Rotate vels#614
Conversation
- This is required for our BOEM-funded Arctic project. It needs something similar for SIS2 as well.
- OK to squash merge
Hallberg-NOAA
left a comment
Hallberg-NOAA
left a comment
There was a problem hiding this comment.
These changes look good to me, but I do have one change that I would like to request. In the upcoming set of revisions to give rotational symmetry with Fused-Multiply-Adds (FMAs), we are going to be writing expressions like Ax * Bx + Ay * By as (Ax * Bx) + (Ay * By) so that the compiler does not have the option of fusing the x- or y- parts of the expressions but not the other one. It would be helpful if you could revise the expressions for ssu_east and ssv_north to either add the parentheses to prevent FMAs from adding an asymmetry, or at least write these expressions so that the cos_rot terms come first in both cases. This is, of course, a diagnostic so perhaps we need not be so particular here, but for a new contribution like this, it is probably easier to make this small change now rather than coming back to it later.
Hallberg-NOAA
left a comment
Hallberg-NOAA
left a comment
There was a problem hiding this comment.
Thank you for this updated PR.
|
@kshedstrom Can you enable "Allow rebase merging" in the Settings of ESMG/MOM6? This would allow us to update this PR to the latest dev/gfdl. |
|
It looks like it is already set. Is it not behaving that way? |
|
I'm not able to update this PR. But let's try to sort this out privately. |
|
Gaea regression: https://gitlab.gfdl.noaa.gov/ogrp/MOM6/-/pipelines/23445 ✔️ 🟡 This requires an update to |
|
This PR was rebased into this commit: bb8a653 |
3 participants
This adds an option to save true north ssu and ssv, required for our BOEM contract. There are similar changes to SIS2.