Column of Zeros Dumped in Ez Data , Possible bug with edges of the processor areas? #105

Open keithbennett opened 3 years ago

In GitLab by Jonathan Kennedy on 01 Dec 2020, 10:55 UTC (GitLab issue GL#2284)

Summary

Was simulating two Lasers traveling at angles of +- pi/9 at a plasma surface and seem to be getting whole column of zeros dumped in the Ez data file. I ran the same simulation for np - 32,64,128 and 100. I saw the same dump of zeros for np - 32,64,128 but none for np - 100 so it may be a bug to do with the edges of the processor areas?

Steps to reproduce

input.deck

Please also state the version of EPOCH : EPOCH2D version 4.17.12

Ran on np - 32,64,128 ,100

Fortran compiler and MPI library used along with their version numbers:

module load compilers/gcc/9.3.0 module load mpi/openmpi/4.0.4/gcc-9.3.0

What is the current bug behavior?

Column of Zeros are dumped in the Ez data file

What is the expected correct behaviour?

No column of zeros

Relevant logs and/or screenshots

[EPOCH_Noncolinear_Bug_Test.pdf](../blob/37c47bd3c9805e46ecfa4681221ec19dd3629922/uploads/a5bd817f95201c051d03a2c81f9a53a5/EPOCH_Noncolinear_Bug_Test.pdf)

In GitLab by Jonathan Kennedy on 01 Dec 2020, 11:02 UTC

changed the description

In GitLab by Thomas Goffrey (@TomGoffrey) on 01 Dec 2020, 12:41 UTC

Certainly looks suspicious, but there's not enough information here for us to investigate. Please edit your post, select the bug template, and provide the requested information. Thanks.

In GitLab by Jonathan Kennedy on 01 Dec 2020, 13:42 UTC

changed the description

In GitLab by Jonathan Kennedy on 01 Dec 2020, 13:43 UTC

changed the description

In GitLab by Jonathan Kennedy on 01 Dec 2020, 13:45 UTC

Hi Thomas, Sorry I'm a bit of a novice when it comes to Gitlab. Hopefully I've used the bug template correctly. Best Jonathan

In GitLab by Thomas Goffrey (@TomGoffrey) on 01 Dec 2020, 17:28 UTC

Thanks, that's great.

Quick question, did you download a .tar file of EPOCH, or did you access the code via git?

In GitLab by Jonathan Kennedy on 02 Dec 2020, 12:24 UTC

Hi Thomas, I got it from the repository via : git clone --recursive https://cfsa-pmw.warwick.ac.uk/EPOCH/epoch.git

In GitLab by Thomas Goffrey (@TomGoffrey) on 02 Dec 2020, 12:37 UTC

It might be worth trying the 4.18-devel branch, to see if the issue has been fixed in our next release. I noticed that you are using subsets and I know there are fixes to subsets in that branch.

Are you able to try that? Let me know if you need instructions relating to git.

In GitLab by Jonathan Kennedy on 02 Dec 2020, 12:41 UTC

Hi Thomas, Yeah I think I'd need some help with instructions relating to git

In GitLab by Thomas Goffrey (@TomGoffrey) on 02 Dec 2020, 12:49 UTC

The first thing to check is if you have any local modifications to the EPOCH source code. To do this run git status from the EPOCH directory. That should show something like this:

On branch master
Your branch is up-to-date with 'origin/master'.

Untracked files:

Followed by a list of untracked files (assuming you have any). If the output contains a section like:

Changes not staged for commit:
  (use "git add <file>..." to update what will be committed)
  (use "git checkout -- <file>..." to discard changes in working directory)

    modified:   epoch1d/src/housekeeping/balance.F90

Then we'd need to clean those up, but that should only happen if you have changed any of the source code. If that's the case let me know.

Assuming we get past that stage, you should run the command git checkout 4.18-devel. This checks out the 4.18 version. You should also then run the command git submodule update --recursive to update the I/O modules.

Once you've done this you should be able to run make clean followed by make COMPILER=<normal options etc> to compile the new version of EPOCH. Assuming that runs smoothly you should be then able to test the new version of EPOCH.

Let me know if any of that doesn't make sense.

In GitLab by Jonathan Kennedy on 02 Dec 2020, 13:56 UTC

Hi Thomas, I completed the steps, Should the version of EPOCH be now called 4.18 ? Cause I'm getting " Welcome to EPOCH2D version 4.17.12 (commit v4.17.12-89-g7a6565d6-clean)" I've probably done something wrong here.

In GitLab by Thomas Goffrey (@TomGoffrey) on 02 Dec 2020, 14:05 UTC

That's fine. It's just that as 4.18 hasn't been released yet there's no tag associated with it yet, hence 4.17.12 is printed. The second half of that 7a6565d6 tells me you're running the 4.18 version.

In GitLab by Jonathan Kennedy on 02 Dec 2020, 14:12 UTC

Ah brilliant! Thank you so much for all the help! I'll re-run the input.deck using this version

In GitLab by Jonathan Kennedy on 03 Dec 2020, 12:29 UTC

Hi Thomas, It seems the issue has been fixed! Thank you for all your help EPOCHBug_Test.pdf

In GitLab by Thomas Goffrey (@TomGoffrey) on 07 Dec 2020, 14:00 UTC

Glad to hear that's working for you.

Hopefully we'll get a new release out soon.

In GitLab by Petr Valenta on 21 Dec 2020, 13:05 UTC

mentioned in issue GL#2291

Warwick-Plasma / epoch