fluids: Add PyTorch external DD SGS evaluation #1581
Merged
Conversation
This file contains 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
jrwrigh
force-pushed
the
jrwrigh/pytorch_external_sgs
branch
2 times, most recently
from
30f4204 to
50af404
Compare
jrwrigh
force-pushed
the
jrwrigh/pytorch_external_sgs
branch
4 times, most recently
from
ab6b800 to
984a964
Compare
jrwrigh
commented
May 20, 2024
jrwrigh
force-pushed
the
jrwrigh/pytorch_external_sgs
branch
3 times, most recently
from
26b57d3 to
31dc6d3
Compare
|
This is good for review now. Tests pass on Noether for CPU and CUDA. I'm having some difficulties getting PyTorch+ROCM to build with Spack, but that's not a huge priority since we don't have hardware to run that on right now. |
jrwrigh
force-pushed
the
jrwrigh/pytorch_external_sgs
branch
from
31dc6d3 to
8d82690
Compare
jeremylt
reviewed
Jun 12, 2024
jeremylt
approved these changes
Jun 12, 2024
jrwrigh
force-pushed
the
jrwrigh/pytorch_external_sgs
branch
from
8d82690 to
a8d73e7
Compare
- Rename sequential_internal -> *_ceed
To have the log_events accessbile to torch (in C++), I needed to separate out the header file containing the extern PetscLogEvent declarations. While I was at it, I figured it'd be more clear to have a separate log_events.c file as well to have the actual "storage" of the PetscLogEvents and the RegisterLogEvents function itself.
On Sunspot, on-device inference is not working reliably. I'm not sure exactly why at the moment (whether it's a libCEED backend issue or something else).
jrwrigh
force-pushed
the
jrwrigh/pytorch_external_sgs
branch
3 times, most recently
from
cd4fd42 to
280bdd9
Compare
jrwrigh
force-pushed
the
jrwrigh/pytorch_external_sgs
branch
from
280bdd9 to
637c7b1
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Add PyTorch as a external DD SGS evaluation. This is a "follow up" to #1361 where we use PyTorch to run the data-driven model instead of a native implementation.
ToDos:
PetscLogEvents(particularly for inference and the data transfer steps)USE_LIBTORCHtesting)