Takafumi Arakaki

Ah, I was actually using julia nightly I downloaded a while ago (1.5.0-DEV.464). Now I switched to 1.5.0-DEV.526, I get the hang! I get it with `JULIA_NUM_THREADS=13` but not with...

You are right. I tried to bisect the bug and it hangs with 1.5.0-DEV.465 (and so presumably with 1.5.0-DEV.464) too. While trying bisection I noticed that it can took more...

Great! This is good news! Thanks for figuring out the version that it works. I did think the bug might be in the scheduler but if it weren't then I'd...

Hmm... That's too bad that we can't rely on that `top` is really `top`. > I modified the script to use `ps` instead of `top` This may be OS-dependent, but...

> I decided to try the following commit a few times: > [JuliaLang/julia@65b8e7e](https://github.com/JuliaLang/julia/commit/65b8e7ed72541d1dd5bc1f4dcabc5245e1dba718) Yes, my bisection points to https://github.com/JuliaLang/julia/pull/32599 (which includes https://github.com/JuliaLang/julia/commit/65b8e7ed72541d1dd5bc1f4dcabc5245e1dba718), too. After this PR (https://github.com/JuliaLang/julia/commit/dc46ddd857d9fcf178ee5074f635dc29c5b1ade6) I can get...

Which Julia revision you used to get the segmentation fault?

The last run with https://github.com/JuliaLang/julia/commit/f5dbc471efc3334a8c532f56fb40b06427fa52c7 (_before_ JuliaLang/julia#32599) caused another segmentation fault ``` seed[] += 1 = 247 [ Info: %CPU = 1167.0 [ Info: 45.550472802933335 minutes without a hang.... signal...

Looking at the code in your stacktrace https://github.com/tkf/ThreadsX.jl/issues/104#issuecomment-606736472, I'm puzzled why it'd cause a segmentation fault. Is it possible that the stacktrace is broken? `-` at [./int.jl:52](https://github.com/JuliaLang/julia/blob/65b8e7ed72541d1dd5bc1f4dcabc5245e1dba718/base/int.jl#L52) `unsafe_length` at [./range.jl:517](https://github.com/JuliaLang/julia/blob/65b8e7ed72541d1dd5bc1f4dcabc5245e1dba718/base/range.jl#L517)...

I'm running the MWE with `--check-bounds=yes` and it's `seed[] += 1 = 423` now. No hang or segfault so far. Edit: I gave up after `seed[] += 1 = 1077`

So far, I think we discovered that: 1. ThreadsX.jl and/or `julia` _before_ JuliaLang/julia#32599 have a bug that is manifested as a segfault when quicksort is run for a very long...