Conversation
…(memmapped maybe) to PrismProbe.calculateProbesFromS, and use einsum. 256x256 probes is now about 5 minutes
…ropping real-space before cropping)
…create a dummy HAADFData object, get the mask from it, do the summation, and load the result into the HAADFData object as we go
…_prism medium FOV 256x256 now runs without memmapping in 1.5GB ram, wowwzers
pulling straight-to-main changes back to beta
Staging to beta, hopeful on to main next
…e is rolling: prep,potential,rolling,ffting are 16,16,67,1 percent of time
…t does not go in calculator, and preserves the original exit-wave in wf_data.probability for reuse
Staging to beta
Collaborator
Author
|
performance update (in how probes are pixel-shifted when min_dk is used, 5x speed-up when this is used for large systems), and electron counts functionality added (addressing issue #80). tests pass, and changes moved to beta this morning. so these changes can be included in this PR |
Critical fixes:
- Remove debug plt.savefig("probecropping.png") in applyShifts
- Fix PrismProbe empty slice when i1==0 (a[0:-0] is empty)
- Fix whichtimestep typo in wf_data.py plot_reciprocal
- Fix haadf_data.py in-place mutation of _wf_array
- Remove debug print("summation took...") in wf_data.py
Ported from main:
- Correct Kirkland Eq C.1 prefactor (47.878 V·Å²) for potential normalization
- Device-aware tensor creation (explicit device= on all zeros/ones calls)
- Remove global torch.set_default_device()
- Infer device from probe in Propagate when not specified
Regenerated test baselines for new normalization.
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
2 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
tested in staging, moved to beta, tested in beta
memory savings tricks include:
some of these have already made it to main, but were broken in certain scenarios. there is now more rigorous testing via tests/04_haadf_otf.py, which runs through all combinations of these arguments to make sure they run. note that in some cases, the result will differ, as some things are an approximation (prism, min_dk).
this will close the following issues: #75 #68 #67 #35 #33, and make #30 almost irrelevant.