JDK-8027476 : Improve performance of Stringtable unlink
  • Type: Enhancement
  • Component: hotspot
  • Sub-Component: gc
  • Affected Version: hs25,8
  • Priority: P4
  • Status: Resolved
  • Resolution: Fixed
  • Submitted: 2013-10-29
  • Updated: 2014-07-29
  • Resolved: 2014-01-20
The Version table provides details related to the release that this issue/RFE will be addressed.

Unresolved : Release in which this issue/RFE will be addressed.
Resolved: Release in which this issue/RFE has been resolved.
Fixed : Release in which this issue/RFE has been fixed. The release containing this fix may be available for download as an Early Access Release or a General Availability Release.

To download the current JDK release, click here.
JDK 7 JDK 8 JDK 9
7u60Fixed 8u20Fixed 9 b03Fixed
Related Reports
Relates :  
JDK-8033443 - Test8000311 fails after latest changes to parallelize string and symbol table unlink
Relates :  
JDK-8033106 - Wrong predicate for checking whether the correct amount of symbol table entries have been processed in G1
Relates :  
JDK-8031472 - Fine tune string/symbol table parallelization
Relates :  
JDK-6996604 - Use more efficient (not linear) data structure for searching buckets in SymbolTable
Relates :  
JDK-8031565 - Use serial/waiting time to perform independent tasks during young gc in G1
Relates :  
JDK-8027454 - Do not traverse string table during G1 remark when treating them as strong roots during initial mark
Relates :  
JDK-8031473 - Parallelize symbol table and string table unlink in non-G1 parallel collectors
Description
At the end of a GC that treated strings as a weak root for traversal, the string table is scanned for unreferenced entries, which are then deleted.

This can be time-consuming if the string table is large; this RFE is about improving the performance.

One way to do this is to parallelize the effort; JDK-8004888 implemented parallel traversal by claiming hash table buckets on a per-thread basis. A similar method could be applied in this case, and provided for all uses in a parallel collector (at least CMS, the parallel gcs, G1 full gc and during the G1 remark pause).