adding MPI dependency in CMakeLists.txt of mgbf and using Russ's gsi_… #833
Conversation
|
@TingLei-NOAA , successfully building |
|
@RussTreadon-NOAA Thanks for this clarificaiton! |
There was a problem hiding this comment.
Installed TingLei-daprediction:feature/mgbf4gnu at 0c1279c on Hera. Set COMPILER=gnu. Execute ush/build.sh. The gsi.x and enkf.x were successfully built.
Approve.
|
@TingLei-daprediction , I will merge this PR into |
There was a problem hiding this comment.
Changes look good. Thanks @TingLei-daprediction @RussTreadon-NOAA! Approve.
There was a problem hiding this comment.
Remove commented lines. Update working copy of TingLei-daprediction:feature/mgbf4gnu to 41f325f. Rerun ush/build.sh for COMPILER=gnu. gsi.x and enkf.x successfully built on Hera.
Approve.
|
@RussTreadon-NOAA and @DavidHuber-NOAA Thanks a lot for your help! |
Description
This PR is to fix the GSI GNU compiling broken on hera
Resolves #831
Added the correct specification of MPI dependency in cmake file for mgbf and added @RussTreadon-NOAA 's changes in gsi_hera.gnu.lua to make right modules loaded on hera.
Co-author : @RussTreadon-NOAA
<But now, I have no a "control" GSI without the above changes to run ctests on hera since the GSI would fail in building if there are no those changes.
@RussTreadon-NOAA, what are your recommendation for ctests problem for this PR.
(Now, I would keep this PR at "draft" mode") >
Update : Russ clarified (see his following comment) that the successful building of GSI using GNU is enough for this PR.