-
Bug
-
Resolution: Fixed
-
P3
-
13
-
b17
See discussion in:
http://mail.openjdk.java.net/pipermail/zgc-dev/2019-March/000605.html
This is output from runs from that thread:
> service.gc.0:[24304.545s][debug][membername,table] ResolvedMethod entries counted 94768 removed 92352
> service.gc.0:[36944.749s][debug][membername,table] ResolvedMethod entries counted 64041 removed 61616
> service.gc.0:[36944.749s][debug][membername,table] ResolvedMethod entries counted 64041 removed 61616
> service.gc.3:[42177.904s][debug][membername,table] ResolvedMethod entries counted 19346 removed 16930
> service.gc.1:[47231.731s][debug][membername,table] ResolvedMethod entries counted 20110 removed 17694
> service.gc.1:[47231.731s][debug][membername,table] ResolvedMethod entries counted 20110 removed 17694
> service.gc.3:[52267.616s][debug][membername,table] ResolvedMethod entries counted 27426 removed 25010
> service.gc.1:[57078.463s][debug][membername,table] ResolvedMethod entries counted 30835 removed 28419
> service.gc.3:[61907.835s][debug][membername,table] ResolvedMethod entries counted 32403 removed 29987
> service.gc.2:[66916.223s][debug][membername,table] ResolvedMethod entries counted 34545 removed 32129
> service.gc.2:[71786.212s][debug][membername,table] ResolvedMethod entries counted 39112 removed 36696
> service.gc.2:[76619.460s][debug][membername,table] ResolvedMethod entries counted 34305 removed 31889
The ResolvedMethodTable size is fixed and too small:
class ResolvedMethodTable : public Hashtable<ClassLoaderWeakHandle, mtClass> {
enum Constants {
_table_size = 1007
};
This will cause performance degradation for lookups.
Time to update this to use the ConcurrentHashTable?
http://mail.openjdk.java.net/pipermail/zgc-dev/2019-March/000605.html
This is output from runs from that thread:
> service.gc.0:[24304.545s][debug][membername,table] ResolvedMethod entries counted 94768 removed 92352
> service.gc.0:[36944.749s][debug][membername,table] ResolvedMethod entries counted 64041 removed 61616
> service.gc.0:[36944.749s][debug][membername,table] ResolvedMethod entries counted 64041 removed 61616
> service.gc.3:[42177.904s][debug][membername,table] ResolvedMethod entries counted 19346 removed 16930
> service.gc.1:[47231.731s][debug][membername,table] ResolvedMethod entries counted 20110 removed 17694
> service.gc.1:[47231.731s][debug][membername,table] ResolvedMethod entries counted 20110 removed 17694
> service.gc.3:[52267.616s][debug][membername,table] ResolvedMethod entries counted 27426 removed 25010
> service.gc.1:[57078.463s][debug][membername,table] ResolvedMethod entries counted 30835 removed 28419
> service.gc.3:[61907.835s][debug][membername,table] ResolvedMethod entries counted 32403 removed 29987
> service.gc.2:[66916.223s][debug][membername,table] ResolvedMethod entries counted 34545 removed 32129
> service.gc.2:[71786.212s][debug][membername,table] ResolvedMethod entries counted 39112 removed 36696
> service.gc.2:[76619.460s][debug][membername,table] ResolvedMethod entries counted 34305 removed 31889
The ResolvedMethodTable size is fixed and too small:
class ResolvedMethodTable : public Hashtable<ClassLoaderWeakHandle, mtClass> {
enum Constants {
_table_size = 1007
};
This will cause performance degradation for lookups.
Time to update this to use the ConcurrentHashTable?
- relates to
-
JDK-8221872 Remove uses of ClassLoaderWeakHandle typedef in protection domain table
- Resolved
-
JDK-8235485 mlvm deoptimize stress test is slower with intermittent timeouts after the 8221393 fix
- Closed
-
JDK-8185525 Add JFR event for DictionarySizes
- Resolved
-
JDK-8222333 fastdebug build broken after JDK-8221393 (phase_mapping[] doesn't match enum Phase in WeakProcessorPhases)
- Resolved
-
JDK-8233453 MLVM deoptimize stress test timed out
- Resolved
-
JDK-8222550 runtime/MemberName/MemberNameLeak.java times out
- Resolved
(1 relates to)