moooonk opened a new pull request, #2092: URL: https://github.com/apache/fury/pull/2092
<!-- **Thanks for contributing to Fury.** **If this is your first time opening a PR on fury, you can refer to [CONTRIBUTING.md](https://github.com/apache/fury/blob/main/CONTRIBUTING.md).** Contribution Checklist - The **Apache Fury (incubating)** community has restrictions on the naming of pr titles. You can also find instructions in [CONTRIBUTING.md](https://github.com/apache/fury/blob/main/CONTRIBUTING.md). - Fury has a strong focus on performance. If the PR you submit will have an impact on performance, please benchmark it first and provide the benchmark result here. --> ## What does this PR do? <!-- Describe the purpose of this PR. --> After testing, it was observed that ThreadPoolFury experiences prolonged blocking during cold starts under high-concurrency scenarios. Analysis revealed that improper usage of locks in ClassLoaderFuryPooled was the root cause. This PR refactors the implementation of ClassLoaderFuryPooled by significantly reducing the granularity of locks, thereby drastically minimizing blocking time during cold starts. ## Related issues <!-- Is there any related issue? Please attach here. - #xxxx0 - #xxxx1 - #xxxx2 --> [](url)https://github.com/apache/fury/issues/2087 ## Does this PR introduce any user-facing change? <!-- If any user-facing interface changes, please [open an issue](https://github.com/apache/fury/issues/new/choose) describing the need to do so and update the document if necessary. --> - [ ] Does this PR introduce any public API change? - [ ] Does this PR introduce any binary protocol compatibility change? ## Benchmark <!-- When the PR has an impact on performance (if you don't know whether the PR will have an impact on performance, you can submit the PR first, and if it will have impact on performance, the code reviewer will explain it), be sure to attach a benchmark data here. --> CPU:9950X Class:org.apache.fury.benchmark.ThreadPoolFurySuite.java old: Percentiles, ms/op: p(0.0000) = 0.001 ms/op p(50.0000) = 0.001 ms/op p(90.0000) = 1587.388 ms/op p(95.0000) = 1587.388 ms/op p(99.0000) = 1587.388 ms/op p(99.9000) = 1587.388 ms/op p(99.9900) = 1587.388 ms/op p(99.9990) = 1587.388 ms/op p(99.9999) = 1587.388 ms/op p(100.0000) = 1587.388 ms/op new: Percentiles, ms/op: p(0.0000) = 0.001 ms/op p(50.0000) = 0.001 ms/op p(90.0000) = 62.746 ms/op p(95.0000) = 62.746 ms/op p(99.0000) = 62.746 ms/op p(99.9000) = 62.746 ms/op p(99.9900) = 62.746 ms/op p(99.9990) = 62.746 ms/op p(99.9999) = 62.746 ms/op p(100.0000) = 62.746 ms/op -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
