDevices = 0; HIP_CHECK(hipGetDeviceCount(&numDevices)); diff --git a/projects/hip-tests/catch/unit/graph/hipStreamBeginCapture.cc b/projects/hip-tests/catch/unit/graph/hipStreamBeginCapture.cc index cb70192049..23da547f39 100644 --- a/projects/hip-tests/catch/unit/graph/hipStreamBeginCapture.cc +++ b/projects/hip-tests/catch/unit/graph/hipStreamBeginCapture.cc @@ -1064,7 +1064,7 @@ TEST_CASE("Unit_hipStreamBeginCapture_Negative_EndingCapwhenCapInProg") { * ------------------------ * - HIP_VERSION >= 5.2 */ -TEST_CASE("Unit_hipStreamBeginCapture_Positive_MultiGPU") { +TEST_CASE("Unit_hipStreamBeginCapture_Positive_MultiGPU", "[multigpu]") { int devcount = 0; HIP_CHECK(hipGetDeviceCount(&devcount)); // If only single GPU is detected then return diff --git a/projects/hip-tests/catch/unit/graph/hipStreamBeginCapture_old.cc b/projects/hip-tests/catch/unit/graph/hipStreamBeginCapture_old.cc index 86cf9a59a0..bc7ddd7516 100644 --- a/projects/hip-tests/catch/unit/graph/hipStreamBeginCapture_old.cc +++ b/projects/hip-tests/catch/unit/graph/hipStreamBeginCapture_old.cc @@ -931,7 +931,7 @@ TEST_CASE("Unit_hipStreamBeginCapture_EndingCapturewhenCaptureInProgress") { /* Test scenario 15 */ -TEST_CASE("Unit_hipStreamBeginCapture_MultiGPU") { +TEST_CASE("Unit_hipStreamBeginCapture_MultiGPU", "[multigpu]") { int devcount = 0; HIP_CHECK(hipGetDeviceCount(&devcount)); // If only single GPU is detected then return diff --git a/projects/hip-tests/catch/unit/kernel/hipPrintfKernel.cc b/projects/hip-tests/catch/unit/kernel/hipPrintfKernel.cc index 35ac895518..90f7424488 100644 --- a/projects/hip-tests/catch/unit/kernel/hipPrintfKernel.cc +++ b/projects/hip-tests/catch/unit/kernel/hipPrintfKernel.cc @@ -45,7 +45,7 @@ __global__ void run_printf() { printf("Hello World\n"); } * ------------------------ * - HIP_VERSION >= 5.6 */ -TEST_CASE("Unit_kernel_ChkPrintf") { +TEST_CASE("Unit_kernel_ChkPrintf", "[multigpu]") { int device_count = 0; CaptureStream capture(stdout); HIP_CHECK(hipGetDeviceCount(&device_count)); diff --git a/projects/hip-tests/catch/unit/memory/hipArrayCreate.cc b/projects/hip-tests/catch/unit/memory/hipArrayCreate.cc index 269f6252ec..609048eb56 100644 --- a/projects/hip-tests/catch/unit/memory/hipArrayCreate.cc +++ b/projects/hip-tests/catch/unit/memory/hipArrayCreate.cc @@ -85,7 +85,7 @@ This testcase verifies the hipArrayCreate API in multithreaded scenario by launching threads in parallel on multiple GPUs and verifies the hipArrayCreate API with small and big chunks data */ -TEST_CASE("Unit_hipArrayCreate_MultiThread") { +TEST_CASE("Unit_hipArrayCreate_MultiThread", "[multigpu]") { CHECK_IMAGE_SUPPORT std::vector threadlist; diff --git a/projects/hip-tests/catch/unit/memory/hipArrayGetDescriptor.cc b/projects/hip-tests/catch/unit/memory/hipArrayGetDescriptor.cc index 4b5a9e9d7b..130bcdf5b2 100644 --- a/projects/hip-tests/catch/unit/memory/hipArrayGetDescriptor.cc +++ b/projects/hip-tests/catch/unit/memory/hipArrayGetDescriptor.cc @@ -233,7 +233,7 @@ float* funcToChkArray(hipArray_t array) { * ------------------------ * - HIP_VERSION >= 5.6 */ -TEST_CASE("Unit_hipArrayGetDescriptor_1D_2D_ArrayParameterChk") { +TEST_CASE("Unit_hipArrayGetDescriptor_1D_2D_ArrayParameterChk", "[multigpu]") { CHECK_IMAGE_SUPPORT int numDevices = 0; @@ -318,7 +318,8 @@ TEST_CASE("Unit_hipArrayGetDescriptor_1D_2D_ArrayParameterChk") { * ------------------------ * - HIP_VERSION >= 5.6 */ -TEST_CASE("Unit_hipArrayGetDescriptor_MultiThreadScenarioFor1D_2D_Array") { +TEST_CASE("Unit_hipArrayGetDescriptor_MultiThreadScenarioFor1D_2D_Array", + "[multigpu]") { CHECK_IMAGE_SUPPORT int numDevices = 0; @@ -368,7 +369,7 @@ TEST_CASE("Unit_hipArrayGetDescriptor_MultiThreadScenarioFor1D_2D_Array") { * ------------------------ * - HIP_VERSION >= 5.6 */ -TEST_CASE("Unit_hipArrayGetDescriptor_Host2Array_Array2Host") { +TEST_CASE("Unit_hipArrayGetDescriptor_Host2Array_Array2Host", "[multigpu]") { CHECK_IMAGE_SUPPORT int numDevices = 0; diff --git a/projects/hip-tests/catch/unit/memory/hipDeviceGetMemPool.cc b/projects/hip-tests/catch/unit/memory/hipDeviceGetMemPool.cc index 265482e794..d1a62ba333 100644 --- a/projects/hip-tests/catch/unit/memory/hipDeviceGetMemPool.cc +++ b/projects/hip-tests/catch/unit/memory/hipDeviceGetMemPool.cc @@ -137,7 +137,7 @@ TEST_CASE("Unit_hipDeviceGetMemPool_Functional") { * ------------------------ * - HIP_VERSION >= 6.2 */ -TEST_CASE("Unit_hipDeviceGetMemPool_Multidevice") { +TEST_CASE("Unit_hipDeviceGetMemPool_Multidevice", "[multigpu]") { int num_devices; HIP_CHECK(hipGetDeviceCount(&num_devices)); diff --git a/projects/hip-tests/catch/unit/memory/hipDeviceSetMemPool.cc b/projects/hip-tests/catch/unit/memory/hipDeviceSetMemPool.cc index eb88338100..689bdf8486 100644 --- a/projects/hip-tests/catch/unit/memory/hipDeviceSetMemPool.cc +++ b/projects/hip-tests/catch/unit/memory/hipDeviceSetMemPool.cc @@ -108,7 +108,7 @@ TEST_CASE("Unit_hipDeviceSetMemPool_Basic") { * ------------------------ * - HIP_VERSION >= 6.2 */ -TEST_CASE("Unit_hipDeviceSetMemPool_DestroyCurrentMempool") { +TEST_CASE("Unit_hipDeviceSetMemPool_DestroyCurrentMempool", "[multigpu]") { int num_devices; HIP_CHECK(hipGetDeviceCount(&num_devices)); for (int dev = 0; dev < num_devices; dev++) { diff --git a/projects/hip-tests/catch/unit/memory/hipDrvMemcpy2DUnaligned.cc b/projects/hip-tests/catch/unit/memory/hipDrvMemcpy2DUnaligned.cc index 6d2af7922f..abf470a7a8 100644 --- a/projects/hip-tests/catch/unit/memory/hipDrvMemcpy2DUnaligned.cc +++ b/projects/hip-tests/catch/unit/memory/hipDrvMemcpy2DUnaligned.cc @@ -269,7 +269,7 @@ TEST_CASE("Unit_hipDrvMemcpy2DUnaligned_FuncTst") { * ------------------------ * - HIP_VERSION >= 6.0 */ -TEST_CASE("Unit_hipDrvMemcpy2DUnaligned_Positive_Basic") { +TEST_CASE("Unit_hipDrvMemcpy2DUnaligned_Positive_Basic", "[multigpu]") { CHECK_IMAGE_SUPPORT SECTION("Device to Device") { diff --git a/projects/hip-tests/catch/unit/memory/hipDrvMemcpy3DAsync_old.cc b/projects/hip-tests/catch/unit/memory/hipDrvMemcpy3DAsync_old.cc index b866626c7f..768dd1008c 100644 --- a/projects/hip-tests/catch/unit/memory/hipDrvMemcpy3DAsync_old.cc +++ b/projects/hip-tests/catch/unit/memory/hipDrvMemcpy3DAsync_old.cc @@ -542,7 +542,7 @@ TEST_CASE("Unit_hipDrvMemcpy3DAsync_ExtentValidation") { * - HIP_VERSION >= 6.0 */ -TEST_CASE("Unit_hipDrvMemcpy3DAsync_H2DDeviceContextChange") { +TEST_CASE("Unit_hipDrvMemcpy3DAsync_H2DDeviceContextChange", "[multigpu]") { CHECK_IMAGE_SUPPORT int numDevices = 0; HIP_CHECK(hipGetDeviceCount(&numDevices)); @@ -567,7 +567,8 @@ TEST_CASE("Unit_hipDrvMemcpy3DAsync_H2DDeviceContextChange") { * - HIP_VERSION >= 6.0 */ -TEST_CASE("Unit_hipDrvMemcpy3DAsync_Host2ArrayDeviceContextChange") { +TEST_CASE("Unit_hipDrvMemcpy3DAsync_Host2ArrayDeviceContextChange", + "[multigpu]") { CHECK_IMAGE_SUPPORT int numDevices = 0; HIP_CHECK(hipGetDeviceCount(&numDevices)); @@ -595,7 +596,8 @@ TEST_CASE("Unit_hipDrvMemcpy3DAsync_Host2ArrayDeviceContextChange") { * - HIP_VERSION >= 6.0 */ -TEST_CASE("Unit_hipDrvMemcpy3DAsync_multiDevice_Basic_Size_Test") { +TEST_CASE("Unit_hipDrvMemcpy3DAsync_multiDevice_Basic_Size_Test", + "[multigpu]") { CHECK_IMAGE_SUPPORT constexpr int size_128b = 128, size_256b = 256; int numDevices = 0; diff --git a/projects/hip-tests/catch/unit/memory/hipDrvMemcpy3D_old.cc b/projects/hip-tests/catch/unit/memory/hipDrvMemcpy3D_old.cc index 2e1e011b44..d27e32b283 100644 --- a/projects/hip-tests/catch/unit/memory/hipDrvMemcpy3D_old.cc +++ b/projects/hip-tests/catch/unit/memory/hipDrvMemcpy3D_old.cc @@ -524,7 +524,7 @@ TEST_CASE("Unit_hipDrvMemcpy3D_ExtentValidation") { * - HIP_VERSION >= 6.0 */ -TEST_CASE("Unit_hipDrvMemcpy3D_H2DDeviceContextChange") { +TEST_CASE("Unit_hipDrvMemcpy3D_H2DDeviceContextChange", "[multigpu]") { CHECK_IMAGE_SUPPORT int numDevices = 0; HIP_CHECK(hipGetDeviceCount(&numDevices)); @@ -549,7 +549,7 @@ TEST_CASE("Unit_hipDrvMemcpy3D_H2DDeviceContextChange") { * - HIP_VERSION >= 6.0 */ -TEST_CASE("Unit_hipDrvMemcpy3D_Host2ArrayDeviceContextChange") { +TEST_CASE("Unit_hipDrvMemcpy3D_Host2ArrayDeviceContextChange", "[multigpu]") { CHECK_IMAGE_SUPPORT int numDevices = 0; HIP_CHECK(hipGetDeviceCount(&numDevices)); @@ -577,7 +577,7 @@ TEST_CASE("Unit_hipDrvMemcpy3D_Host2ArrayDeviceContextChange") { * - HIP_VERSION >= 6.0 */ -TEST_CASE("Unit_hipDrvMemcpy3D_multiDevice_Basic_Size_Test") { +TEST_CASE("Unit_hipDrvMemcpy3D_multiDevice_Basic_Size_Test", "[multigpu]") { CHECK_IMAGE_SUPPORT constexpr int size_128b = 128, size_256b = 256; int numDevices = 0; diff --git a/projects/hip-tests/catch/unit/memory/hipGetProcAddressMemoryApis.cc b/projects/hip-tests/catch/unit/memory/hipGetProcAddressMemoryApis.cc index 4ff37b23bb..56348f163d 100644 --- a/projects/hip-tests/catch/unit/memory/hipGetProcAddressMemoryApis.cc +++ b/projects/hip-tests/catch/unit/memory/hipGetProcAddressMemoryApis.cc @@ -2705,7 +2705,7 @@ TEST_CASE("Unit_hipGetProcAddress_MemoryApisGetMemInfoRelated") { * ------------------------ * - HIP_VERSION >= 6.2 */ -TEST_CASE("Unit_hipGetProcAddress_MemoryApisMemcpy2DRelated") { +TEST_CASE("Unit_hipGetProcAddress_MemoryApisMemcpy2DRelated", "[multigpu]") { CHECK_IMAGE_SUPPORT void* hipMemcpy2D_ptr = nullptr; @@ -6008,7 +6008,7 @@ TEST_CASE("Unit_hipGetProcAddress_MemoryApisStreamOrderedMemory") { * ------------------------ * - HIP_VERSION >= 6.2 */ -TEST_CASE("Unit_hipGetProcAddress_MemoryApisPeerToPeer") { +TEST_CASE("Unit_hipGetProcAddress_MemoryApisPeerToPeer", "[multigpu]") { int deviceCount = 0; HIP_CHECK(hipGetDeviceCount(&deviceCount)); diff --git a/projects/hip-tests/catch/unit/memory/hipHostRegister.cc b/projects/hip-tests/catch/unit/memory/hipHostRegister.cc index 37c91dac2b..87dc32ceaa 100644 --- a/projects/hip-tests/catch/unit/memory/hipHostRegister.cc +++ b/projects/hip-tests/catch/unit/memory/hipHostRegister.cc @@ -107,7 +107,8 @@ void doMemCopy(size_t numElements, int offset, T* A, T* Bh, T* Bd, bool internal * ------------------------ * - HIP_VERSION >= 5.2 */ -TEMPLATE_TEST_CASE("Unit_hipHostRegister_ReferenceFromKernelandhipMemset", "", int, float, double) { +TEMPLATE_TEST_CASE("Unit_hipHostRegister_ReferenceFromKernelandhipMemset", + "[multigpu]", int, float, double) { size_t sizeBytes{LEN * sizeof(TestType)}; TestType *A, **Ad; int num_devices = 0; @@ -214,7 +215,8 @@ TEMPLATE_TEST_CASE("Unit_hipHostRegister_DirectReferenceFromKernel", "", int, fl * ------------------------ * - HIP_VERSION >= 5.6 */ -TEMPLATE_TEST_CASE("Unit_hipHostRegister_DirectReferenceMultGpu", "", int, float, double) { +TEMPLATE_TEST_CASE("Unit_hipHostRegister_DirectReferenceMultGpu", "[multigpu]", + int, float, double) { // 1 refers to doing hipHostRegister once for all devices // 0 refers to doing hipHostRegister for each device auto register_once = GENERATE(0, 1); diff --git a/projects/hip-tests/catch/unit/memory/hipMalloc3D.cc b/projects/hip-tests/catch/unit/memory/hipMalloc3D.cc index 5e33a3aba0..8844c06392 100644 --- a/projects/hip-tests/catch/unit/memory/hipMalloc3D.cc +++ b/projects/hip-tests/catch/unit/memory/hipMalloc3D.cc @@ -127,7 +127,7 @@ This testcase verifies the hipMalloc3D API in multithreaded scenario by launching threads in parallel on multiple GPUs and verifies the hipMalloc3D API with small and big chunks data */ -TEST_CASE("Unit_hipMalloc3D_MultiThread") { +TEST_CASE("Unit_hipMalloc3D_MultiThread", "[multigpu]") { CHECK_IMAGE_SUPPORT std::vector threadlist; diff --git a/projects/hip-tests/catch/unit/memory/hipMalloc3DArray.cc b/projects/hip-tests/catch/unit/memory/hipMalloc3DArray.cc index 35bd0dff86..b2506f44cb 100644 --- a/projects/hip-tests/catch/unit/memory/hipMalloc3DArray.cc +++ b/projects/hip-tests/catch/unit/memory/hipMalloc3DArray.cc @@ -78,7 +78,7 @@ This testcase verifies the hipMalloc3DArray API in multithreaded scenario by launching threads in parallel on multiple GPUs and verifies the hipMalloc3DArray API with small and big chunks data */ -TEST_CASE("Unit_hipMalloc3DArray_MultiThread") { +TEST_CASE("Unit_hipMalloc3DArray_MultiThread", "[multigpu]") { CHECK_IMAGE_SUPPORT std::vector threadlist; diff --git a/projects/hip-tests/catch/unit/memory/hipMallocArray.cc b/projects/hip-tests/catch/unit/memory/hipMallocArray.cc index 2d8d55dec6..142de1d28e 100644 --- a/projects/hip-tests/catch/unit/memory/hipMallocArray.cc +++ b/projects/hip-tests/catch/unit/memory/hipMallocArray.cc @@ -80,7 +80,7 @@ This testcase verifies the hipMallocArray API in multithreaded scenario by launching threads in parallel on multiple GPUs and verifies the hipMallocArray API with small and big chunks data */ -TEST_CASE("Unit_hipMallocArray_MultiThread") { +TEST_CASE("Unit_hipMallocArray_MultiThread", "[multigpu]") { CHECK_IMAGE_SUPPORT std::vector threadlist; diff --git a/projects/hip-tests/catch/unit/memory/hipMallocAsync.cc b/projects/hip-tests/catch/unit/memory/hipMallocAsync.cc index ea25d19a58..f4bd8cedc0 100644 --- a/projects/hip-tests/catch/unit/memory/hipMallocAsync.cc +++ b/projects/hip-tests/catch/unit/memory/hipMallocAsync.cc @@ -290,7 +290,7 @@ TEST_CASE("Unit_hipMallocAsync_StreamEvent_CrissCross") { * ------------------------ * - HIP_VERSION >= 6.2 */ -TEST_CASE("Unit_hipMallocAsync_Multidevice") { +TEST_CASE("Unit_hipMallocAsync_Multidevice", "[multigpu]") { int num_devices; HIP_CHECK(hipGetDeviceCount(&num_devices)); for (int i = 0; i < num_devices; i++) { @@ -330,7 +330,7 @@ static void threadQAsyncCommands(streamMemAllocTest* testObj, hipStream_t strm, testObj->freeDevBuf(strm); } -TEST_CASE("Unit_hipMallocAsync_Multidevice_Concurrent") { +TEST_CASE("Unit_hipMallocAsync_Multidevice_Concurrent", "[multigpu]") { int num_devices; HIP_CHECK(hipGetDeviceCount(&num_devices)); checkIfMultiDev(num_devices) hipStream_t* stream_buf = new hipStream_t[num_devices]; @@ -379,7 +379,7 @@ TEST_CASE("Unit_hipMallocAsync_Multidevice_Concurrent") { * ------------------------ * - HIP_VERSION >= 6.2 */ -TEST_CASE("Unit_hipMallocAsync_Multidevice_MultiStream") { +TEST_CASE("Unit_hipMallocAsync_Multidevice_MultiStream", "[multigpu]") { int num_devices; HIP_CHECK(hipGetDeviceCount(&num_devices)); checkIfMultiDev(num_devices) diff --git a/projects/hip-tests/catch/unit/memory/hipMallocConcurrency.cc b/projects/hip-tests/catch/unit/memory/hipMallocConcurrency.cc index 7743aeb664..1e347a51ba 100644 --- a/projects/hip-tests/catch/unit/memory/hipMallocConcurrency.cc +++ b/projects/hip-tests/catch/unit/memory/hipMallocConcurrency.cc @@ -303,7 +303,7 @@ TEST_CASE("Unit_hipMalloc_AllocateAndPoolBuffers") { * Exercise hipMalloc() api parellely on all gpus from * multiple threads and regress the api. */ -TEST_CASE("Unit_hipMalloc_Multithreaded_MultiGPU") { +TEST_CASE("Unit_hipMalloc_Multithreaded_MultiGPU", "[multigpu]") { std::vector threadlist; int devCnt; diff --git a/projects/hip-tests/catch/unit/memory/hipMallocFromPoolAsync.cc b/projects/hip-tests/catch/unit/memory/hipMallocFromPoolAsync.cc index b1b5a84179..b0eca5a6a0 100644 --- a/projects/hip-tests/catch/unit/memory/hipMallocFromPoolAsync.cc +++ b/projects/hip-tests/catch/unit/memory/hipMallocFromPoolAsync.cc @@ -340,7 +340,7 @@ TEST_CASE("Unit_hipMallocFromPoolAsync_hipStreamPerThread") { * ------------------------ * - HIP_VERSION >= 6.2 */ -TEST_CASE("Unit_hipMallocFromPoolAsync_ReleaseThreshold_Mgpu") { +TEST_CASE("Unit_hipMallocFromPoolAsync_ReleaseThreshold_Mgpu", "[multigpu]") { constexpr int N = 1 << 20; int numDevices = 0; HIP_CHECK(hipGetDeviceCount(&numDevices)); @@ -575,7 +575,7 @@ static bool checkReuseAllowOtherFlags(int N, hipMemPoolAttr attr, enum eTestValu * - HIP_VERSION >= 6.2 */ #if HT_AMD -TEST_CASE("Unit_hipMallocFromPoolAsync_Multidevice_Concurrent") { +TEST_CASE("Unit_hipMallocFromPoolAsync_Multidevice_Concurrent", "[multigpu]") { auto testType = GENERATE(testdefault, testMaximum); constexpr int N = 1 << 20; int num_devices; @@ -627,7 +627,7 @@ TEST_CASE("Unit_hipMallocFromPoolAsync_Multidevice_Concurrent") { * ------------------------ * - HIP_VERSION >= 6.2 */ -TEST_CASE("Unit_hipMallocFromPoolAsync_Multidevice_MultiStream") { +TEST_CASE("Unit_hipMallocFromPoolAsync_Multidevice_MultiStream", "[multigpu]") { int num_devices; auto testType = GENERATE(testdefault, testMaximum); constexpr int N = 1 << 20; diff --git a/projects/hip-tests/catch/unit/memory/hipMallocManagedFlagsTst.cc b/projects/hip-tests/catch/unit/memory/hipMallocManagedFlagsTst.cc index 4199837862..5891de5fbe 100644 --- a/projects/hip-tests/catch/unit/memory/hipMallocManagedFlagsTst.cc +++ b/projects/hip-tests/catch/unit/memory/hipMallocManagedFlagsTst.cc @@ -31,7 +31,7 @@ __global__ void MallcMangdFlgTst(int n, float* x, float* y) { } // The following section tests working of hipMallocManaged with flag parameters -TEST_CASE("Unit_hipMallocManaged_FlgParam") { +TEST_CASE("Unit_hipMallocManaged_FlgParam", "[multigpu]") { auto managed = HmmAttrPrint(); if (managed != 1) { HipTest::HIP_SKIP_TEST("GPU doesn't support managed memory so skipping test."); @@ -119,7 +119,7 @@ TEST_CASE("Unit_hipMallocManaged_FlgParam") { // The following function tests Memory access allocated using hipMallocManaged // in multiple streams -TEST_CASE("Unit_hipMallocManaged_AccessMultiStream") { +TEST_CASE("Unit_hipMallocManaged_AccessMultiStream", "[multigpu]") { auto managed = HmmAttrPrint(); if (managed != 1) { HipTest::HIP_SKIP_TEST("GPU doesn't support managed memory so skipping test."); diff --git a/projects/hip-tests/catch/unit/memory/hipMallocManaged_MultiScenario.cc b/projects/hip-tests/catch/unit/memory/hipMallocManaged_MultiScenario.cc index d828d65971..551fdf0a4d 100644 --- a/projects/hip-tests/catch/unit/memory/hipMallocManaged_MultiScenario.cc +++ b/projects/hip-tests/catch/unit/memory/hipMallocManaged_MultiScenario.cc @@ -161,7 +161,7 @@ TEST_CASE("Unit_hipMallocManaged_MultiChunkSingleDevice") { // Equal parts of Hmm is accessed on available gpus and // kernel is launched on acessed chunk of hmm memory // and checks if there are any inconsistencies or access issues -TEST_CASE("Unit_hipMallocManaged_MultiChunkMultiDevice") { +TEST_CASE("Unit_hipMallocManaged_MultiChunkMultiDevice", "[multigpu]") { auto managed = HmmAttrPrint(); if (managed != 1) { HipTest::HIP_SKIP_TEST("GPU doesn't support managed memory so skipping test."); @@ -304,7 +304,8 @@ TEST_CASE("Unit_hipMallocManaged_Negative") { // Allocate two pointers using hipMallocManaged(), initialize, // then launch kernel using these pointers directly and // later validate the content without using any Memcpy. -TEMPLATE_TEST_CASE("Unit_hipMallocManaged_TwoPointers", "", int, float, double) { +TEMPLATE_TEST_CASE("Unit_hipMallocManaged_TwoPointers", "[multigpu]", int, + float, double) { auto managed = HmmAttrPrint(); if (managed != 1) { HipTest::HIP_SKIP_TEST("GPU doesn't support managed memory so skipping test."); @@ -344,8 +345,8 @@ TEMPLATE_TEST_CASE("Unit_hipMallocManaged_TwoPointers", "", int, float, double) // to all other devices. This include verification and Device two Device // transfers and kernel launch o discover if there any access issues. -TEMPLATE_TEST_CASE("Unit_hipMallocManaged_DeviceContextChange", "", unsigned char, int, float, - double) { +TEMPLATE_TEST_CASE("Unit_hipMallocManaged_DeviceContextChange", "[multigpu]", + unsigned char, int, float, double) { auto managed = HmmAttrPrint(); if (managed != 1) { HipTest::HIP_SKIP_TEST("GPU doesn't support managed memory so skipping test."); diff --git a/projects/hip-tests/catch/unit/memory/hipMallocMipmappedArray.cc b/projects/hip-tests/catch/unit/memory/hipMallocMipmappedArray.cc index aebd4bee16..1789142db6 100644 --- a/projects/hip-tests/catch/unit/memory/hipMallocMipmappedArray.cc +++ b/projects/hip-tests/catch/unit/memory/hipMallocMipmappedArray.cc @@ -85,7 +85,7 @@ This testcase verifies the hipMallocMipmappedArray API in multithreaded scenario by launching threads in parallel on multiple GPUs and verifies the hipMallocMipmappedArray API with small and big chunks data */ -TEST_CASE("Unit_hipMallocMipmappedArray_MultiThread") { +TEST_CASE("Unit_hipMallocMipmappedArray_MultiThread", "[multigpu]") { std::vector threadlist; int devCnt = 0; devCnt = HipTest::getDeviceCount(); diff --git a/projects/hip-tests/catch/unit/memory/hipMallocMngdMultiThread.cc b/projects/hip-tests/catch/unit/memory/hipMallocMngdMultiThread.cc index 03f2b02fc2..7ee31832d2 100644 --- a/projects/hip-tests/catch/unit/memory/hipMallocMngdMultiThread.cc +++ b/projects/hip-tests/catch/unit/memory/hipMallocMngdMultiThread.cc @@ -297,7 +297,7 @@ static void AllocateHmmMemory(int flag, int device) { } } -TEST_CASE("Unit_hipMallocManaged_MultiThread") { +TEST_CASE("Unit_hipMallocManaged_MultiThread", "[multigpu]") { auto managed = HmmAttrPrint(); if (managed != 1) { HipTest::HIP_SKIP_TEST("GPU doesn't support managed memory so skipping test."); @@ -351,7 +351,7 @@ TEST_CASE("Unit_hipMallocManaged_MultiThread") { // The following test checks what happens when same Hmm memory is used to // launch multiple threads over multiple gpus -TEST_CASE("Unit_hipMallocManaged_MGpuMThread") { +TEST_CASE("Unit_hipMallocManaged_MGpuMThread", "[multigpu]") { auto managed = HmmAttrPrint(); if (managed != 1) { HipTest::HIP_SKIP_TEST("GPU doesn't support managed memory so skipping test."); diff --git a/projects/hip-tests/catch/unit/memory/hipMallocPitch.cc b/projects/hip-tests/catch/unit/memory/hipMallocPitch.cc index c02f16c965..bcac741d32 100644 --- a/projects/hip-tests/catch/unit/memory/hipMallocPitch.cc +++ b/projects/hip-tests/catch/unit/memory/hipMallocPitch.cc @@ -459,7 +459,7 @@ scenario by launching threads in parallel on multiple GPUs and verifies the hipMallocPitch API with small and big chunks data */ -TEST_CASE("Unit_hipMallocPitch_MultiThread", "") { +TEST_CASE("Unit_hipMallocPitch_MultiThread", "[multigpu]") { CHECK_IMAGE_SUPPORT std::vector threadlist; diff --git a/projects/hip-tests/catch/unit/memory/hipMemAdvise.cc b/projects/hip-tests/catch/unit/memory/hipMemAdvise.cc index 0436f981ba..39c379e580 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemAdvise.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemAdvise.cc @@ -181,7 +181,7 @@ TEST_CASE("Unit_hipMemAdvise_Flags_Do_Not_Cause_Prefetch") { #endif } -TEST_CASE("Unit_hipMemAdvise_Read_Write_After_Advise") { +TEST_CASE("Unit_hipMemAdvise_Read_Write_After_Advise", "[multigpu]") { auto supported_devices = GetDevicesWithAdviseSupport(); if (supported_devices.empty()) { HipTest::HIP_SKIP_TEST("Test needs at least 1 device that supports managed memory"); diff --git a/projects/hip-tests/catch/unit/memory/hipMemAdvise_old.cc b/projects/hip-tests/catch/unit/memory/hipMemAdvise_old.cc index ece1ad9598..7d92bbb5a7 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemAdvise_old.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemAdvise_old.cc @@ -243,7 +243,7 @@ TEST_CASE("Unit_hipMemAdvise_NegtveTsts") { // The following function tests various scenarios around the flag // 'hipMemAdviseSetPreferredLocation' using HMM memory and hipMemAdvise() api -TEST_CASE("Unit_hipMemAdvise_PrefrdLoc") { +TEST_CASE("Unit_hipMemAdvise_PrefrdLoc", "[multigpu]") { int MangdMem = HmmAttrPrint(); if (MangdMem == 1) { // Check that when a page fault occurs for the memory region set to devPtr, @@ -428,7 +428,7 @@ TEST_CASE("Unit_hipMemAdvise_TstFlgOverrideEffect") { // The following function tests if peers can set hipMemAdviseSetAccessedBy flag // on HMM memory prefetched on each of the other gpus #if HT_AMD -TEST_CASE("Unit_hipMemAdvise_TstAccessedByPeer") { +TEST_CASE("Unit_hipMemAdvise_TstAccessedByPeer", "[multigpu]") { int MangdMem = HmmAttrPrint(); if (MangdMem == 1) { bool IfTestPassed = true; @@ -732,7 +732,7 @@ TEST_CASE("Unit_hipMemAdvise_TstMemAdvisePrefrdLoc") { to device1, probe for hipMemRangeAttributeLastPrefetchLocation using hipMemRangeGetAttribute(), we should get 1*/ -TEST_CASE("Unit_hipMemAdvise_TstMemAdviseLstPreftchLoc") { +TEST_CASE("Unit_hipMemAdvise_TstMemAdviseLstPreftchLoc", "[multigpu]") { int NumDevs = 0; HIP_CHECK(hipGetDeviceCount(&NumDevs)); if (NumDevs >= 2) { @@ -802,7 +802,7 @@ TEST_CASE("Unit_hipMemAdvise_TstMemAdviseMultiFlag") { access denial case arising due to setting ReadMostly only to a particular gpu*/ -TEST_CASE("Unit_hipMemAdvise_ReadMosltyMgpuTst") { +TEST_CASE("Unit_hipMemAdvise_ReadMosltyMgpuTst", "[multigpu]") { int managed = HmmAttrPrint(); if (managed == 1) { int Ngpus = 0; diff --git a/projects/hip-tests/catch/unit/memory/hipMemAdvise_v2.cc b/projects/hip-tests/catch/unit/memory/hipMemAdvise_v2.cc index bc1e9f50d8..94d43a2388 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemAdvise_v2.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemAdvise_v2.cc @@ -70,7 +70,7 @@ static std::vector getSupportedDevices() { * ------------------------ * - HIP_VERSION >= 7.1 */ -TEST_CASE("Unit_hipMemAdvise_v2_Device_Host") { +TEST_CASE("Unit_hipMemAdvise_v2_Device_Host", "[multigpu]") { auto supportedDevices = getSupportedDevices(); if (supportedDevices.empty()) { HipTest::HIP_SKIP_TEST( diff --git a/projects/hip-tests/catch/unit/memory/hipMemAllocHost.cc b/projects/hip-tests/catch/unit/memory/hipMemAllocHost.cc index 50cfa5bf25..92236b0fec 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemAllocHost.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemAllocHost.cc @@ -93,7 +93,7 @@ TEST_CASE("Unit_hipMemAllocHost_Negative") { /* * Verify that a device can read/write to the memory of another device */ -TEST_CASE("Unit_hipMemAllocHost_VerifyAccess") { +TEST_CASE("Unit_hipMemAllocHost_VerifyAccess", "[multigpu]") { int devices_number = 0; HIP_CHECK(hipGetDeviceCount(&devices_number)); std::vector devices_memories(devices_number); diff --git a/projects/hip-tests/catch/unit/memory/hipMemPoolCreate.cc b/projects/hip-tests/catch/unit/memory/hipMemPoolCreate.cc index 5b309990a9..87d5bf92ef 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemPoolCreate.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemPoolCreate.cc @@ -157,7 +157,7 @@ static __global__ void setKer(int* devptr) { * ------------------------ * - HIP_VERSION >= 6.2 */ -TEST_CASE("Unit_hipMemPoolCreate_DeviceTest") { +TEST_CASE("Unit_hipMemPoolCreate_DeviceTest", "[multigpu]") { checkMempoolSupported(0) int num_devices = 0; HIP_CHECK(hipGetDeviceCount(&num_devices)); checkIfMultiDev(num_devices) diff --git a/projects/hip-tests/catch/unit/memory/hipMemPoolSetGetAccess.cc b/projects/hip-tests/catch/unit/memory/hipMemPoolSetGetAccess.cc index bbad79748e..8668568148 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemPoolSetGetAccess.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemPoolSetGetAccess.cc @@ -98,7 +98,7 @@ int CheckP2PMemPoolSupport(int src_device, int dst_device) { * ------------------------ * - HIP_VERSION >= 6.2 */ -TEST_CASE("Unit_hipMemPoolSetGetAccess_Positive_MultipleGPU") { +TEST_CASE("Unit_hipMemPoolSetGetAccess_Positive_MultipleGPU", "[multigpu]") { const auto device_count = HipTest::getDeviceCount(); if (device_count < 2) { HipTest::HIP_SKIP_TEST("Skipping because devices < 2"); @@ -212,7 +212,7 @@ void MemPoolSetGetAccess_P2P(const MemPools mempool_type) { * ------------------------ * - HIP_VERSION >= 6.2 */ -TEST_CASE("Unit_hipMemPoolSetGetAccess_Positive_P2P") { +TEST_CASE("Unit_hipMemPoolSetGetAccess_Positive_P2P", "[multigpu]") { const auto device_count = HipTest::getDeviceCount(); if (device_count < 2) { HipTest::HIP_SKIP_TEST("Skipping because devices < 2"); @@ -406,7 +406,7 @@ static void getDevicePairs(std::vector>* p2p_pairs, int numD * ------------------------ * - HIP_VERSION >= 6.2 */ -TEST_CASE("Unit_hipMemPoolSetAccess_SetAccess") { +TEST_CASE("Unit_hipMemPoolSetAccess_SetAccess", "[multigpu]") { constexpr int N = 1 << 14; int numDevices = 0; HIP_CHECK(hipGetDeviceCount(&numDevices)); diff --git a/projects/hip-tests/catch/unit/memory/hipMemPrefetchAsync.cc b/projects/hip-tests/catch/unit/memory/hipMemPrefetchAsync.cc index 777c52249a..8f0e292bae 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemPrefetchAsync.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemPrefetchAsync.cc @@ -47,7 +47,7 @@ __global__ void MemPrefetchAsyncKernel(int* C_d, const int* A_d, size_t N) { } } -TEST_CASE("Unit_hipMemPrefetchAsync_Basic") { +TEST_CASE("Unit_hipMemPrefetchAsync_Basic", "[multigpu]") { const auto supported_devices = GetDevicesWithPrefetchSupport(); if (supported_devices.empty()) { HipTest::HIP_SKIP_TEST("Test need at least one device with managed memory support"); diff --git a/projects/hip-tests/catch/unit/memory/hipMemPrefetchAsync_v2.cc b/projects/hip-tests/catch/unit/memory/hipMemPrefetchAsync_v2.cc index e6efd86723..7f2373c1a6 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemPrefetchAsync_v2.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemPrefetchAsync_v2.cc @@ -70,7 +70,7 @@ static std::vector getSupportedDevices() { * ------------------------ * - HIP_VERSION >= 7.1 */ -TEST_CASE("Unit_hipMemPrefetchAsync_v2_Device_Host") { +TEST_CASE("Unit_hipMemPrefetchAsync_v2_Device_Host", "[multigpu]") { auto supportedDevices = getSupportedDevices(); if (supportedDevices.empty()) { HipTest::HIP_SKIP_TEST( diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpy2D.cc b/projects/hip-tests/catch/unit/memory/hipMemcpy2D.cc index 7a1e995d81..6ef94e0b88 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpy2D.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpy2D.cc @@ -27,7 +27,7 @@ THE SOFTWARE. #include #include -TEST_CASE("Unit_hipMemcpy2D_Positive_Basic") { +TEST_CASE("Unit_hipMemcpy2D_Positive_Basic", "[multigpu]") { CHECK_IMAGE_SUPPORT constexpr bool async = false; diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpy2DAsync.cc b/projects/hip-tests/catch/unit/memory/hipMemcpy2DAsync.cc index 312e7a1a02..ef6af19c15 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpy2DAsync.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpy2DAsync.cc @@ -27,7 +27,7 @@ THE SOFTWARE. #include #include -TEST_CASE("Unit_hipMemcpy2DAsync_Positive_Basic") { +TEST_CASE("Unit_hipMemcpy2DAsync_Positive_Basic", "[multigpu]") { using namespace std::placeholders; constexpr bool async = true; diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpy2DAsync_old.cc b/projects/hip-tests/catch/unit/memory/hipMemcpy2DAsync_old.cc index dac1d0e322..f5a5c51169 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpy2DAsync_old.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpy2DAsync_old.cc @@ -172,7 +172,8 @@ TEMPLATE_TEST_CASE("Unit_hipMemcpy2DAsync_Host&PinnedMem", "", int, float, doubl * - HIP_VERSION >= 5.2 */ -TEMPLATE_TEST_CASE("Unit_hipMemcpy2DAsync_multiDevice-Host&PinnedMem", "", int, float, double) { +TEMPLATE_TEST_CASE("Unit_hipMemcpy2DAsync_multiDevice-Host&PinnedMem", + "[multigpu]", int, float, double) { CHECK_IMAGE_SUPPORT auto mem_type = GENERATE(0, 1); int numDevices = 0; @@ -264,7 +265,8 @@ TEMPLATE_TEST_CASE("Unit_hipMemcpy2DAsync_multiDevice-Host&PinnedMem", "", int, * - HIP_VERSION >= 5.2 */ -TEMPLATE_TEST_CASE("Unit_hipMemcpy2DAsync_multiDevice-StreamOnDiffDevice", "", int, float, double) { +TEMPLATE_TEST_CASE("Unit_hipMemcpy2DAsync_multiDevice-StreamOnDiffDevice", + "[multigpu]", int, float, double) { CHECK_IMAGE_SUPPORT auto mem_type = GENERATE(0, 1); int numDevices = 0; @@ -506,7 +508,7 @@ static void hipMemcpy2DAsync_Basic_Size_Test(size_t inc) { * - HIP_VERSION >= 6.0 */ -TEST_CASE("Unit_hipMemcpy2DAsync_multiDevice_Basic_Size_Test") { +TEST_CASE("Unit_hipMemcpy2DAsync_multiDevice_Basic_Size_Test", "[multigpu]") { CHECK_IMAGE_SUPPORT size_t input = 1 << 20; int numDevices = 0; diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpy2DFromArray.cc b/projects/hip-tests/catch/unit/memory/hipMemcpy2DFromArray.cc index 05ab0811ba..7e4cfffa94 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpy2DFromArray.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpy2DFromArray.cc @@ -34,8 +34,7 @@ invalid #include #include - -TEST_CASE("Unit_hipMemcpy2DFromArray_Positive_Default") { +TEST_CASE("Unit_hipMemcpy2DFromArray_Positive_Default", "[multigpu]") { CHECK_IMAGE_SUPPORT using namespace std::placeholders; diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpy2DFromArrayAsync.cc b/projects/hip-tests/catch/unit/memory/hipMemcpy2DFromArrayAsync.cc index 0fa4a14219..14a92eece2 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpy2DFromArrayAsync.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpy2DFromArrayAsync.cc @@ -34,7 +34,7 @@ of hipMemcpy2DFromArrayAsync api when parameters are invalid #include #include -TEST_CASE("Unit_hipMemcpy2DFromArrayAsync_Positive_Default") { +TEST_CASE("Unit_hipMemcpy2DFromArrayAsync_Positive_Default", "[multigpu]") { CHECK_IMAGE_SUPPORT using namespace std::placeholders; diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpy2DFromArrayAsync_old.cc b/projects/hip-tests/catch/unit/memory/hipMemcpy2DFromArrayAsync_old.cc index 5c94dffebd..1591f807a8 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpy2DFromArrayAsync_old.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpy2DFromArrayAsync_old.cc @@ -195,7 +195,8 @@ TEST_CASE("Unit_hipMemcpy2DFromArrayAsync_PinnedHostMemSameGpu") { then A_d-->E_h in GPU1 * OUTPUT: validating the result by comparing A_h and E_h */ -TEST_CASE("Unit_hipMemcpy2DFromArrayAsync_multiDevicePinnedHostMem") { +TEST_CASE("Unit_hipMemcpy2DFromArrayAsync_multiDevicePinnedHostMem", + "[multigpu]") { CHECK_IMAGE_SUPPORT int numDevices = 0; @@ -254,7 +255,8 @@ TEST_CASE("Unit_hipMemcpy2DFromArrayAsync_multiDevicePinnedHostMem") { * --> A_h host variable * and verifying A_h with Phi * */ -TEST_CASE("Unit_hipMemcpy2DFromArrayAsync_multiDeviceContextChange") { +TEST_CASE("Unit_hipMemcpy2DFromArrayAsync_multiDeviceContextChange", + "[multigpu]") { CHECK_IMAGE_SUPPORT int numDevices = 0; diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpy2DFromArray_old.cc b/projects/hip-tests/catch/unit/memory/hipMemcpy2DFromArray_old.cc index 65dd09acfc..4e10f3b433 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpy2DFromArray_old.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpy2DFromArray_old.cc @@ -165,7 +165,8 @@ TEST_CASE("Unit_hipMemcpy2DFromArray_PinnedMemSameGPU") { * --> E_h host variable * and verifying A_h with E_h */ -TEST_CASE("Unit_hipMemcpy2DFromArray_multiDevicePinnedMemPeerGpu") { +TEST_CASE("Unit_hipMemcpy2DFromArray_multiDevicePinnedMemPeerGpu", + "[multigpu]") { CHECK_IMAGE_SUPPORT int numDevices = 0; @@ -218,7 +219,7 @@ TEST_CASE("Unit_hipMemcpy2DFromArray_multiDevicePinnedMemPeerGpu") { * --> A_h host variable * and verifying A_h with Phi * */ -TEST_CASE("Unit_hipMemcpy2DFromArray_multiDeviceContextChange") { +TEST_CASE("Unit_hipMemcpy2DFromArray_multiDeviceContextChange", "[multigpu]") { CHECK_IMAGE_SUPPORT int numDevices = 0; diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpy2DToArray.cc b/projects/hip-tests/catch/unit/memory/hipMemcpy2DToArray.cc index 4ab2a48cc0..c5ad586013 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpy2DToArray.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpy2DToArray.cc @@ -33,8 +33,7 @@ unsuccessful execution of hipMemcpy2DToArray api when parameters are invalid #include #include - -TEST_CASE("Unit_hipMemcpy2DToArray_Positive_Default") { +TEST_CASE("Unit_hipMemcpy2DToArray_Positive_Default", "[multigpu]") { CHECK_IMAGE_SUPPORT using namespace std::placeholders; diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpy2DToArrayAsync.cc b/projects/hip-tests/catch/unit/memory/hipMemcpy2DToArrayAsync.cc index e2adcfff45..c90ab755be 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpy2DToArrayAsync.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpy2DToArrayAsync.cc @@ -34,8 +34,7 @@ of hipMemcpy2DToArrayAsync api when parameters are invalid #include #include - -TEST_CASE("Unit_hipMemcpy2DToArrayAsync_Positive_Default") { +TEST_CASE("Unit_hipMemcpy2DToArrayAsync_Positive_Default", "[multigpu]") { CHECK_IMAGE_SUPPORT using namespace std::placeholders; diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpy2DToArrayAsync_old.cc b/projects/hip-tests/catch/unit/memory/hipMemcpy2DToArrayAsync_old.cc index 7c154f979f..cd2e55613e 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpy2DToArrayAsync_old.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpy2DToArrayAsync_old.cc @@ -194,7 +194,8 @@ TEST_CASE("Unit_hipMemcpy2DToArrayAsync_PinnedHostMemSameGpu") { * --> A_h host variable * and verifying A_h with E_h[0]+i(i.e., 10+i) */ -TEST_CASE("Unit_hipMemcpy2DToArrayAsync_multiDevicePinnedHostMem") { +TEST_CASE("Unit_hipMemcpy2DToArrayAsync_multiDevicePinnedHostMem", + "[multigpu]") { CHECK_IMAGE_SUPPORT int numDevices = 0; @@ -254,7 +255,8 @@ TEST_CASE("Unit_hipMemcpy2DToArrayAsync_multiDevicePinnedHostMem") { * --> A_h host variable * and verifying A_h with Phi * */ -TEST_CASE("Unit_hipMemcpy2DToArrayAsync_multiDeviceDeviceContextChange") { +TEST_CASE("Unit_hipMemcpy2DToArrayAsync_multiDeviceDeviceContextChange", + "[multigpu]") { CHECK_IMAGE_SUPPORT int numDevices = 0; diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpy2DToArray_old.cc b/projects/hip-tests/catch/unit/memory/hipMemcpy2DToArray_old.cc index 29f70d65ec..e302d606b1 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpy2DToArray_old.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpy2DToArray_old.cc @@ -164,7 +164,7 @@ TEST_CASE("Unit_hipMemcpy2DToArray_PinnedMemSameGPU") { * --> A_h host variable * and verifying A_h with E_h[0]+i(i.e., 10+i) */ -TEST_CASE("Unit_hipMemcpy2DToArray_multiDevicePinnedMemPeerGpu") { +TEST_CASE("Unit_hipMemcpy2DToArray_multiDevicePinnedMemPeerGpu", "[multigpu]") { CHECK_IMAGE_SUPPORT int numDevices = 0; @@ -218,7 +218,8 @@ TEST_CASE("Unit_hipMemcpy2DToArray_multiDevicePinnedMemPeerGpu") { * --> A_h host variable * and verifying A_h with Phi * */ -TEST_CASE("Unit_hipMemcpy2DToArray_multiDeviceDeviceContextChange") { +TEST_CASE("Unit_hipMemcpy2DToArray_multiDeviceDeviceContextChange", + "[multigpu]") { CHECK_IMAGE_SUPPORT int numDevices = 0; diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpy2D_old.cc b/projects/hip-tests/catch/unit/memory/hipMemcpy2D_old.cc index b5c20ea9aa..d0afaa1dfb 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpy2D_old.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpy2D_old.cc @@ -309,7 +309,8 @@ TEMPLATE_TEST_CASE("Unit_hipMemcpy2D_H2D-D2D-D2H_Managed_WithOffset", "", int, f * - HIP_VERSION >= 6.0 */ -TEMPLATE_TEST_CASE("Unit_hipMemcpy2D_multiDevice-D2D", "", int, float, double) { +TEMPLATE_TEST_CASE("Unit_hipMemcpy2D_multiDevice-D2D", "[multigpu]", int, float, + double) { CHECK_IMAGE_SUPPORT auto mem_type = GENERATE(0, 1); int numDevices = 0; @@ -524,7 +525,7 @@ static void hipMemcpy2D_Basic_Size_Test(size_t inc) { * - HIP_VERSION >= 6.0 */ -TEST_CASE("Unit_hipMemcpy2D_multiDevice_Basic_Size_Test") { +TEST_CASE("Unit_hipMemcpy2D_multiDevice_Basic_Size_Test", "[multigpu]") { CHECK_IMAGE_SUPPORT size_t input = 1 << 20; int numDevices = 0; diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpy3DAsync_old.cc b/projects/hip-tests/catch/unit/memory/hipMemcpy3DAsync_old.cc index 7d3fddf905..82984bc55a 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpy3DAsync_old.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpy3DAsync_old.cc @@ -715,7 +715,7 @@ TEST_CASE("Unit_hipMemcpy3DAsync_multiDevice-Negative") { * - HIP_VERSION >= 6.0 */ -TEST_CASE("Unit_hipMemcpy3DAsync_multiDevice-D2D") { +TEST_CASE("Unit_hipMemcpy3DAsync_multiDevice-D2D", "[multigpu]") { CHECK_IMAGE_SUPPORT int numDevices = 0; HIP_CHECK(hipGetDeviceCount(&numDevices)); @@ -747,7 +747,7 @@ TEST_CASE("Unit_hipMemcpy3DAsync_multiDevice-D2D") { * - HIP_VERSION >= 6.0 */ -TEST_CASE("Unit_hipMemcpy3DAsync_multiDevice-DiffStream") { +TEST_CASE("Unit_hipMemcpy3DAsync_multiDevice-DiffStream", "[multigpu]") { CHECK_IMAGE_SUPPORT int numDevices = 0; HIP_CHECK(hipGetDeviceCount(&numDevices)); diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpy3D_old.cc b/projects/hip-tests/catch/unit/memory/hipMemcpy3D_old.cc index da0ec091af..17ce84533a 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpy3D_old.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpy3D_old.cc @@ -609,7 +609,7 @@ TEST_CASE("Unit_hipMemcpy3D_multiDevice-Negative") { * - HIP_VERSION >= 5.2 */ -TEST_CASE("Unit_hipMemcpy3D_multiDevice-OnPeerDevice") { +TEST_CASE("Unit_hipMemcpy3D_multiDevice-OnPeerDevice", "[multigpu]") { CHECK_IMAGE_SUPPORT int numDevices = 0; HIP_CHECK(hipGetDeviceCount(&numDevices)); @@ -644,7 +644,7 @@ TEST_CASE("Unit_hipMemcpy3D_multiDevice-OnPeerDevice") { * - HIP_VERSION >= 6.0 */ -TEST_CASE("Unit_hipMemcpy3D_multiDevice_Basic_Size_Test") { +TEST_CASE("Unit_hipMemcpy3D_multiDevice_Basic_Size_Test", "[multigpu]") { CHECK_IMAGE_SUPPORT constexpr int size_128b = 128, size_256b = 256; int numDevices = 0; diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpyAsync_old.cc b/projects/hip-tests/catch/unit/memory/hipMemcpyAsync_old.cc index 9ce903dd06..507d9d17a9 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpyAsync_old.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpyAsync_old.cc @@ -128,7 +128,8 @@ This testcase verifies the following scenarios 4. Device context change 5. H2D-D2D-D2H peer GPU */ -TEMPLATE_TEST_CASE("Unit_hipMemcpyAsync_H2H-H2D-D2H-H2PinMem", "", char, int, float, double) { +TEMPLATE_TEST_CASE("Unit_hipMemcpyAsync_H2H-H2D-D2H-H2PinMem", "[multigpu]", + char, int, float, double) { TestType *A_d{nullptr}, *B_d{nullptr}; TestType *A_h{nullptr}, *B_h{nullptr}; TestType *A_Ph{nullptr}, *B_Ph{nullptr}; @@ -288,7 +289,8 @@ This testcase verifies hipMemcpy API with pinnedMemory and hostRegister along with kernel launches */ -TEMPLATE_TEST_CASE("Unit_hipMemcpyAsync_PinnedRegMemWithKernelLaunch", "", int, float, double) { +TEMPLATE_TEST_CASE("Unit_hipMemcpyAsync_PinnedRegMemWithKernelLaunch", + "[multigpu]", int, float, double) { int numDevices = 0; HIP_CHECK(hipGetDeviceCount(&numDevices)); if (numDevices < 2) { diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpyAtoH_old.cc b/projects/hip-tests/catch/unit/memory/hipMemcpyAtoH_old.cc index 353bb06da2..c76ec5e77e 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpyAtoH_old.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpyAtoH_old.cc @@ -90,8 +90,8 @@ Output:"B_h" host variable output of hipMemcpyAtoH API is then validated with "hData" */ #if HT_AMD -TEMPLATE_TEST_CASE("Unit_hipMemcpyAtoH_multiDevice-PeerDeviceContext", "[hipMemcpyAtoH]", char, int, - float) { +TEMPLATE_TEST_CASE("Unit_hipMemcpyAtoH_multiDevice-PeerDeviceContext", + "[hipMemcpyAtoH][multigpu]", char, int, float) { CHECK_IMAGE_SUPPORT int numDevices = 0; diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpyDtoD.cc b/projects/hip-tests/catch/unit/memory/hipMemcpyDtoD.cc index c0c8056d12..77e94d1cff 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpyDtoD.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpyDtoD.cc @@ -38,7 +38,8 @@ This testcase verifies hipMemcpyDtoD API 6.Kernel Launch 7.DtoH copy and validating the result */ -TEMPLATE_TEST_CASE("Unit_hipMemcpyDtoD_Basic", "", int, float, double) { +TEMPLATE_TEST_CASE("Unit_hipMemcpyDtoD_Basic", "[multigpu]", int, float, + double) { size_t Nbytes = NUM_ELM * sizeof(TestType); int numDevices = 0; TestType *A_d{nullptr}, *B_d{nullptr}, *C_d{nullptr}, *X_d{nullptr}, *Y_d{nullptr}, *Z_d{nullptr}; diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpyDtoDAsync.cc b/projects/hip-tests/catch/unit/memory/hipMemcpyDtoDAsync.cc index 01480c688e..1e22d8e7bd 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpyDtoDAsync.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpyDtoDAsync.cc @@ -40,7 +40,8 @@ This testcase verifies hipMemcpyDtoDAsync API 7.DtoH copy and validating the result */ -TEMPLATE_TEST_CASE("Unit_hipMemcpyDtoDAsync_Basic", "", int, float, double) { +TEMPLATE_TEST_CASE("Unit_hipMemcpyDtoDAsync_Basic", "[multigpu]", int, float, + double) { size_t Nbytes = NUM_ELM * sizeof(TestType); int numDevices = 0; TestType *A_d{nullptr}, *B_d{nullptr}, *C_d{nullptr}, *X_d{nullptr}, *Y_d{nullptr}, *Z_d{nullptr}; diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpyHtoAAsync.cc b/projects/hip-tests/catch/unit/memory/hipMemcpyHtoAAsync.cc index 6bd1b1f363..170286267b 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpyHtoAAsync.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpyHtoAAsync.cc @@ -166,7 +166,7 @@ TEST_CASE("Unit_hipMemcpyHtoAAsync_BasicTstsWithDiffStreams") { * ------------------------ * - HIP_VERSION >= 6.2 */ -TEST_CASE("Unit_hipMemcpyHtoAAsync_MultiDevice") { +TEST_CASE("Unit_hipMemcpyHtoAAsync_MultiDevice", "[multigpu]") { #if HT_NVIDIA HipTest::HIP_SKIP_TEST("API currently unsupported on nvidia, skipping..."); return; diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpyHtoA_old.cc b/projects/hip-tests/catch/unit/memory/hipMemcpyHtoA_old.cc index d922133a72..e02c405e4f 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpyHtoA_old.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpyHtoA_old.cc @@ -93,8 +93,8 @@ Output: "A_d" output of hipMemcpyHtoA is copied to "hData" host variable validated the result with "B_h" */ #if HT_AMD -TEMPLATE_TEST_CASE("Unit_hipMemcpyHtoA_multiDevice-PeerDeviceContext", "[hipMemcpyHtoA]", char, int, - float) { +TEMPLATE_TEST_CASE("Unit_hipMemcpyHtoA_multiDevice-PeerDeviceContext", + "[hipMemcpyHtoA][multigpu]", char, int, float) { CHECK_IMAGE_SUPPORT int numDevices = 0; HIP_CHECK(hipGetDeviceCount(&numDevices)); diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpyParam2D.cc b/projects/hip-tests/catch/unit/memory/hipMemcpyParam2D.cc index e3c14610fc..060065c74b 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpyParam2D.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpyParam2D.cc @@ -27,7 +27,7 @@ THE SOFTWARE. #include #include -TEST_CASE("Unit_hipMemcpyParam2D_Positive_Basic") { +TEST_CASE("Unit_hipMemcpyParam2D_Positive_Basic", "[multigpu]") { CHECK_IMAGE_SUPPORT constexpr bool async = false; diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpyParam2DAsync.cc b/projects/hip-tests/catch/unit/memory/hipMemcpyParam2DAsync.cc index e7f219204c..585f0b31e3 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpyParam2DAsync.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpyParam2DAsync.cc @@ -27,7 +27,7 @@ THE SOFTWARE. #include #include -TEST_CASE("Unit_hipMemcpyParam2DAsync_Positive_Basic") { +TEST_CASE("Unit_hipMemcpyParam2DAsync_Positive_Basic", "[multigpu]") { using namespace std::placeholders; constexpr bool async = true; diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpyParam2DAsync_old.cc b/projects/hip-tests/catch/unit/memory/hipMemcpyParam2DAsync_old.cc index a4bab56b88..c1d1b50633 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpyParam2DAsync_old.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpyParam2DAsync_old.cc @@ -44,7 +44,8 @@ static constexpr size_t NUM_H{10}; * */ TEMPLATE_TEST_CASE("Unit_hipMemcpyParam2DAsync_multiDevice-StreamOnDiffDevice", - "[hipMemcpyParam2DAsync]", char, float, int, double, long double) { + "[hipMemcpyParam2DAsync][multigpu]", char, float, int, + double, long double) { CHECK_IMAGE_SUPPORT int numDevices = 0; @@ -118,8 +119,9 @@ TEMPLATE_TEST_CASE("Unit_hipMemcpyParam2DAsync_multiDevice-StreamOnDiffDevice", * it with the initalized data "C_h". * */ -TEMPLATE_TEST_CASE("Unit_hipMemcpyParam2DAsync_multiDevice-D2D", "[hipMemcpyParam2DAsync]", char, - int, float, double, long double) { +TEMPLATE_TEST_CASE("Unit_hipMemcpyParam2DAsync_multiDevice-D2D", + "[hipMemcpyParam2DAsync][multigpu]", char, int, float, + double, long double) { CHECK_IMAGE_SUPPORT int numDevices = 0; @@ -195,8 +197,9 @@ TEMPLATE_TEST_CASE("Unit_hipMemcpyParam2DAsync_multiDevice-D2D", "[hipMemcpyPara * * Validating the result by comparing "A_h" to "C_h" */ -TEMPLATE_TEST_CASE("Unit_hipMemcpyParam2DAsync_multiDevice-H2D-D2H", "[hipMemcpyParam2DAsync]", - char, int, float, double, long double) { +TEMPLATE_TEST_CASE("Unit_hipMemcpyParam2DAsync_multiDevice-H2D-D2H", + "[hipMemcpyParam2DAsync][multigpu]", char, int, float, + double, long double) { CHECK_IMAGE_SUPPORT // 1 refers to pinned host memory and 0 refers diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpyParam2D_old.cc b/projects/hip-tests/catch/unit/memory/hipMemcpyParam2D_old.cc index 41f7e4f05c..0240b10b94 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpyParam2D_old.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpyParam2D_old.cc @@ -39,8 +39,9 @@ static constexpr size_t NUM_H{10}; * it with the initalized data "C_h". * */ -TEMPLATE_TEST_CASE("Unit_hipMemcpyParam2D_multiDevice-D2D", "[hipMemcpyParam2D]", char, float, int, - double, long double) { +TEMPLATE_TEST_CASE("Unit_hipMemcpyParam2D_multiDevice-D2D", + "[hipMemcpyParam2D][multigpu]", char, float, int, double, + long double) { CHECK_IMAGE_SUPPORT int numDevices = 0; diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpyPeer.cc b/projects/hip-tests/catch/unit/memory/hipMemcpyPeer.cc index aa285d3b32..1cec76cfb2 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpyPeer.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpyPeer.cc @@ -45,7 +45,7 @@ THE SOFTWARE. * - Multi-device * - HIP_VERSION >= 5.2 */ -TEST_CASE("Unit_hipMemcpyPeer_Positive_Default") { +TEST_CASE("Unit_hipMemcpyPeer_Positive_Default", "[multigpu]") { const auto device_count = HipTest::getDeviceCount(); if (device_count < 2) { HipTest::HIP_SKIP_TEST("Skipping because devices < 2"); @@ -110,7 +110,8 @@ TEST_CASE("Unit_hipMemcpyPeer_Positive_Default") { * - Multi-device * - HIP_VERSION >= 5.2 */ -TEST_CASE("Unit_hipMemcpyPeer_Positive_Synchronization_Behavior") { +TEST_CASE("Unit_hipMemcpyPeer_Positive_Synchronization_Behavior", + "[multigpu]") { HIP_CHECK(hipDeviceSynchronize()); const auto device_count = HipTest::getDeviceCount(); @@ -158,7 +159,7 @@ TEST_CASE("Unit_hipMemcpyPeer_Positive_Synchronization_Behavior") { * - Multi-device * - HIP_VERSION >= 5.2 */ -TEST_CASE("Unit_hipMemcpyPeer_Positive_ZeroSize") { +TEST_CASE("Unit_hipMemcpyPeer_Positive_ZeroSize", "[multigpu]") { const auto device_count = HipTest::getDeviceCount(); if (device_count < 2) { HipTest::HIP_SKIP_TEST("Skipping because devices < 2"); @@ -241,7 +242,7 @@ TEST_CASE("Unit_hipMemcpyPeer_Positive_ZeroSize") { * - Multi-device * - HIP_VERSION >= 5.2 */ -TEST_CASE("Unit_hipMemcpyPeer_Negative_Parameters") { +TEST_CASE("Unit_hipMemcpyPeer_Negative_Parameters", "[multigpu]") { const auto device_count = HipTest::getDeviceCount(); if (device_count < 2) { HipTest::HIP_SKIP_TEST("Skipping because devices < 2"); diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpyPeerAsync.cc b/projects/hip-tests/catch/unit/memory/hipMemcpyPeerAsync.cc index feeb7322ad..a2bdf638c6 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpyPeerAsync.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpyPeerAsync.cc @@ -45,7 +45,7 @@ THE SOFTWARE. * - Multi-device * - HIP_VERSION >= 5.2 */ -TEST_CASE("Unit_hipMemcpyPeerAsync_Positive_Default") { +TEST_CASE("Unit_hipMemcpyPeerAsync_Positive_Default", "[multigpu]") { const auto device_count = HipTest::getDeviceCount(); if (device_count < 2) { HipTest::HIP_SKIP_TEST("Skipping because devices < 2"); @@ -114,7 +114,8 @@ TEST_CASE("Unit_hipMemcpyPeerAsync_Positive_Default") { * - Multi-device * - HIP_VERSION >= 5.2 */ -TEST_CASE("Unit_hipMemcpyPeerAsync_Positive_Synchronization_Behavior") { +TEST_CASE("Unit_hipMemcpyPeerAsync_Positive_Synchronization_Behavior", + "[multigpu]") { HIP_CHECK(hipDeviceSynchronize()); const auto device_count = HipTest::getDeviceCount(); @@ -165,7 +166,7 @@ TEST_CASE("Unit_hipMemcpyPeerAsync_Positive_Synchronization_Behavior") { * - Multi-device * - HIP_VERSION >= 5.2 */ -TEST_CASE("Unit_hipMemcpyPeerAsync_Positive_ZeroSize") { +TEST_CASE("Unit_hipMemcpyPeerAsync_Positive_ZeroSize", "[multigpu]") { const auto device_count = HipTest::getDeviceCount(); if (device_count < 2) { HipTest::HIP_SKIP_TEST("Skipping because devices < 2"); @@ -250,7 +251,7 @@ TEST_CASE("Unit_hipMemcpyPeerAsync_Positive_ZeroSize") { * - Multi-device * - HIP_VERSION >= 5.2 */ -TEST_CASE("Unit_hipMemcpyPeerAsync_Negative_Parameters") { +TEST_CASE("Unit_hipMemcpyPeerAsync_Negative_Parameters", "[multigpu]") { const auto device_count = HipTest::getDeviceCount(); if (device_count < 2) { HipTest::HIP_SKIP_TEST("Skipping because devices < 2"); diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpyPeerAsync_old.cc b/projects/hip-tests/catch/unit/memory/hipMemcpyPeerAsync_old.cc index ae1d95ffbe..4044c7036a 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpyPeerAsync_old.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpyPeerAsync_old.cc @@ -32,7 +32,7 @@ This testfile verifies the following scenarios of hipMemcpyPeerAsync API /*This testcase verifies the negative scenarios of hipmemcpypeerAsync */ -TEST_CASE("Unit_hipMemcpyPeerAsync_Negative") { +TEST_CASE("Unit_hipMemcpyPeerAsync_Negative", "[multigpu]") { constexpr auto numElements{10}; constexpr auto copy_bytes{numElements * sizeof(int)}; int numDevices = 0; @@ -99,7 +99,7 @@ TEST_CASE("Unit_hipMemcpyPeerAsync_Negative") { * Then performs the addition and validates the sum */ -TEST_CASE("Unit_hipMemcpyPeerAsync_Basic") { +TEST_CASE("Unit_hipMemcpyPeerAsync_Basic", "[multigpu]") { constexpr auto numElements{10}; constexpr auto copy_bytes{numElements * sizeof(int)}; @@ -175,7 +175,7 @@ TEST_CASE("Unit_hipMemcpyPeerAsync_Basic") { * where stream is created in GPU-1 * Then performs the addition and validates the sum */ -TEST_CASE("Unit_hipMemcpyPeerAsync_StreamOnDiffDevice") { +TEST_CASE("Unit_hipMemcpyPeerAsync_StreamOnDiffDevice", "[multigpu]") { constexpr auto numElements{10}; constexpr auto copy_bytes{numElements * sizeof(int)}; int numDevices = 0; diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpyPeer_old.cc b/projects/hip-tests/catch/unit/memory/hipMemcpyPeer_old.cc index acd028b7a8..2a0a3006fc 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpyPeer_old.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpyPeer_old.cc @@ -28,7 +28,7 @@ This testfile verifies the following scenarios of hipMemcpyPeer API /*This testcase verifies the negative scenarios of hipmemcpypeer */ -TEST_CASE("Unit_hipMemcpyPeer_Negative") { +TEST_CASE("Unit_hipMemcpyPeer_Negative", "[multigpu]") { constexpr auto numElements{10}; constexpr auto copy_bytes{numElements * sizeof(int)}; int numDevices = 0; @@ -91,7 +91,7 @@ TEST_CASE("Unit_hipMemcpyPeer_Negative") { * Copies the data from GPU-0 to GPU-1 using hipMemcpyPeer API * Then performs the addition and validates the sum */ -TEST_CASE("Unit_hipMemcpyPeer_Basic") { +TEST_CASE("Unit_hipMemcpyPeer_Basic", "[multigpu]") { constexpr auto numElements{10}; constexpr auto copy_bytes{numElements * sizeof(int)}; int numDevices = 0; diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpyWithStreamMultiThread.cc b/projects/hip-tests/catch/unit/memory/hipMemcpyWithStreamMultiThread.cc index 45d455e5c2..41b4d846a2 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpyWithStreamMultiThread.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpyWithStreamMultiThread.cc @@ -580,7 +580,7 @@ void HipMemcpyWithStreamMultiThreadtests::TestkindHtoH(bool& val_res) { HIP_CHECK_THREAD(hipStreamDestroy(stream)); } -TEST_CASE("Unit_hipMemcpyWithStream_MultiThread") { +TEST_CASE("Unit_hipMemcpyWithStream_MultiThread", "[multigpu]") { const auto Threadcount{10}; bool ret_val[Threadcount]; std::thread th[Threadcount]; diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpyWithStream_old.cc b/projects/hip-tests/catch/unit/memory/hipMemcpyWithStream_old.cc index e5a68affb4..1841b5690e 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpyWithStream_old.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpyWithStream_old.cc @@ -520,13 +520,20 @@ TEST_CASE("Unit_hipMemcpyWithStream_TestkindDtoH") { TestkindDtoH(); } TEST_CASE("Unit_hipMemcpyWithStream_TestkindHtoH") { TestkindHtoH(); } -TEST_CASE("Unit_hipMemcpyWithStream_TestkindDtoD") { TestkindDtoD(); } +TEST_CASE("Unit_hipMemcpyWithStream_TestkindDtoD", "[multigpu]") { + TestkindDtoD(); +} -TEST_CASE("Unit_hipMemcpyWithStream_TestOnMultiGPUwithOneStream") { TestOnMultiGPUwithOneStream(); } +TEST_CASE("Unit_hipMemcpyWithStream_TestOnMultiGPUwithOneStream", + "[multigpu]") { + TestOnMultiGPUwithOneStream(); +} TEST_CASE("Unit_hipMemcpyWithStream_TestkindDefault") { TestkindDefault(); } #ifndef __HIP_PLATFORM_NVIDIA__ -TEST_CASE("Unit_hipMemcpyWithStream_TestkindDefaultForDtoD") { TestkindDefaultForDtoD(); } +TEST_CASE("Unit_hipMemcpyWithStream_TestkindDefaultForDtoD", "[multigpu]") { + TestkindDefaultForDtoD(); +} #endif TEST_CASE("Unit_hipMemcpyWithStream_TestDtoDonSameDevice") { TestDtoDonSameDevice(); } diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpy_EdgeCases.cc b/projects/hip-tests/catch/unit/memory/hipMemcpy_EdgeCases.cc index 1d8fa6b689..e432fd2845 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpy_EdgeCases.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpy_EdgeCases.cc @@ -399,7 +399,8 @@ This testcase verifies the following scenarios 4. Device context change 5. H2D-D2D-D2H peer GPU */ -TEMPLATE_TEST_CASE("Unit_hipMemcpy_H2H-H2D-D2H-H2PinMem", "", int, float, double) { +TEMPLATE_TEST_CASE("Unit_hipMemcpy_H2H-H2D-D2H-H2PinMem", "[multigpu]", int, + float, double) { TestType *A_d{nullptr}, *B_d{nullptr}; TestType *A_h{nullptr}, *B_h{nullptr}; TestType *A_Ph{nullptr}, *B_Ph{nullptr}; @@ -495,7 +496,8 @@ This testcase verifies hipMemcpy API with pinnedMemory and hostRegister along with kernel launches */ -TEMPLATE_TEST_CASE("Unit_hipMemcpy_PinnedRegMemWithKernelLaunch", "", int, float, double) { +TEMPLATE_TEST_CASE("Unit_hipMemcpy_PinnedRegMemWithKernelLaunch", "[multigpu]", + int, float, double) { int numDevices = 0; HIP_CHECK(hipGetDeviceCount(&numDevices)); if (numDevices < 2) { diff --git a/projects/hip-tests/catch/unit/memory/hipMemcpy_MultiThread.cc b/projects/hip-tests/catch/unit/memory/hipMemcpy_MultiThread.cc index 9550f8d4fd..468412111d 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemcpy_MultiThread.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemcpy_MultiThread.cc @@ -280,7 +280,7 @@ void Thread_func(bool& ret_val) { } -TEST_CASE("Unit_hipMemcpy_MultiThread-AllAPIs") { +TEST_CASE("Unit_hipMemcpy_MultiThread-AllAPIs", "[multigpu]") { std::thread Thrd[NUM_THREADS]; bool ret_val[NUM_THREADS]; for (int i = 0; i < NUM_THREADS; i++) Thrd[i] = std::thread(Thread_func, std::ref(ret_val[i])); diff --git a/projects/hip-tests/catch/unit/memory/hipMemset3DRegressMultiThread.cc b/projects/hip-tests/catch/unit/memory/hipMemset3DRegressMultiThread.cc index 7af99d132c..ce535ec9a2 100644 --- a/projects/hip-tests/catch/unit/memory/hipMemset3DRegressMultiThread.cc +++ b/projects/hip-tests/catch/unit/memory/hipMemset3DRegressMultiThread.cc @@ -186,7 +186,7 @@ bool loopRegression(bool bAsync) { * Perform regression of hipMemset3D api with device memory allocated * on different gpus. */ -TEST_CASE("Unit_hipMemset3D_RegressInLoop") { +TEST_CASE("Unit_hipMemset3D_RegressInLoop", "[multigpu]") { CHECK_IMAGE_SUPPORT bool TestPassed = false; @@ -199,7 +199,7 @@ TEST_CASE("Unit_hipMemset3D_RegressInLoop") { * Perform regression of hipMemset3DAsync api with device memory allocated * on different gpus. */ -TEST_CASE("Unit_hipMemset3DAsync_RegressInLoop") { +TEST_CASE("Unit_hipMemset3DAsync_RegressInLoop", "[multigpu]") { CHECK_IMAGE_SUPPORT bool TestPassed = false; diff --git a/projects/hip-tests/catch/unit/memory/hipPointerGetAttribute.cc b/projects/hip-tests/catch/unit/memory/hipPointerGetAttribute.cc index 97fc15237e..63659a81a9 100644 --- a/projects/hip-tests/catch/unit/memory/hipPointerGetAttribute.cc +++ b/projects/hip-tests/catch/unit/memory/hipPointerGetAttribute.cc @@ -150,7 +150,7 @@ TEST_CASE("Unit_hipPointerGetAttribute_KernelUpdation") { * from peer GPU device.It validates the memory type and * device ordinal in peer GPU */ -TEST_CASE("Unit_hipPointerGetAttribute_PeerGPU") { +TEST_CASE("Unit_hipPointerGetAttribute_PeerGPU", "[multigpu]") { HIP_CHECK(hipSetDevice(0)); size_t Nbytes = 0; Nbytes = N * sizeof(int); diff --git a/projects/hip-tests/catch/unit/memory/hipPointerGetAttributes.cc b/projects/hip-tests/catch/unit/memory/hipPointerGetAttributes.cc index bff4ca2333..792ff0f6a0 100644 --- a/projects/hip-tests/catch/unit/memory/hipPointerGetAttributes.cc +++ b/projects/hip-tests/catch/unit/memory/hipPointerGetAttributes.cc @@ -310,7 +310,7 @@ TEST_CASE("Unit_hipPointerGetAttributes_Basic") { * - HIP_VERSION >= 5.7 */ -TEST_CASE("Unit_hipPointerGetAttributes_ClusterAlloc") { +TEST_CASE("Unit_hipPointerGetAttributes_ClusterAlloc", "[multigpu]") { srand(0x100); printf("\n=============================================\n"); clusterAllocs(100, 1024 * 1, 1024 * 1024); @@ -327,7 +327,7 @@ TEST_CASE("Unit_hipPointerGetAttributes_ClusterAlloc") { * - HIP_VERSION >= 5.7 */ -TEST_CASE("Unit_hipPointerGetAttributes_TinyClusterAlloc") { +TEST_CASE("Unit_hipPointerGetAttributes_TinyClusterAlloc", "[multigpu]") { srand(0x200); printf("\n=============================================\n"); clusterAllocs(1000, 1, 10); // Many tiny allocations; @@ -337,7 +337,7 @@ TEST_CASE("Unit_hipPointerGetAttributes_TinyClusterAlloc") { // IN : serialize will force the test to run in serial fashion. #if 0 // FIXME_jatinx These need to be ported to HIP_CHECK_THREAD. Disabling it for now -TEST_CASE("Unit_hipPointerGetAttributes_MultiThread") { +TEST_CASE("Unit_hipPointerGetAttributes_MultiThread", "[multigpu]") { srand(0x300); auto serialize = 1; printf("\n=============================================\n"); @@ -393,7 +393,7 @@ TEST_CASE("Unit_hipPointerGetAttributes_Negative") { * ------------------------ * - HIP_VERSION >= 6.0 */ -TEST_CASE("Unit_hipPointerGetAttributes_GpuIter") { +TEST_CASE("Unit_hipPointerGetAttributes_GpuIter", "[multigpu]") { int deviceCount{0}; HIP_CHECK(hipGetDeviceCount(&deviceCount)); REQUIRE(deviceCount != 0); @@ -453,7 +453,8 @@ TEST_CASE("Unit_hipPointerGetAttributes_GpuIter") { * ------------------------ * - HIP_VERSION >= 6.0 */ -TEST_CASE("Unit_hipPointerGetAttributes_GpuIter_Managed__Memory") { +TEST_CASE("Unit_hipPointerGetAttributes_GpuIter_Managed__Memory", + "[multigpu]") { int deviceCount{0}; HIP_CHECK(hipGetDeviceCount(&deviceCount)); REQUIRE(deviceCount != 0); @@ -490,7 +491,8 @@ TEST_CASE("Unit_hipPointerGetAttributes_GpuIter_Managed__Memory") { * ------------------------ * - HIP_VERSION >= 6.0 */ -TEST_CASE("Unit_hipPointerGetAttributes_GpuIter_Unregistered_Memory") { +TEST_CASE("Unit_hipPointerGetAttributes_GpuIter_Unregistered_Memory", + "[multigpu]") { int deviceCount{0}; HIP_CHECK(hipGetDeviceCount(&deviceCount)); REQUIRE(deviceCount != 0); diff --git a/projects/hip-tests/catch/unit/memory/hipPtrGetAttribute.cc b/projects/hip-tests/catch/unit/memory/hipPtrGetAttribute.cc index 3ab24f8ec4..6698e19021 100644 --- a/projects/hip-tests/catch/unit/memory/hipPtrGetAttribute.cc +++ b/projects/hip-tests/catch/unit/memory/hipPtrGetAttribute.cc @@ -29,7 +29,7 @@ THE SOFTWARE. #include // Run few simple cases including host pointer arithmetic: -TEST_CASE("Unit_hipPtrGetAttribute_Simple") { +TEST_CASE("Unit_hipPtrGetAttribute_Simple", "[multigpu]") { HIP_CHECK(hipSetDevice(0)); size_t Nbytes = 0; constexpr size_t N{1000000}; diff --git a/projects/hip-tests/catch/unit/memory/hipSVMTestByteGranularity.cpp b/projects/hip-tests/catch/unit/memory/hipSVMTestByteGranularity.cpp index 512c554d2d..343a235501 100644 --- a/projects/hip-tests/catch/unit/memory/hipSVMTestByteGranularity.cpp +++ b/projects/hip-tests/catch/unit/memory/hipSVMTestByteGranularity.cpp @@ -72,7 +72,7 @@ __global__ void sum_neighbor_locations(char* a, unsigned int num_devices, * - Fine grain access and atomics supported on devices * - HIP_VERSION >= 5.7 */ -TEST_CASE("test_svm_byte_granularity") { +TEST_CASE("test_svm_byte_granularity", "[multigpu]") { int pcieAtomic = 0; HIP_CHECK(hipDeviceGetAttribute(&pcieAtomic, hipDeviceAttributeHostNativeAtomicSupported, 0)); if (!pcieAtomic) { diff --git a/projects/hip-tests/catch/unit/memory/hipSVMTestFineGrainMemoryConsistency.cpp b/projects/hip-tests/catch/unit/memory/hipSVMTestFineGrainMemoryConsistency.cpp index ed990a1d80..0f7505873d 100644 --- a/projects/hip-tests/catch/unit/memory/hipSVMTestFineGrainMemoryConsistency.cpp +++ b/projects/hip-tests/catch/unit/memory/hipSVMTestFineGrainMemoryConsistency.cpp @@ -227,7 +227,7 @@ void launch_kernels_and_verify(std::vector& streams, unsigned int n * - Fine grain access and atomics supported on devices and host * - HIP_VERSION >= 5.7 */ -TEST_CASE("test_svm_fine_grain_memory_consistency") { +TEST_CASE("test_svm_fine_grain_memory_consistency", "[multigpu]") { int num_devices = 0; HIP_CHECK(hipGetDeviceCount(&num_devices)); diff --git a/projects/hip-tests/catch/unit/memory/hipSVMTestSharedAddressSpaceFineGrain.cpp b/projects/hip-tests/catch/unit/memory/hipSVMTestSharedAddressSpaceFineGrain.cpp index e00fc58792..d7fee91325 100644 --- a/projects/hip-tests/catch/unit/memory/hipSVMTestSharedAddressSpaceFineGrain.cpp +++ b/projects/hip-tests/catch/unit/memory/hipSVMTestSharedAddressSpaceFineGrain.cpp @@ -157,7 +157,7 @@ void verify_linked_lists_on_device(hipStream_t stream, Node* pNodes, unsigned in * - Fine grain access supported on devices and host * - HIP_VERSION >= 5.7 */ -TEST_CASE("test_svm_shared_address_space_fine_grain_buffers") { +TEST_CASE("test_svm_shared_address_space_fine_grain_buffers", "[multigpu]") { const unsigned int num_elements = 1024; int num_devices = 0; HIP_CHECK(hipGetDeviceCount(&num_devices)); @@ -247,7 +247,7 @@ TEST_CASE("test_svm_shared_address_space_fine_grain_buffers") { * - System fine grain access supported on devices * - HIP_VERSION >= 5.7 */ -TEST_CASE("test_svm_shared_address_space_fine_grain_system") { +TEST_CASE("test_svm_shared_address_space_fine_grain_system", "[multigpu]") { int num_devices = 0; HIP_CHECK(hipGetDeviceCount(&num_devices)); diff --git a/projects/hip-tests/catch/unit/memory/hipStreamAttachMemAsync.cc b/projects/hip-tests/catch/unit/memory/hipStreamAttachMemAsync.cc index f2b0a9a0ea..d20667e3cb 100644 --- a/projects/hip-tests/catch/unit/memory/hipStreamAttachMemAsync.cc +++ b/projects/hip-tests/catch/unit/memory/hipStreamAttachMemAsync.cc @@ -61,7 +61,7 @@ TEST_CASE("Unit_hipStreamAttachMemAsync_Positive_Pageable") { // CUDA docs: // If the cudaMemAttachGlobal flag is specified, the memory can be accessed by any stream on any // device. -TEST_CASE("Unit_hipStreamAttachMemAsync_Positive_AttachGlobal") { +TEST_CASE("Unit_hipStreamAttachMemAsync_Positive_AttachGlobal", "[multigpu]") { if (!DeviceAttributesSupport(0, hipDeviceAttributeManagedMemory)) { HipTest::HIP_SKIP_TEST("Managed memory is not supported"); return; diff --git a/projects/hip-tests/catch/unit/module/hipExtLaunchMultiKernelMultiDevice.cc b/projects/hip-tests/catch/unit/module/hipExtLaunchMultiKernelMultiDevice.cc index 8302b71548..c1c4e36d86 100644 --- a/projects/hip-tests/catch/unit/module/hipExtLaunchMultiKernelMultiDevice.cc +++ b/projects/hip-tests/catch/unit/module/hipExtLaunchMultiKernelMultiDevice.cc @@ -53,7 +53,7 @@ __global__ void vector_square(float* C_d, float* A_d, size_t N) { } } -TEST_CASE("Unit_hipExtLaunchMultiKernelMultiDevice_Functional") { +TEST_CASE("Unit_hipExtLaunchMultiKernelMultiDevice_Functional", "[multigpu]") { constexpr int MAX_GPUS = 8; float *A_d[MAX_GPUS], *C_d[MAX_GPUS]; float *A_h, *C_h; diff --git a/projects/hip-tests/catch/unit/module/hipExtModuleLaunchKernel.cc b/projects/hip-tests/catch/unit/module/hipExtModuleLaunchKernel.cc index 2df5dc3ae6..b25617e7ea 100644 --- a/projects/hip-tests/catch/unit/module/hipExtModuleLaunchKernel.cc +++ b/projects/hip-tests/catch/unit/module/hipExtModuleLaunchKernel.cc @@ -277,7 +277,7 @@ TEST_CASE("Unit_hipExtModuleLaunchKernel_Positive_Parameters") { } } -TEST_CASE("Unit_hipExtModuleLaunchKernel_Negative_Parameters") { +TEST_CASE("Unit_hipExtModuleLaunchKernel_Negative_Parameters", "[multigpu]") { ModuleLaunchKernelNegativeParameters(true); } /** diff --git a/projects/hip-tests/catch/unit/module/hipGetFuncBySymbol.cc b/projects/hip-tests/catch/unit/module/hipGetFuncBySymbol.cc index e8de712ed7..becd95e0df 100644 --- a/projects/hip-tests/catch/unit/module/hipGetFuncBySymbol.cc +++ b/projects/hip-tests/catch/unit/module/hipGetFuncBySymbol.cc @@ -207,7 +207,7 @@ TEST_CASE("Unit_hipGetFuncBySymbol_InChildProcess") { * ------------------------ * - HIP_VERSION >= 6.2 */ -TEST_CASE("Unit_hipGetFuncBySymbol_MultiDev") { +TEST_CASE("Unit_hipGetFuncBySymbol_MultiDev", "[multigpu]") { int deviceCount = 0; HIP_CHECK(hipGetDeviceCount(&deviceCount)); if (deviceCount < 2) { @@ -323,7 +323,7 @@ void MultiThreadMultiDevFunc(int DevId) { * ------------------------ * - HIP_VERSION >= 6.2 */ -TEST_CASE("Unit_hipGetFuncBySymbol_MultiDevMultiThread") { +TEST_CASE("Unit_hipGetFuncBySymbol_MultiDevMultiThread", "[multigpu]") { int deviceCount = 0; HIP_CHECK(hipGetDeviceCount(&deviceCount)); if (deviceCount < 2) { diff --git a/projects/hip-tests/catch/unit/module/hipGetProcAddressModuleApis.cc b/projects/hip-tests/catch/unit/module/hipGetProcAddressModuleApis.cc index 7af86ef31d..dbd883f6e7 100644 --- a/projects/hip-tests/catch/unit/module/hipGetProcAddressModuleApis.cc +++ b/projects/hip-tests/catch/unit/module/hipGetProcAddressModuleApis.cc @@ -328,7 +328,7 @@ TEST_CASE("Unit_hipGetProcAddress_ModuleApisLoadData") { * ------------------------ * - HIP_VERSION >= 6.2 */ -TEST_CASE("Unit_hipGetProcAddress_ModuleApisCooperativeKernels") { +TEST_CASE("Unit_hipGetProcAddress_ModuleApisCooperativeKernels", "[multigpu]") { if (!DeviceAttributesSupport(0, hipDeviceAttributeCooperativeLaunch)) { HipTest::HIP_SKIP_TEST("CooperativeLaunch not supported"); return; diff --git a/projects/hip-tests/catch/unit/module/hipManagedKeyword.cc b/projects/hip-tests/catch/unit/module/hipManagedKeyword.cc index 0f1e86c7ff..902a71bfcd 100644 --- a/projects/hip-tests/catch/unit/module/hipManagedKeyword.cc +++ b/projects/hip-tests/catch/unit/module/hipManagedKeyword.cc @@ -43,7 +43,7 @@ constexpr auto fileName = "managed_kernel.code"; * - HIP_VERSION >= 5.6 */ -TEST_CASE("Unit_hipModuleGetGlobal_Functional") { +TEST_CASE("Unit_hipModuleGetGlobal_Functional", "[multigpu]") { bool testStatus = true; int numDevices = 0; hipDeviceptr_t x; diff --git a/projects/hip-tests/catch/unit/module/hipModuleGetFunction.cc b/projects/hip-tests/catch/unit/module/hipModuleGetFunction.cc index 3686d1da09..2ccf0b8758 100644 --- a/projects/hip-tests/catch/unit/module/hipModuleGetFunction.cc +++ b/projects/hip-tests/catch/unit/module/hipModuleGetFunction.cc @@ -71,7 +71,7 @@ TEST_CASE("Unit_hipModuleGetFunction_Negative_Parameters") { // Test description: Loading kernel function from different device than the one on which the module // is loaded -TEST_CASE("Unit_hipModuleGetFunction_DiffDevice") { +TEST_CASE("Unit_hipModuleGetFunction_DiffDevice", "[multigpu]") { int numDevices = 0; HIP_CHECK(hipGetDeviceCount(&numDevices)); if (numDevices < 2) { diff --git a/projects/hip-tests/catch/unit/module/hipModuleGetGlobal.cc b/projects/hip-tests/catch/unit/module/hipModuleGetGlobal.cc index d7cd231494..c3dd06fd0d 100644 --- a/projects/hip-tests/catch/unit/module/hipModuleGetGlobal.cc +++ b/projects/hip-tests/catch/unit/module/hipModuleGetGlobal.cc @@ -148,7 +148,7 @@ TEST_CASE("Unit_hipModuleGetGlobal_Negative_Dptr_And_Bytes_Are_Nullptr") { // Test description: Loading device ptr from different device than the one on which the module // is loaded -TEST_CASE("Unit_hipModuleGetGlobal_DiffDevice") { +TEST_CASE("Unit_hipModuleGetGlobal_DiffDevice", "[multigpu]") { int numDevices = 0; HIP_CHECK(hipGetDeviceCount(&numDevices)); if (numDevices < 2) { diff --git a/projects/hip-tests/catch/unit/module/hipModuleLaunchCooperativeKernelMultiDevice.cc b/projects/hip-tests/catch/unit/module/hipModuleLaunchCooperativeKernelMultiDevice.cc index d9740b13cf..f8680ff49c 100644 --- a/projects/hip-tests/catch/unit/module/hipModuleLaunchCooperativeKernelMultiDevice.cc +++ b/projects/hip-tests/catch/unit/module/hipModuleLaunchCooperativeKernelMultiDevice.cc @@ -48,7 +48,8 @@ THE SOFTWARE. * ------------------------ * - HIP_VERSION >= 5.5 */ -TEST_CASE("Unit_hipModuleLaunchCooperativeKernelMultiDevice_Positive_Basic") { +TEST_CASE("Unit_hipModuleLaunchCooperativeKernelMultiDevice_Positive_Basic", + "[multigpu]") { auto mg = ModuleGuard::InitModule("launch_kernel_module.code"); if (!DeviceAttributesSupport(0, hipDeviceAttributeCooperativeLaunch)) { HipTest::HIP_SKIP_TEST("CooperativeLaunch not supported"); @@ -106,7 +107,9 @@ TEST_CASE("Unit_hipModuleLaunchCooperativeKernelMultiDevice_Positive_Basic") { * ------------------------ * - HIP_VERSION >= 5.5 */ -TEST_CASE("Unit_hipModuleLaunchCooperativeKernelMultiDevice_Negative_Parameters") { +TEST_CASE( + "Unit_hipModuleLaunchCooperativeKernelMultiDevice_Negative_Parameters", + "[multigpu]") { auto mg = ModuleGuard::InitModule("launch_kernel_module.code"); if (!DeviceAttributesSupport(0, hipDeviceAttributeCooperativeLaunch)) { HipTest::HIP_SKIP_TEST("CooperativeLaunch not supported"); @@ -230,7 +233,9 @@ TEST_CASE("Unit_hipModuleLaunchCooperativeKernelMultiDevice_Negative_Parameters" * ------------------------ * - HIP_VERSION >= 5.5 */ -TEST_CASE("Unit_hipModuleLaunchCooperativeKernelMultiDevice_Negative_MultiKernelSameDevice") { +TEST_CASE("Unit_hipModuleLaunchCooperativeKernelMultiDevice_Negative_" + "MultiKernelSameDevice", + "[multigpu]") { auto mg = ModuleGuard::InitModule("launch_kernel_module.code"); if (!DeviceAttributesSupport(0, hipDeviceAttributeCooperativeLaunch)) { HipTest::HIP_SKIP_TEST("CooperativeLaunch not supported"); diff --git a/projects/hip-tests/catch/unit/module/hipModuleLaunchKernel.cc b/projects/hip-tests/catch/unit/module/hipModuleLaunchKernel.cc index 82193ebf15..93c6a37962 100644 --- a/projects/hip-tests/catch/unit/module/hipModuleLaunchKernel.cc +++ b/projects/hip-tests/catch/unit/module/hipModuleLaunchKernel.cc @@ -42,7 +42,7 @@ TEST_CASE("Unit_hipModuleLaunchKernel_Positive_Parameters") { ModuleLaunchKernelPositiveParameters(); } -TEST_CASE("Unit_hipModuleLaunchKernel_Negative_Parameters") { +TEST_CASE("Unit_hipModuleLaunchKernel_Negative_Parameters", "[multigpu]") { HIP_CHECK(hipFree(nullptr)); ModuleLaunchKernelNegativeParameters(); } diff --git a/projects/hip-tests/catch/unit/module/hipModuleLoadMultProcessOnMultGPU.cc b/projects/hip-tests/catch/unit/module/hipModuleLoadMultProcessOnMultGPU.cc index 243284c92f..625549138b 100644 --- a/projects/hip-tests/catch/unit/module/hipModuleLoadMultProcessOnMultGPU.cc +++ b/projects/hip-tests/catch/unit/module/hipModuleLoadMultProcessOnMultGPU.cc @@ -44,7 +44,7 @@ THE SOFTWARE. * ------------------------ * - HIP_VERSION >= 5.6 */ -TEST_CASE("Unit_hipModuleLoad_MultProcess_MultGPU") { +TEST_CASE("Unit_hipModuleLoad_MultProcess_MultGPU", "[multigpu]") { int deviceCount{0}; HIP_CHECK(hipGetDeviceCount(&deviceCount)); REQUIRE(deviceCount != 0); diff --git a/projects/hip-tests/catch/unit/occupancy/hipOccupancyMaxPotentialBlockSizeVariableSMemWithFlags.cc b/projects/hip-tests/catch/unit/occupancy/hipOccupancyMaxPotentialBlockSizeVariableSMemWithFlags.cc index 01081d04df..7ff7f7e7f8 100644 --- a/projects/hip-tests/catch/unit/occupancy/hipOccupancyMaxPotentialBlockSizeVariableSMemWithFlags.cc +++ b/projects/hip-tests/catch/unit/occupancy/hipOccupancyMaxPotentialBlockSizeVariableSMemWithFlags.cc @@ -108,7 +108,8 @@ TEST_CASE("Unit_hipOccupancyMaxPotBlkSizeVariableSMemWithFlags_chkRange") { - for 0 < block_size_limit < attr.maxThreadsPerBlock - for block_size_limit > attr.maxThreadsPerBlock */ -TEST_CASE("Unit_hipOccupancyMaxPotBlkSizeVariableSMemWithFlags_mgpu") { +TEST_CASE("Unit_hipOccupancyMaxPotBlkSizeVariableSMemWithFlags_mgpu", + "[multigpu]") { int devcount = 0; HIP_CHECK(hipGetDeviceCount(&devcount)); // If only single GPU is detected then return diff --git a/projects/hip-tests/catch/unit/printf/hipPrintfManyDevices.cc b/projects/hip-tests/catch/unit/printf/hipPrintfManyDevices.cc index 9277e2716c..f5441121e4 100644 --- a/projects/hip-tests/catch/unit/printf/hipPrintfManyDevices.cc +++ b/projects/hip-tests/catch/unit/printf/hipPrintfManyDevices.cc @@ -47,7 +47,7 @@ __global__ void print_things() { * ------------------------ * - HIP_VERSION >= 6.2 */ -TEST_CASE("Unit_Printf_ManyDevicesTest") { +TEST_CASE("Unit_Printf_ManyDevicesTest", "[multigpu]") { int pcieAtomic = 0; HIP_CHECK(hipDeviceGetAttribute(&pcieAtomic, hipDeviceAttributeHostNativeAtomicSupported, 0)); if (!pcieAtomic) { diff --git a/projects/hip-tests/catch/unit/stream/hipLaunchHostFunc.cc b/projects/hip-tests/catch/unit/stream/hipLaunchHostFunc.cc index 6d18d859f8..cb474e57ab 100644 --- a/projects/hip-tests/catch/unit/stream/hipLaunchHostFunc.cc +++ b/projects/hip-tests/catch/unit/stream/hipLaunchHostFunc.cc @@ -313,7 +313,7 @@ TEST_CASE("Unit_hipLaunchHostFunc_KernelHost") { // Test scenario 5 // scenario that validates the host launch function on multi device // environment. -TEST_CASE("Unit_hipLaunchHostFunc_multidevice") { +TEST_CASE("Unit_hipLaunchHostFunc_multidevice", "[multigpu]") { int num_devices; HIP_CHECK(hipGetDeviceCount(&num_devices)); if (num_devices < 2) { diff --git a/projects/hip-tests/catch/unit/stream/hipStreamGetDevice.cc b/projects/hip-tests/catch/unit/stream/hipStreamGetDevice.cc index 9671d49dd2..5c9de18529 100644 --- a/projects/hip-tests/catch/unit/stream/hipStreamGetDevice.cc +++ b/projects/hip-tests/catch/unit/stream/hipStreamGetDevice.cc @@ -74,7 +74,7 @@ TEST_CASE("Unit_hipStreamGetDevice_Negative") { * - HIP_VERSION >= 5.6 */ -TEST_CASE("Unit_hipStreamGetDevice_Usecase") { +TEST_CASE("Unit_hipStreamGetDevice_Usecase", "[multigpu]") { int device_count = 0; HIP_CHECK(hipGetDeviceCount(&device_count)); REQUIRE(device_count != 0); @@ -186,7 +186,7 @@ TEST_CASE("Unit_hipStreamGetDevice_MThread") { REQUIRE(true == test_hipStreamGet * - HIP_VERSION >= 5.6 */ -TEST_CASE("Unit_hipStreamGetDevice_SetDiffDevice") { +TEST_CASE("Unit_hipStreamGetDevice_SetDiffDevice", "[multigpu]") { hipDevice_t device_from_stream; int device_count = 0; HIP_CHECK(hipGetDeviceCount(&device_count)); @@ -225,7 +225,7 @@ TEST_CASE("Unit_hipStreamGetDevice_SetDiffDevice") { * Test to be run only on AMD machine as it's failing in CUDA. */ #if HT_AMD -TEST_CASE("Unit_hipStreamGetDevice_NullStream") { +TEST_CASE("Unit_hipStreamGetDevice_NullStream", "[multigpu]") { int device_count = 0; HIP_CHECK(hipGetDeviceCount(&device_count)); REQUIRE(device_count != 0); diff --git a/projects/hip-tests/catch/unit/stream/hipStreamGetId.cc b/projects/hip-tests/catch/unit/stream/hipStreamGetId.cc index 9c1e6798fc..1da810b678 100644 --- a/projects/hip-tests/catch/unit/stream/hipStreamGetId.cc +++ b/projects/hip-tests/catch/unit/stream/hipStreamGetId.cc @@ -191,7 +191,7 @@ TEST_CASE("Unit_hipStreamGetId_MultipleThreads") { * ------------------------ * - HIP_VERSION >= 7.1 */ -TEST_CASE("Unit_hipStreamGetId_MultiDevice") { +TEST_CASE("Unit_hipStreamGetId_MultiDevice", "[multigpu]") { int deviceCount = 0; HIP_CHECK(hipGetDeviceCount(&deviceCount)); if (deviceCount < 2) { diff --git a/projects/hip-tests/catch/unit/stream/hipStreamLegacy_Ext.cc b/projects/hip-tests/catch/unit/stream/hipStreamLegacy_Ext.cc index abde78deb3..e8aa3193f7 100644 --- a/projects/hip-tests/catch/unit/stream/hipStreamLegacy_Ext.cc +++ b/projects/hip-tests/catch/unit/stream/hipStreamLegacy_Ext.cc @@ -295,7 +295,7 @@ TEST_CASE("Unit_hipStreamLegacy_WithStreamPerThread") { * ------------------------ * - HIP_VERSION >= 6.3 */ -TEST_CASE("Unit_hipStreamLegacy_MultiDevice") { +TEST_CASE("Unit_hipStreamLegacy_MultiDevice", "[multigpu]") { int deviceCount = 0; HIP_CHECK(hipGetDeviceCount(&deviceCount)); if (deviceCount < 2) { @@ -418,7 +418,7 @@ TEST_CASE("Unit_hipStreamLegacy_H2H_H2D_D2D_D2H_Default") { * ------------------------ * - HIP_VERSION >= 6.3 */ -TEST_CASE("Unit_hipStreamLegacy_MultiDeviceMultiOperation") { +TEST_CASE("Unit_hipStreamLegacy_MultiDeviceMultiOperation", "[multigpu]") { int deviceCount = 0; HIP_CHECK(hipGetDeviceCount(&deviceCount)); if (deviceCount < 2) { @@ -594,7 +594,7 @@ TEST_CASE("Unit_hipStreamLegacy_TwoThreadsEachOneDiffOperation") { * ------------------------ * - HIP_VERSION >= 6.3 */ -TEST_CASE("Unit_hipStreamLegacy_TwoDevicesEachOneDiffOperation") { +TEST_CASE("Unit_hipStreamLegacy_TwoDevicesEachOneDiffOperation", "[multigpu]") { int deviceCount = 0; HIP_CHECK(hipGetDeviceCount(&deviceCount)); if (deviceCount < 2) { @@ -679,7 +679,8 @@ static void operationsInDev1(int* devArrDev1, int* hostArrDst) { * ------------------------ * - HIP_VERSION >= 6.3 */ -TEST_CASE("Unit_hipStreamLegacy_TwoThreadsInTwoDevicesEachOneDiffOperation") { +TEST_CASE("Unit_hipStreamLegacy_TwoThreadsInTwoDevicesEachOneDiffOperation", + "[multigpu]") { int deviceCount = 0; HIP_CHECK(hipGetDeviceCount(&deviceCount)); if (deviceCount < 2) { diff --git a/projects/hip-tests/catch/unit/synchronization/cache_coherency_gpu_gpu.cc b/projects/hip-tests/catch/unit/synchronization/cache_coherency_gpu_gpu.cc index 751fa88684..eda2cc4651 100644 --- a/projects/hip-tests/catch/unit/synchronization/cache_coherency_gpu_gpu.cc +++ b/projects/hip-tests/catch/unit/synchronization/cache_coherency_gpu_gpu.cc @@ -284,7 +284,7 @@ static bool gpu_to_gpu_coherency() { * - Test to be run only on AMD. */ -TEST_CASE("Unit_cache_coherency_gpu_gpu") { +TEST_CASE("Unit_cache_coherency_gpu_gpu", "[multigpu]") { bool passed = true; // Coherency between GPUs accessing local or remote FB. REQUIRE(passed == gpu_to_gpu_coherency()); diff --git a/projects/hip-tests/catch/unit/threadfence/__threadfence.cc b/projects/hip-tests/catch/unit/threadfence/__threadfence.cc index fa58e9b50f..0013866689 100644 --- a/projects/hip-tests/catch/unit/threadfence/__threadfence.cc +++ b/projects/hip-tests/catch/unit/threadfence/__threadfence.cc @@ -164,7 +164,7 @@ TEST_CASE("Unit___threadfence_Positive_Basic_Managed") { * ------------------------ * - HIP_VERSION >= 5.2 */ -TEST_CASE("Unit___threadfence_Positive_Basic_Peer") { +TEST_CASE("Unit___threadfence_Positive_Basic_Peer", "[multigpu]") { const auto device_count = HipTest::getDeviceCount(); if (device_count < 2) { HipTest::HIP_SKIP_TEST("At least 2 devices are required"); diff --git a/projects/hip-tests/catch/unit/threadfence/__threadfence_block.cc b/projects/hip-tests/catch/unit/threadfence/__threadfence_block.cc index 757cb55905..23e8223633 100644 --- a/projects/hip-tests/catch/unit/threadfence/__threadfence_block.cc +++ b/projects/hip-tests/catch/unit/threadfence/__threadfence_block.cc @@ -164,7 +164,7 @@ TEST_CASE("Unit___threadfence_block_Positive_Basic_Managed") { * ------------------------ * - HIP_VERSION >= 5.2 */ -TEST_CASE("Unit___threadfence_block_Positive_Basic_Peer") { +TEST_CASE("Unit___threadfence_block_Positive_Basic_Peer", "[multigpu]") { const auto device_count = HipTest::getDeviceCount(); if (device_count < 2) { HipTest::HIP_SKIP_TEST("At least 2 devices are required"); diff --git a/projects/hip-tests/catch/unit/threadfence/__threadfence_system.cc b/projects/hip-tests/catch/unit/threadfence/__threadfence_system.cc index 74cc803b35..9b54b0b8c1 100644 --- a/projects/hip-tests/catch/unit/threadfence/__threadfence_system.cc +++ b/projects/hip-tests/catch/unit/threadfence/__threadfence_system.cc @@ -60,7 +60,7 @@ __global__ void ReadKernel(int* out, int* in) { * ------------------------ * - HIP_VERSION >= 5.2 */ -TEST_CASE("Unit___threadfence_system_Positive_Basic_Peer") { +TEST_CASE("Unit___threadfence_system_Positive_Basic_Peer", "[multigpu]") { const auto device_count = HipTest::getDeviceCount(); if (device_count < 2) { HipTest::HIP_SKIP_TEST("At least 2 devices are required"); diff --git a/projects/hip-tests/catch/unit/virtualMemoryManagement/hipMemGetHandleForAddressRange.cc b/projects/hip-tests/catch/unit/virtualMemoryManagement/hipMemGetHandleForAddressRange.cc index ea4ccc4a33..027f8f6e61 100644 --- a/projects/hip-tests/catch/unit/virtualMemoryManagement/hipMemGetHandleForAddressRange.cc +++ b/projects/hip-tests/catch/unit/virtualMemoryManagement/hipMemGetHandleForAddressRange.cc @@ -362,7 +362,8 @@ TEST_CASE("Unit_hipMemGetHandleForAddressRange_VM") { * ------------------------ * - HIP_VERSION >= 7.0 */ -TEST_CASE("Unit_hipMemGetHandleForAddressRange_DeviceMemory_InAnotherDevice") { +TEST_CASE("Unit_hipMemGetHandleForAddressRange_DeviceMemory_InAnotherDevice", + "[multigpu]") { int deviceCount = 0; HIP_CHECK(hipGetDeviceCount(&deviceCount)); if (deviceCount < 2) { @@ -414,7 +415,8 @@ TEST_CASE("Unit_hipMemGetHandleForAddressRange_DeviceMemory_InAnotherDevice") { * ------------------------ * - HIP_VERSION >= 7.0 */ -TEST_CASE("Unit_hipMemGetHandleForAddressRange_VM_InAnotherDevice") { +TEST_CASE("Unit_hipMemGetHandleForAddressRange_VM_InAnotherDevice", + "[multigpu]") { int deviceCount = 0; HIP_CHECK(hipGetDeviceCount(&deviceCount)); if (deviceCount < 2) { diff --git a/projects/hip-tests/catch/unit/virtualMemoryManagement/hipMemMap.cc b/projects/hip-tests/catch/unit/virtualMemoryManagement/hipMemMap.cc index aa51b35224..63c0075713 100644 --- a/projects/hip-tests/catch/unit/virtualMemoryManagement/hipMemMap.cc +++ b/projects/hip-tests/catch/unit/virtualMemoryManagement/hipMemMap.cc @@ -364,7 +364,7 @@ void physicalMemoryReuse_MultiDev (hipMemAllocationProp prop) { * ------------------------ * - HIP_VERSION >= 7.0 */ -TEST_CASE("Unit_hipMemMap_PhysicalMemoryReuse_MultiDev") { +TEST_CASE("Unit_hipMemMap_PhysicalMemoryReuse_MultiDev", "[multigpu]") { CHECK_P2P_SUPPORT SECTION("Memory Allocation Type as hipMemAllocationTypePinned") { hipMemAllocationProp prop{}; @@ -554,7 +554,7 @@ void vMMMemoryReuse_MultiGPU (hipMemAllocationProp prop) { * ------------------------ * - HIP_VERSION >= 7.0 */ -TEST_CASE("Unit_hipMemMap_VMMMemoryReuse_MultiGPU") { +TEST_CASE("Unit_hipMemMap_VMMMemoryReuse_MultiGPU", "[multigpu]") { CHECK_P2P_SUPPORT SECTION("Memory Allocation Type as hipMemAllocationTypePinned") { hipMemAllocationProp prop{}; diff --git a/projects/hip-tests/catch/unit/virtualMemoryManagement/hipMemSetGetAccess.cc b/projects/hip-tests/catch/unit/virtualMemoryManagement/hipMemSetGetAccess.cc index b61f0d0b6c..8dbd329800 100644 --- a/projects/hip-tests/catch/unit/virtualMemoryManagement/hipMemSetGetAccess.cc +++ b/projects/hip-tests/catch/unit/virtualMemoryManagement/hipMemSetGetAccess.cc @@ -340,7 +340,7 @@ TEST_CASE("Unit_hipMemGetAccess_NegTst") { * ------------------------ * - HIP_VERSION >= 6.1 */ -TEST_CASE("Unit_hipMemSetAccess_FuncTstOnMultDev") { +TEST_CASE("Unit_hipMemSetAccess_FuncTstOnMultDev", "[multigpu]") { size_t granularity = 0; constexpr int N = DATA_SIZE; size_t buffer_size = N * sizeof(int); @@ -638,7 +638,7 @@ TEST_CASE("Unit_hipMemSetAccess_Vmm2DevMemCpy") { * ------------------------ * - HIP_VERSION >= 6.1 */ -TEST_CASE("Unit_hipMemSetAccess_Vmm2PeerDevMemCpy") { +TEST_CASE("Unit_hipMemSetAccess_Vmm2PeerDevMemCpy", "[multigpu]") { size_t granularity = 0; constexpr int N = DATA_SIZE; size_t buffer_size = N * sizeof(int); @@ -729,7 +729,7 @@ TEST_CASE("Unit_hipMemSetAccess_Vmm2PeerDevMemCpy") { * ------------------------ * - HIP_VERSION >= 6.1 */ -TEST_CASE("Unit_hipMemSetAccess_Vmm2PeerPeerMemCpy") { +TEST_CASE("Unit_hipMemSetAccess_Vmm2PeerPeerMemCpy", "[multigpu]") { size_t granularity = 0; constexpr int N = DATA_SIZE; size_t buffer_size = N * sizeof(int); @@ -891,7 +891,7 @@ TEST_CASE("Unit_hipMemSetAccess_Vmm2VMMMemCpy") { * ------------------------ * - HIP_VERSION >= 6.1 */ -TEST_CASE("Unit_hipMemSetAccess_Vmm2VMMInterDevMemCpy") { +TEST_CASE("Unit_hipMemSetAccess_Vmm2VMMInterDevMemCpy", "[multigpu]") { size_t granularity = 0; constexpr int N = DATA_SIZE; size_t buffer_size = N * sizeof(int);