RFR: 8358179: Performance regression in Math.cbrt
Mohamed Issa
missa at openjdk.org
Wed Jun 25 18:41:41 UTC 2025
The changes described below are meant to resolve the performance regression introduced by the **x86_64 cbrt** double precision floating point scalar intrinsic in #24470.
1. Check for +0, -0, +INF, -INF, and NaN before any other input values.
2. If these special values are found, return immediately with minimal modifications to the result register.
The commands to run all relevant micro-benchmarks are posted below.
`make test TEST="micro:CbrtPerf.CbrtPerfRanges"`
`make test TEST="micro:CbrtPerf.CbrtPerfSpecialValues"`
The results of all tests posted below were captured with an [Intel® Xeon 8488C](https://www.intel.com/content/www/us/en/products/sku/231730/intel-xeon-platinum-8480c-processor-105m-cache-2-00-ghz/specifications.html) using [OpenJDK v26-b1](https://github.com/openjdk/jdk/releases/tag/jdk-26%2B1) as the baseline version. The term _baseline1_ refers to runs with the intrinsic enabled and _baseline2_ refers to runs with the intrinsic disabled.
Each result is the mean of 8 individual runs, and the input ranges used match those from the original Java implementation. Overall, the changes provide a significant uplift over _baseline1_ except for a mild regression in the (**2^(-1022) <= |x| < INF**) input range, which is expected due to the extra checks. When comparing against _baseline2_, the modified intrinsic significantly still outperforms for the inputs (**-INF < x < INF**) that require heavy compute. However, the special value inputs that trigger fast path returns still perform better with _baseline2_.
| Input range(s) | Baseline1 (ops/ms) | Change (ops/ms) | Change vs baseline1 (%) |
| :-------------------------------------: | :-------------------: | :------------------: | :--------------------------: |
| [-2^(-1022), 2^(-1022)] | 18470 | 20847 | +12.87 |
| (-INF, -2^(-1022)], [2^(-1022), INF) | 210538 | 198925 | -5.52 |
| [0] | 344990 | 627561 | +81.91 |
| [-0] | 291983 | 629941 | +115.75 |
| [INF] | 382685 | 542211 | +41.68 |
| [-INF] | 386174 | 542291 | +40.43 |
| [NaN] | 421700 | 615157 | +45.88 |
| Input range(s) | Baseline2 (ops/ms) | Change (ops/ms) | Change vs baseline2 (%) |
| :-------------------------------------: | :-------------------: | :------------------: | :--------------------------: |
| [-2^(-1022), 2^(-1022)] | 7072 | 20847 | +194.78 |
| (-INF, -2^(-1022)], [2^(-1022), INF) | 147884 | 198925 | +34.51 |
| [0] | 1890520 | 627561 | -66.80 |
| [-0] | 1890404 | 629941 | -66.68 |
| [INF] | 1247633 | 542211 | -56.54 |
| [-INF] | 1242287 | 542291 | -56.35 |
| [NaN] | 1253700 | 615157 | -50.93 |
Finally, the `jtreg:test/jdk/java/lang/Math/CubeRootTests.java` test passed with the changes.
-------------
Commit messages:
- Make absolute mask memory constant 16 byte aligned for compatibility with andpd instruction
- Check for special values first in x86_64 cbrt intrinsic
Changes: https://git.openjdk.org/jdk/pull/25962/files
Webrev: https://webrevs.openjdk.org/?repo=jdk&pr=25962&range=00
Issue: https://bugs.openjdk.org/browse/JDK-8358179
Stats: 49 lines in 1 file changed: 10 ins; 36 del; 3 mod
Patch: https://git.openjdk.org/jdk/pull/25962.diff
Fetch: git fetch https://git.openjdk.org/jdk.git pull/25962/head:pull/25962
PR: https://git.openjdk.org/jdk/pull/25962
More information about the hotspot-compiler-dev
mailing list