This repository was archived by the owner on Dec 22, 2022. It is now read-only.
a2b_ord4 with less stencils but still validating and not slower #493
Conversation
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
This reverts commit 17f45f9.
Contributor
Author
|
launch jenkins |
Contributor
Author
|
launch jenkins |
twicki
suggested changes
May 20, 2021
| qin[1, 1, 0], | ||
| ) | ||
| qout = (ec1 + ec2 + ec3) * (1.0 / 3.0) | ||
| tmp = 0.0 |
Contributor
There was a problem hiding this comment.
is it still an issue to have regions first?
If we need this, should we put a TODO comment here?
Contributor
Author
There was a problem hiding this comment.
it is still an issue -- a horizontalIf error raises . I added a TODO
fv3core/stencils/a2b_ord4.py
Outdated
| from __externals__ import i_end, i_start | ||
|
|
||
| with computation(PARALLEL), interval(...): | ||
| # ppm_volume_mean_x |
Contributor
There was a problem hiding this comment.
can we make this a docstring?
Contributor
Author
There was a problem hiding this comment.
I now use a gtscript function instead
… of memory, but is slower, 5 seconds for some reason
Contributor
Author
|
blarg is 10x slower?! |
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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Outdated suggestions cannot be applied.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.
Purpose
a2b_ord4 was made into a class, but the stencils were only merged a bit towards the single stencil pr version. This takes another swing at it to try to improve speed and gpu utilization. It still seems to need to be split in multiple stencils to validate on the gt backends, but removed the blocking computation dependence, getting the different regions to not rely on each other by recomputing components rather than using a previously computed offset value. Removed qxx and qyy temporaries. corner computations are done in a larger stencil rather than one at a time. Putting 2 together worked and no longer ran into the nan issue that had been happening before using gtscript functions in regions. Putting all 4 together results in an error with the gtcuda backend:
excessive recursion at instantiation of class "gridtools::meta::lazy::lfold<gri dtools::meta::dedup_step_implCode changes:
Checklist
Before submitting this PR, please make sure: