On Wed, 26 Feb 2025 16:25:46 GMT, Vladimir Ivanov wrote:
> Yes, exceptions reported for runs with size=1024. The test support max
> size=512 and have no checks for passed params.
The change makes sense to me. Thanks for your fixing!
-
PR Comment: https://git.openjdk.org/jdk/pull/2
On Tue, 25 Feb 2025 19:06:05 GMT, Vladimir Ivanov wrote:
> Array initialization by parameter was added. Extra constant was used to align
> cycle step with used arrays.
Marked as reviewed by xgong (Committer).
-
PR Review: https://git.openjdk.org/jdk/pull/23783#pullrequestreview-26
On Tue, 25 Feb 2025 19:06:05 GMT, Vladimir Ivanov wrote:
> Array initialization by parameter was added. Extra constant was used to align
> cycle step with used arrays.
Yes, exceptions reported for runs with size=1024. The test support max size=512
and have no checks for passed params.
---
On Wed, 26 Feb 2025 06:59:27 GMT, Xiaohong Gong wrote:
> Hi @IvaVladimir , thank you for fixing this benchmark. Could you please tell
> more information about the IOOBE crash (e.g. `size`, `benchmark name`, `arch
> info`, .etc) ? I still cannot figure out why it can fail with IOOBE. Thanks
> s
On Tue, 25 Feb 2025 19:06:05 GMT, Vladimir Ivanov wrote:
> Array initialization by parameter was added. Extra constant was used to align
> cycle step with used arrays.
Hi @IvaVladimir , thank you for fixing this benchmark. Could you please tell
more information about the IOOBE crash (e.g. `siz
Array initialization by parameter was added. Extra constant was used to align
cycle step with used arrays.
-
Commit messages:
- 8350682 [JMH] vector.IndexInRangeBenchmark failed with
IndexOutOfBoundsException for size=1024
Changes: https://git.openjdk.org/jdk/pull/23783/files
We
