[
https://issues.apache.org/jira/browse/ARROW-12533?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17332892#comment-17332892
]
Yibo Cai commented on ARROW-12533:
----------------------------------
As clang libc++ is not adopted widely, I think it makes sense to implement a
random real distribution code in Arrow test.
I looked at <random> implementation in gcc and clang. I plan to port clang
libc++ code to Arrow. Clang adopts Apache 2.0 license, and the code looks more
optimized than gcc.
We need three function/classes.
- {{std::generate_canonical}} function, produces random reals in {{[0, 1)}}.
This function does the dirty job.
- {{std::uniform_real_distribution}} class, calls {{std::generate_canonical}}
to produce reals in any value range.
- {{std::bernoulli_distribution}} class, a shallow wrap of
{{std::uniform_real_distribution}}.
> [C++] Random real generator is slow on Arm64 Linux when built with clang
> ------------------------------------------------------------------------
>
> Key: ARROW-12533
> URL: https://issues.apache.org/jira/browse/ARROW-12533
> Project: Apache Arrow
> Issue Type: Improvement
> Components: C++
> Reporter: Yibo Cai
> Priority: Minor
>
> Many benchmarks run very slow on Arm64 Linux when built with clang.
> Most time is spent in preparing test data, not the test itself.
> Per my investigation, it boils down to poor performance of
> `std::uniform_real_distribution`, which uses software emulated `long double`
> arithmetic on Arm64 [1].
> Apple M1 doesn't have this issue. Clang aarch64 sets `long double` size to 64
> bits on MacOS, but 128 on Linux [2].
> Gcc aarch64 doesn't have this issue. It doesn't use `long double` to generate
> random reals [1]. Guess clang uses algorithms with better randomness.
> clang `-ffast-math` option removes the `long double` arithmetic (and adds
> other simplifications to floating point arithmetic), it improves speed 100x
> on Arm64 in generating random reals.
> It may deserve some effort to study if `long double` is really necessary, and
> if `-ffast-math` is acceptable for generating test bits.
> [1] [https://godbolt.org/z/Y3Tc6MTME]
> [2] [https://en.wikipedia.org/wiki/Long_double]
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
