openjdk test cases is different between Mode150 and Mode650?

In aqa-tests all openjdk tests are running with both mode150 and mode650. That is the number of testcase should be same for the two mode? However it feels like not. Maybe I misunderstand something or investigation should be taken to figure it out?

For example: compiler/rtm/locking/TestRTMAbortRatio.java is running with Mode650 and is not running with Mode150 on AIX.
jvm_compile_1 Test results: passed: 839; failed: 1
20:33:30 Report written to /home/jenkins/workspace/Test_openjdk11_hs_extended.openjdk_ppc64_aix_testList_0/jvmtest/openjdk/report/html/report.html
20:33:30 Results written to /home/jenkins/workspace/Test_openjdk11_hs_extended.openjdk_ppc64_aix_testList_0/aqa-tests/TKG/output_16504090503753/jvm_compiler_1/work
20:33:30 Error: Some tests failed or other problems occurred.
20:33:30
20:33:30 jvm_compiler_1_FAILED
https://ci.adoptopenjdk.net/job/Test_openjdk11_hs_extended.openjdk_ppc64_aix_testList_0/37/consoleFull

jvm_compile_0 Test results: passed: 821
https://ci.adoptopenjdk.net/job/Test_openjdk11_hs_extended.openjdk_ppc64_aix_testList_1/37/consoleFull

Looking at the test compiler/rtm/locking/TestRTMAbortRatio.java itself doesn't look like compressedRef is filtered.

Start with compiler/rtm/locking/TestRTMAbortRatio.java and compare the Mode150 and Mode650 test lists and AIX results referenced in the issue. Check the jvm_compile_0 and jvm_compiler_1 console reports to determine why the test is included in one mode but not the other. Done means documenting whether the difference is expected or identifying the filtering/configuration issue.

Written by the indexing model from the issue text.