[jdk21u-dev] RFR: 8325553: Parallel: Use per-marker cache for marking stats during Full GC

[Rui Li]

A clean backport for https://bugs.openjdk.org/browse/JDK-8325553.

Mainly performance improvement: instead of using atomic operation to update shared data, use per marker cache.

Tests are running. On tip for over a year.

-------------

Commit messages:
 - Backport 2d3c9c5e673cf72fd4873fb5139e7d7a3d65a097

Changes: https://git.openjdk.org/jdk21u-dev/pull/2152/files
  Webrev: https://webrevs.openjdk.org/?repo=jdk21u-dev&pr=2152&range=00
  Issue: https://bugs.openjdk.org/browse/JDK-8325553
  Stats: 120 lines in 5 files changed: 118 ins; 1 del; 1 mod
  Patch: https://git.openjdk.org/jdk21u-dev/pull/2152.diff
  Fetch: git fetch https://git.openjdk.org/jdk21u-dev.git pull/2152/head:pull/2152

PR: https://git.openjdk.org/jdk21u-dev/pull/2152