On Fri, 8 Aug 2025 22:11:32 GMT, Francesco Andreuzzi <d...@openjdk.org> wrote:
>> In this PR I propose to refresh the included headers in hotspot >> `precompiled.hpp`. The current set of precompiled headers was refreshed in >> 2018, 7 years ago. I repeated the same operations and measurements after >> refreshing the set of precompiled headers according to the current usage >> frequency. >> >> These are the results I observed. Depending on the platform, the improvement >> is between 10 and 20% in terms of total work (user+sys). The results are in >> seconds. >> >> >> linux-x64 GCC >> master real 81.39 user 3352.15 sys 287.49 >> JDK-8365053 real 81.94 user 3030.24 sys 295.82 >> >> linux-x64 Clang >> master real 43.44 user 2082.93 sys 130.70 >> JDK-8365053 real 38.44 user 1723.80 sys 117.68 >> >> linux-aarch64 GCC >> master real 1188.08 user 2015.22 sys 175.53 >> JDK-8365053 real 1019.85 user 1667.45 sys 171.86 >> >> linux-aarch64 clang >> master real 981.77 user 1645.05 sys 118.60 >> JDK-8365053 real 791.96 user 1262.92 sys 101.50 > > Francesco Andreuzzi has updated the pull request incrementally with two > additional commits since the last revision: > > - magic number: 400 > - inline I'm starting to think the high variability is a quirk of my environment. I ran the builds in a virtualized machine and I'm probably experiencing erratic behavior under high load. I re-ran each combination fewer times (3, should still be enough to get an idea), and I'm observing more reasonable results: | Feature | Baseline `precompiled.hpp` | Updated `precompiled.hpp` | |----------------------------|-----------------|-----------------| | epsilongc | 458.88 +/- 0.35 | 459.69 +/- 0.34 | | epsilongc cds | 501.63 +/- 0.35 | 501.58 +/- 0.57 | | epsilongc compiler1 | 505.66 +/- 0.61 | 505.06 +/- 0.51 | | epsilongc compiler2 | 725.91 +/- 0.41 | 692.25 +/- 0.40 | | g1gc | 630.73 +/- 0.02 | 624.24 +/- 0.35 | | epsilongc services jfr | 625.54 +/- 1.45 | 592.37 +/- 1.00 | | epsilongc jni-check | 461.19 +/- 0.14 | 461.79 +/- 0.89 | | epsilongc compiler1 jvmci | 541.31 +/- 0.42 | 540.31 +/- 0.70 | | epsilongc compiler2 jvmci | 758.13 +/- 0.71 | 724.80 +/- 0.21 | | epsilongc services jvmti | 499.64 +/- 0.18 | 499.69 +/- 0.52 | | epsilongc management | 450.82 +/- 0.62 | 451.38 +/- 0.59 | | epsilongc opt-size | 446.44 +/- 0.56 | 446.16 +/- 0.12 | | parallelgc | 467.23 +/- 0.19 | 466.89 +/- 0.41 | | serialgc | 454.17 +/- 0.40 | 454.22 +/- 0.12 | | epsilongc services | 453.82 +/- 0.21 | 453.59 +/- 0.51 | | shenandoahgc | 658.17 +/- 0.50 | 647.44 +/- 0.40 | [This](https://github.com/fandreuz/jdk-stuff/blob/master/script) is the script I used. ------------- PR Comment: https://git.openjdk.org/jdk/pull/26681#issuecomment-3177333952
