Exception during GC on Circuitscape #41524
Comments
vchuravy
added
bug
|
Hello @ShabnamS93 to make it easier to reproduce the bug, could you post @ranjanan @ViralBShah is Circuitscape doing anything fishy with memory? |
vchuravy
added
the
needs more info
|
It is all straightforward Julia - but it may be using Mmapped memory for parallel runs. |
|
@ShabnamS93 Can you send us the ini file and the input files to reproduce? |
|
Hi @vchuravy and @ViralBShah
Thanks for your reply.
Sure.
Here are the files you requested.
As for my `Pkg.status()`, Julia's not working for me to run the code. I
even re-installed it it just does not work.
…On Fri, Jul 9, 2021 at 5:19 PM Viral B. Shah ***@***.***> wrote:
@ShabnamS93 <https://github.com/ShabnamS93> Can you send us the ini file
and the input files to reproduce?
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#41524 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AKVDZRULGYAYQQUIQXOUX43TW3V45ANCNFSM5ACBW5VA>
.
|
|
@ShabnamS93 you'll have to log in to github and put the files in the issue thread. If you attach them to the email, we cannot see them. Alternatively, you can put them up in a google drive link and reply via email. |
|
I have attached them to the email. But, now, here they are: |
|
Thanks, checking it out. |
|
Seems to work on my laptop. But I'll run this on a server for longer (seems like you saw this at 1700 pairs). |
|
I wonder if one needs to replicate this on windows. This doesn't appear to parallelize - I note |
|
This does not parallelize because a new graph is being generated for every pair. In this case however, it's probably worth parallelizing given each generated graph isn't too large. I'll run this for a while on a Windows machine too. |
|
Oh, I didn't think of that as an issue. Thanks a lot. I'll try it and let you know. |
|
I just ran this job over the weekend and it completed without issues on a Windows machine (Julia 1.6.0) and Ubuntu Linux (Julia 1.61). Also, @ShabnamS93 just to clarify, parallelization isn't implemented for the mode you're running Circuitscape in. I was just making a comment about whether it's worth implementing (it is). |
|
Hi everyone
Yeah, it's my laptop. I'm having it fixed and still haven't done the codes.
ASA I get it done, I'll let you know how it goes.
…On Mon, Jul 12, 2021, 20:10 Ranjan Anantharaman ***@***.***> wrote:
I just ran this job over the weekend and it completed without issues on a
Windows machine (Julia 1.6.0) and Ubuntu Linux (Julia 1.61).
Also, @ShabnamS93 <https://github.com/ShabnamS93> just to clarify,
parallelization isn't implemented for the mode you're running Circuitscape
in. I was just making a comment about whether it's worth implementing (it
is).
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#41524 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AKVDZRQCAJFUIW4KPNH5QA3TXMEFHANCNFSM5ACBW5VA>
.
|
|
I tried to test and got: |
|
Hi. ERROR: LoadError: On worker 31: |
|
This seems unrelated here - can you file a new issue and mention the OS, input files and settings to reproduce? |
Hi.
Here I'm trying to run the Circuitscape package, but I got the following bug. Can it be caused because of my heavy files? My system runs on an 8-GB Ram, has a Core i7 processor, and the files are heavy (almost 17000 nodes to compute).
After it was through 1/10 of the process, I faced this issue (and Julia is not working anymore):
The text was updated successfully, but these errors were encountered: