Page History

...

This is SPECjvm2008 modified to run with JDK24. Scores are ops/min, more is better. All benchmarks have been run on a AWS c5.9xlarge instance.

JVM flags: -Xms8g -Xms8g Xmx8g -XX:+AlwaysPreTouch

Notes: The CryptoRsa benchmark shows a legit regression of ~2%. I confirmed this by running 50 times (most benchmarks are run 3 times). The regression disappears when improving oopDesc::klass().

...

This is renaissance-jmh-0.15.0, excluding Als, ChiSquare, DecTree, GaussMix, LogRegression, MovieLens, NaiveBayes, PageRank and DbShootout, which are not compatible with JDK24 due to SecurityManager restriction. Scores are ms/op, less is better. All benchmarks have been run on a AWS c5.9xlarge instance.

JVM flags: -Xms8g -Xms8g Xmx8g -XX:+AlwaysPreTouch

SPECjbb2015

SPECjbb2015 in composite mode, run on a AWS c5.9xlarge instance, 10 times in a row, scores averaged over all runs. More is better.

JVM flags: -Xms20g -Xms20g Xmx20g -XX:+AlwaysPreTouch

GC Pause times

Young GC time, measured over 10 runs of SPECjbb2015 each, the total number of cycles is given, too.

Young GC time, baseline: avg: 71.4059 ms, #cycles: 22909
Young GC time, -UCOH:    avg: 73.2457 ms, #cycles: 23032
Young GC time, +UCOH:    avg: 69.8878 ms, #cycles: 20619

Concurrent marking time, measured over a run of SPECjvm compiler.compiler benchmark.

Concurrent marking time, baseline: avg: 1757.43 ms, #cycles: 69
Concurrent marking time, -UCOH:    avg: 1724.96 ms, #cycles: 78
Concurrent marking time, +UCOH:    avg: 1664.67 ms, #cycles: 62

SPECjbb2015 (run at Red Hat)

The following tests were run at Red Hat on a bare metal RHEL9 x64 i7-4770 8core machine, with the benchmark process isolated on 6 cores, 10 consecutive runs, all scores averaged. Perf statistics were done for the full benchmark process run, therefore include warmup and report generation.

JVM flags: -Xshare:off -Xlog:gc* -XX:+UseG1GC -Xms4g -Xms4gXmx4g

...