gmtb-gfsphysics optimization: speedup of ccpp init#48
Merged
climbfuji merged 5 commits intoMar 15, 2018
Merged
Conversation
…MP_generic_pre.f90
…calpreciptype.f90
…) derived type so that it gets allocated in the right place in memory (first touch principle): GFS_layer/GFS_driver.F90
…ata_block to avoid reading SDF multiple times, cleanup and formatting changes: IPD_layer/IPD_CCPP_driver.F90
… into gmtb-gfsphysics-optimization
Contributor
|
Dom - My tests today, and yesterday, are not giving bit-for-bit results, but I'm not sure that they should. The branch I used yesterday (gmtb-fv3-cleanup-and-error-handling) may or may not have had all of the same fixes as this one. Is this a concern worth tracking down? If your tests show bit-for-bit before/after, that's sufficient for me. |
Collaborator
Author
|
@llpcarson Did you create a new baseline or use my baseline? With the changes to the optimization flags in gmtb-gfsphysics PR47, the results have changed (see discussion there) and I did not update the "official" baseline yet. The following one works for me: /scratch4/BMC/gmtb/Dom.Heinzeller/gmtb-fv3/reference/intel/C96fv3gfs2016092900/ |
Contributor
|
I was just comparing my run from yesterday and mine from today, so I
believe both of them have the PR47 included, but I'll re-confirm
That was it --- the branch I tested yesterday did not have that PR included... so this one is ready to commit!
…On Wed, Mar 14, 2018 at 3:40 PM, Dom Heinzeller ***@***.***> wrote:
@llpcarson <https://github.com/llpcarson> Did you create a new baseline
or use my baseline? With the changes to the optimization flags in
gmtb-gfsphysics PR47, the results have changed (see discussion there) and I
did not update the "official" baseline yet. The following one works for me:
/scratch4/BMC/gmtb/Dom.Heinzeller/gmtb-fv3/reference/
intel/C96fv3gfs2016092900/
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#48 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AHTrInvKV9vyNS7RvwbdLOr-vOUYKqghks5teY5MgaJpZM4Sq3qH>
.
|
Collaborator
Author
|
My reference is from straight after pr47 went in, i.e. March 12/13, before the merge this morning and the PRs afterwards. I do get bfb identical results for the C96 test case. But I definitely want to understand why you don't! |
Contributor
|
My runs today match that baseline, too. So, the branch I grabbed earlier
was pre-PR47 (which was OK for timing, but not for b4b)
…On Wed, Mar 14, 2018 at 3:47 PM, Dom Heinzeller ***@***.***> wrote:
My reference is from straight after pr47 went in, i.e. March 12/13, before
the merge this morning and the PRs afterwards. I do get bfb identical
results for the C96 test case.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#48 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AHTrImwIzphn7EXSgSTDdIO5JbJnZfrJks5teZAOgaJpZM4Sq3qH>
.
|
Collaborator
Author
|
Great, thank you! |
climbfuji
deleted the
gmtb-gfsphysics-optimization-speedup-of-ccpp-init-rebased
branch
hannahcbarnes
pushed a commit
to hannahcbarnes/ccpp-physics
that referenced
this pull request
Apr 7, 2020
* fv3atm issue NCAR#37: fix the real(8) lat/lon in netcdf file * fv3atm NCAR#35: Reducing background vertical diffusivities in the inversion layers * fv3atm NCAR#24: bug in gfsphysics/physics/moninedmf_hafs.f * fv3atm NCAR#18: Optimize netcdf write component and bugfix for post and samfdeepcnv.f * set (0-1) bounds for ficein_cpl * remove cache_size due to lower netcdf verion 4.5.1 on mars * Change ice falling to 0.9 in gfsphysics/physics/gfdl_cloud_microphys.F90
climbfuji
pushed a commit
to climbfuji/ccpp-physics
that referenced
this pull request
Aug 10, 2020
…r_master Update gsd/develop from NCAR master
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This PR improves the performance of the CCPP implementation in FV3 v0 by
This PR also
The results are bit for bit compatible (tested on Theia/Intel and Macbook/GNU with OpenMP enabled).