MiniMix_aot_5m_0 failed on windows32 with jdk11, jdk17

[sophia-guo]

MiniMix_aot_5m_0 failed on windows32 with with jdk11, jdk17. Might happen to other version? Deep history shows it has been failed since Dec26 (maybe earlier than that, as earlier history is deleted)

12:00:07  STF 11:00:16.343 - Monitoring processes: LT 
12:00:07  LT  Error occurred during initialization of VM
12:00:07  LT  Could not reserve enough space for 2097152KB object heap
12:00:07  STF 11:00:16.380 - **FAILED** Process LT  ended with exit code (1) and not the expected exit code/s (0)
12:00:07  STF 11:00:16.380 - Monitoring Report Summary:
12:00:07  STF 11:00:16.380 -   o Process LT  ended with exit code (1) and not the expected exit code/s (0)
12:00:07  STF 11:00:16.381 - Killing processes: LT 
12:00:07  STF 11:00:16.381 -   o Process LT  pid 44304 is not running
12:00:07  **FAILED** at step 1 (Run mixed unit tests). Expected return value=0 Actual=1 at E:/workspace/Test_openjdk17_hs_extended.system_x86-32_windows/aqa-tests/\TKG\output_16423517415225\MiniMix_aot_5m_0/20220116-110014-MixedLoadTest/execute.pl line 98.
12:00:07  STF 11:00:16.397 - **FAILED** execute script failed. Expected return value=0 Actual=1

https://ci.adoptopenjdk.net/job/Test_openjdk17_hs_extended.system_x86-32_windows/134/consoleFull
https://ci.adoptopenjdk.net/job/Test_openjdk11_hs_extended.system_ppc64le_linux/587/

[smlambert]

like the 6 perf benchmarks that we exclude on win32 (and arm_linux, see #3201), it may be that we permanently exclude some tests on certain platforms that have limitations

[Mesbah-Alam]

Similar behaviour was seen on Openj9 and the test was excluded from Win32 on Openj9 and IBM: #3193.

Quoting Pete on the failure: "It indicates a problem with the test. The test is trying to use a bigger object heap than is available on the 32-bit systems".