RFR: 8295267: Improve G1 scan to merge ratio calculation [v2]
Thomas Schatzl
tschatzl at openjdk.org
Thu Oct 13 13:56:32 UTC 2022
> Hi all,
>
> can I have reviews for this change that fixes the merge-to-scan ratio calculation ("which fraction of the remembered sets creates unique cards") for predictions to make it more accurate and understandable?
>
> Previously the denominator has been the total number of scanned cards; this seems to be a good approximation (trying to make a case for it, it's wrong) if the number of log buffers is low at GC start, but tends to get broken otherwise (i.e. with [JDK-8137022](https://bugs.openjdk.org/browse/JDK-8137022)).
>
> This change fixes this to use the number of cards that were merged in.
>
> This change depends on PR#10656. Please review that one first if not done otherwise.
>
> Testing: tier1, checking that no tests use the "Dirty Cards" string now changed.
>
> Thanks,
> Thomas
Thomas Schatzl has updated the pull request incrementally with one additional commit since the last revision:
ayang review, merge_to_scan_ratio -> scan_to_merge_ratio
-------------
Changes:
- all: https://git.openjdk.org/jdk/pull/10694/files
- new: https://git.openjdk.org/jdk/pull/10694/files/3ff69963..b648df4a
Webrevs:
- full: https://webrevs.openjdk.org/?repo=jdk&pr=10694&range=01
- incr: https://webrevs.openjdk.org/?repo=jdk&pr=10694&range=00-01
Stats: 16 lines in 3 files changed: 0 ins; 0 del; 16 mod
Patch: https://git.openjdk.org/jdk/pull/10694.diff
Fetch: git fetch https://git.openjdk.org/jdk pull/10694/head:pull/10694
PR: https://git.openjdk.org/jdk/pull/10694
More information about the hotspot-gc-dev
mailing list