JDK-8316355 : G1: Improve concurrent undo cycle to be more efficient
  • Type: Enhancement
  • Component: hotspot
  • Sub-Component: gc
  • Affected Version: 22
  • Priority: P4
  • Status: Open
  • Resolution: Unresolved
  • Submitted: 2023-09-15
  • Updated: 2023-09-18
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.
Other
tbdUnresolved
Related Reports
Relates :  
JDK-8240556 - Abort concurrent mark after effective eager reclamation of humongous objects
Description
The concurrent undo cycle may still take a significant amount of time due to clearing of the next bitmap.

Example:

[2023-08-25T08:51:23.515-0700][160596.213s][16441][info ][gc         ] GC(74) Concurrent Undo Cycle
[2023-08-25T08:51:23.518-0700][160596.216s][16441][info ][gc,marking     ] GC(74) Concurrent Cleanup for Next Mark
[2023-08-25T08:51:28.829-0700][160601.526s][16441][info ][gc,marking     ] GC(74) Concurrent Cleanup for Next Mark 5310.224ms
[2023-08-25T08:51:28.829-0700][160601.527s][16441][info ][gc         ] GC(74) Concurrent Undo Cycle 5313.553ms

Think of ways to shorten (or avoid?) (the expensive parts of) the concurrent undo cycle.

(random)options are:
* better track where the marks from VM roots were set and only undo them
* do not mark on the bitmap immediately, but store these somewhere (and do the marks concurrently).