[
https://issues.apache.org/jira/browse/GROOVY-12023?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18085638#comment-18085638
]
ASF GitHub Bot commented on GROOVY-12023:
-----------------------------------------
github-actions[bot] commented on PR #2549:
URL: https://github.com/apache/groovy/pull/2549#issuecomment-4607888429
### JMH summary — indy (commit `c7c8805`)
Speedup vs trailing 90-day baseline on gh-pages. Higher = faster.
`1.00` = in line with history. Per-benchmark ratio, geomean within group.
Time-per-op units inverted so direction is consistent.
| Group | Speedup | n |
|--------|---------|---|
| bench | 1.125 × | 26 |
| core | 0.942 × | 75 |
| grails | 1.081 × | 48 |
<sub>Baseline: <code>dev/bench/jmh/<part>/indy/data.js</code> on
gh-pages, trailing 90 days. <a
href="https://apache.github.io/groovy/dev/bench/jmh/summary.html";>Daily
dashboard</a> · <a
href="https://apache.github.io/groovy/dev/bench/jmh/";>Per-suite raw
data</a></sub>
<!
> create a PIC for indy
> ---------------------
>
> Key: GROOVY-12023
> URL: https://issues.apache.org/jira/browse/GROOVY-12023
> Project: Groovy
> Issue Type: Bug
> Reporter: Jochen Theodorou
> Assignee: Jochen Theodorou
> Priority: Major
>
> What we currently have is essentially a two layer system in which we have the
> target handle and cache for everything else. This can be seen as a PIC of
> size 1 with a fallback cache for megamorphic cases. Here I propose to
> implement a PIC of at least size 3 consisting of method handles where the
> guard will use the next PIC element as fallback instead of going to the
> cache. The goal is to reach a more stable configuration on the callsite for
> JIT to optimize.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
