-
Notifications
You must be signed in to change notification settings - Fork 23
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
Segmentation fault in multi-threaded code #304
Comments
Ah, yeah, I see why/how the data race is happening... there is a lock but on different mutexes. I can get it patched easily and I’ll generate a new release |
Good stuff! Let me know how to get the patched version. |
Jonathan, this is the Neko code, I was discussing the same with Niclas today. Please inform us when there is a change. |
It is being held up by whatever happened to the build system on RedHat in HIP 5.5 and 5.6, seen in #300. Looks like some amdgpu libraries got moved and aren’t being found. I’m trying to get to solving it soon |
Ok, I finally found the time to sort out #300. I’ll get that merged shortly and then addressing this will be quick and easy. There should be a release available tomorrow |
It’s going to be a little while longer until I figure out how to solve the packaging and code coverage routinely running out of disk space. |
Hi @jrmadsen, I also encounters almost the same bug when profiling my multi-threaded program(this one is unordered map, mine is ordered map's RBTree). Is there any estimation about when this bug fix will be brought to release? Thanks! |
I’m tied up with the rocprofiler v2 rewrite right now. @benrichard-amd is looking into fixing #300 so that the testing can pass. Right now it’s just the code coverage job that is failing. If he cannot find a fix soon, I’ll just disable that job so that I can merge it, fix the bug, and generate a release |
Thanks @jrmadsen! waiting for your good news on the fix. |
Just generated the new release. Installers should be available shortly, however I haven't updated the installer generation to provide installers for ROCm 5.7 yet, just FYI. |
I have an application that uses up to 7-threads and I randomly get segmentation faults from omnitrace version 1.10.2 coming from:
This app was executed as:
I believe there might be a race going into this unordered map. This comes from an app that is not trivial to build. Let me know if you'd like to provide more information about the SEGFault or the app itself.
The text was updated successfully, but these errors were encountered: