Charm++ parallel programming system

[Nitin Bhat <nitin AT hpccharm.com>]

Hi Jonathan,

Thanks for reaching out. 

As mentioned in question 5.9.12 in this FAQ (https://charm.readthedocs.io/en/latest/faq/manual.html), the error that you mentioned is most likely a segmentation fault in the program. 

I see that you used the distribution from the Github mirror of Charm++. Are you using the development version (the latest commit) or any specific release? 

Can you send the command that you used to build Charm++ on Golub? 

Which program run resulted in giving you that error? It’ll be helpful if you send the command that you used to run the program. 

Also, do you see the error when you run examples and tests that are present as a part of the distribution? 

Thanks, 

Nitin

On Mar 26, 2019, at 10:44 PM, Jonathan Hoelzel <hoelzel2 AT illinois.edu> wrote:

Hello,

I am having an issue when running my Charm++ program on the UIUC campus cluster.  I have pasted the error below:

Charmrun> error on request socket to node 0 'golub298'--
Socket closed before recv.

I am using the github mirror of Charm++.  I am encountering the same issue when building Charm++ with 'verbs' or with 'net'.  I am wondering if there is something specific I might be doing wrong to cause this error.

Thanks for your help,

Jonathan Hoelzel

Computer Science | College of Engineering

University of Illinois at Urbana-Champaign

hoelzel2 AT illinois.edu