[upstream] Segmentation fault at origin - Kmeans multi-threaded #101
Labels
Comments
|
I was able to reproduce this Jack on my setup. I am taking a look at it. And will update you asap. |
|
Jack I was able to triage the problem. I am going to have an updated version of the upstream kernel later today for you to test. It may roll back some changes until I figure out the full fix. |
jnarf
added a commit
that referenced
this issue
Apr 1, 2020
process_remote_futex was missing pcn_kmsg_done before error exit.
|
Fixed issue. ran with command: ./kmeans -d 1000 -c 10 -p 2000 -s 1000 -n 2 -t 8 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
I ran "./kmeans -d 3 -c 20 -p 50000 -n 2 -t 4" twice. The first execution completed w/o any problem (awesome!). However, the 2nd execution gave me a segmentation fault on the origin node as shown below:
The execution log, dmesg logs on node0/1, and kmeans binary are attached.
kmeans_logs.zip
My setup is 2 VM using old x86.img downloaded from this repository's wiki page. Each VM has 8 vCPU with 40Gb RAM.
Please take a look. Thank you.
The text was updated successfully, but these errors were encountered: