-
Notifications
You must be signed in to change notification settings - Fork 4.7k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Assertion failed 'compCyclesAtEnd >= m_compCyclesAtEndOfInlining fails intermittently on freebsd-x64 Debug build #105472
Comments
Tagging subscribers to this area: @JulieLeeMSFT, @jakobbotsch |
Tried |
This uses PAL |
It seems we could switch all of this to just use |
Is |
Yup, PAL tests seem to have that API covered. If you can reproduce this issue with Debug build (basically |
I assume the |
That one is different. @AustinWise is fixing it for illumos-x64 #105207. We can take a look at it for freebsd-x64 separately. QueryThreadCycleTime intermittent failure is not reproduced with PAL tests. I tried it in my box and got the same results: https://dpaste.com/CGVJZEZSK ( I then removed everything from paltestlist.txt except for @jakobbotsch's idea to avoid reliance on thread times is something we should try. There are only two usages of this PAL API in product code, one in jit and the other one in superpmi. |
I think the ambition here was to avoid charging the jit for time when the thread is suspended. Changing it to a wallclock measure should be ok. |
@am11 can this wait post 9.0 or better to be included in 9.0 ? |
I think this can wait until after .net 9.0; since it's Debug-only and (what I gathered from @Thefrank) it does not block publishing nativeaot apps on FreeBSD with Release build of shared framework. |
Same results with Debug E5-2660 lacks AVX2 amd AVX512 |
It was actually failing the assertion on @Thefrank, please open an issue with details for the two failing tests. |
Dug back into the test:
For both the failed tests. |
It fails on random methods with Debug build, e.g.
then:
etc.
(side note: that warning is bogus because it's definitely a debug build)
cc @Thefrank
The text was updated successfully, but these errors were encountered: