Results from Release-1.1 on Aurora with LLVM-16 built with gcc-12 and Intel UMD 736.25

[colleeneb]

On Aurora with LLVM-16 built with gcc-12 and Intel UMD 736.25, the results I see from Release-1.1 were:

Level0 BE:

Immediate Command Lists

Test Run Command:

chip_be=level0 CHIP_L0_IMM_CMD_LISTS=ON ctest --output-on-failure --timeout 30 --repeat until-fail:1 -j 1 -E "`cat ./test_lists/dgpu_level0_failed_imm_tests.txt`|[Tt]ex"  -O checkpy_dgpu_imm_level0.txt

Result:

99% tests passed, 4 tests failed out of 882

Total Test time (real) = 1138.14 sec
...
The following tests FAILED:
        803 - Unit_hipMultiThreadStreams2 (Subprocess aborted)
        825 - firstTouch (Failed)
        826 - syncthreadsExitedThreads (Timeout)
        875 - TestIndirectMappedHostAlloc (Failed)

Regular Command Lists

Test Run Command:

chip_be=level0 CHIP_L0_IMM_CMD_LISTS=OFF ctest --output-on-failure --timeout 30 --repeat until-fail:1 -j 1 -E "`cat ./test_lists/dgpu_level0_failed_reg_tests.txt`|[Tt]ex"  -O checkpy_dgpu_reg_level0.txt

Result:

99% tests passed, 6 tests failed out of 877

Total Test time (real) = 1177.01 sec
...
The following tests FAILED:
        798 - Unit_hipMultiThreadStreams2 (Subprocess aborted)
        819 - TestHiprtcOptions (Timeout)
        820 - firstTouch (Failed)
        821 - syncthreadsExitedThreads (Timeout)
        847 - TestWholeProgramCompilation (Timeout)
        870 - TestIndirectMappedHostAlloc (Failed)

OpenCL BE:

Test Run Command:

hip_be=opencl ctest --output-on-failure --timeout 30 --repeat until-fail:1 -j 1 -E "`cat ./test_lists/dgpu_opencl_failed_tests.txt`|[Tt]ex"  -O checkpy_dgpu_opencl.txt

Result:

99% tests passed, 5 tests failed out of 871

Total Test time (real) = 1278.99 sec
...
The following tests FAILED:
        813 - TestHiprtcOptions (Timeout)
        814 - firstTouch (Failed)
        815 - syncthreadsExitedThreads (Timeout)
        841 - TestWholeProgramCompilation (Timeout)
        864 - TestIndirectMappedHostAlloc (Failed)