AI/rocm-systems
2
0
포크 0
코드 이슈 풀 리퀘스트 액션 패키지 프로젝트 릴리즈 위키 활동

kfdtest: Improve the printing message for CuMasking tests

소스 검색 
Decimal is better than hex in this case.

Change-Id: Ic15a9373e99160880b98d3dcd6827d551c87b77a
Signed-off-by: Yong Zhao <Yong.Zhao@amd.com>


[ROCm/ROCR-Runtime commit: 3e4c42ef13]
이 커밋은 다음에 포함됨:
Yong Zhao
2019-07-30 20:19:47 -04:00
커밋한 사람 Yong Zhao
부모 8fc23418ea
커밋 768bd50cd2
2개의 변경된 파일과 9개의 추가작업 그리고 9개의 파일을 삭제
패치 파일 다운로드 Diff 파일 다운로드 모든 파일 펼침 모든 파일 접기
projects/rocr-runtime/tests/kfdtest/scripts/kfdtest.exclude
+1 -1
파일 보기
@@ -5,7 +5,7 @@
# stresses memory management in unusual ways by having lots of memory allocated but
# not mapped, which is not relevant to compute applications over ROCr.
#
# CU Masking Linear/Split are not working correctly due to how the HW distributes work over CUs.
# CU Masking Linear are not working correctly due to how the HW distributes work over CUs.
# They are available for testing but are not currently expected to pass on CI/VI/AI.
#
# The CheckZeroInitializationVram test is no longer expected to pass as KFD no longer
projects/rocr-runtime/tests/kfdtest/src/KFDQMTest.cpp
+8 -8
파일 보기
@@ -655,10 +655,10 @@ TEST_F(KFDQMTest, BasicCuMaskingLinear) {
const HsaNodeProperties *pNodeProperties = m_NodeInfo.GetNodeProperties(defaultGPUNode);
uint32_t ActiveCU = (pNodeProperties->NumFComputeCores / pNodeProperties->NumSIMDPerCU);
uint32_t numSEs = pNodeProperties->NumShaderBanks;
LOG() << std::hex << "# Compute cores: 0x" << pNodeProperties->NumFComputeCores << std::endl;
LOG() << std::hex << "# SIMDs per CPU: 0x" << pNodeProperties->NumSIMDPerCU << std::endl;
LOG() << std::hex << "# Shader engines: 0x" << numSEs << std::endl;
LOG() << std::hex << "# Active CUs: 0x" << ActiveCU << std::endl;
LOG() << std::dec << "# Compute cores: " << pNodeProperties->NumFComputeCores << std::endl;
LOG() << std::dec << "# SIMDs per CU: " << pNodeProperties->NumSIMDPerCU << std::endl;
LOG() << std::dec << "# Shader engines: " << numSEs << std::endl;
LOG() << std::dec << "# Active CUs: " << ActiveCU << std::endl;
HSAint64 TimewithCU1, TimewithCU;
uint32_t maskNumDwords = (ActiveCU + 31) / 32; /* Round up to the nearest multiple of 32 */
uint32_t maskNumBits = maskNumDwords * 32;
@@ -718,10 +718,10 @@ TEST_F(KFDQMTest, BasicCuMaskingEven) {
return;
}
LOG() << std::hex << "# Compute cores: 0x" << pNodeProperties->NumFComputeCores << std::endl;
LOG() << std::hex << "# SIMDs per CPU: 0x" << pNodeProperties->NumSIMDPerCU << std::endl;
LOG() << std::hex << "# Shader engines: 0x" << numShaderEngines << std::endl;
LOG() << std::hex << "# Active CUs: 0x" << ActiveCU << std::endl;
LOG() << std::dec << "# Compute cores: " << pNodeProperties->NumFComputeCores << std::endl;
LOG() << std::dec << "# SIMDs per CU: " << pNodeProperties->NumSIMDPerCU << std::endl;
LOG() << std::dec << "# Shader engines: " << numShaderEngines << std::endl;
LOG() << std::dec << "# Active CUs: " << ActiveCU << std::endl;
HSAint64 TimewithCU1, TimewithCU;
uint32_t maskNumDwords = (ActiveCU + 31) / 32; /* Round up to the nearest multiple of 32 */
uint32_t maskNumBits = maskNumDwords * 32;