ENH: Optimize tensor-product B-Spline kernel evaluation #157
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
oesteban
force-pushed
the
fix/auto-lowmem-mode-followup
branch
from
f663d55 to
e4cc7ed
Compare
We're still having memory issues when interpolating with TOPUP-generated coefficients.
This implementation hopes to be more optimal than the previous one. Instead of calculating the tensor-product B-Spline weights of every point w.r.t. every control point, it calculates the weights along each axis and then calculates the tensor-product for the whole grid. In principle, the number of calls to ``np.piecewise`` has to have dramatically dropped. Memory utilization should be also more optimal, as there're only one very short-lived and large array (before it is converted to sparse matrix).
oesteban
force-pushed
the
fix/auto-lowmem-mode-followup
branch
from
e4cc7ed to
f9498fa
Compare
Codecov Report
@@ Coverage Diff @@
## master #157 +/- ##
==========================================
- Coverage 94.49% 93.57% -0.93%
==========================================
Files 19 19
Lines 1181 1229 +48
Branches 153 162 +9
==========================================
+ Hits 1116 1150 +34
- Misses 37 46 +9
- Partials 28 33 +5
Flags with carried forward coverage won't be shown. Click here to find out more.
Continue to review full report at Codecov.
|
oesteban
added a commit
that referenced
this pull request
Dec 24, 2020
2.0.0rc4 * FIX: Convert SEI fieldmaps given in rad/s into Hz (#127) * FIX: Limit ``3dQwarp`` to maximum 4 CPUs for stability reasons (#128) * ENH: Generate a simple mask after correction (#161) * ENH: Increase unit-tests coverage of ``sdcflows.fieldmaps`` (#159) * ENH: Optimize tensor-product B-Spline kernel evaluation (#157) * ENH: Add a memory check to dynamically limit interpolation blocksize (#156) * ENH: Massage TOPUP's fieldcoeff files to be compatible with ours (#154) * ENH: Add a simplistic EPI masking algorithm (#152) * ENH: Utility that returns the ``B0FieldSource`` of a given potential EPI target (#151) * ENH: Write ``fmapid-`` entity in Derivatives (#150) * ENH: Multiplex fieldmap estimation outputs into a single ``outputnode`` (#149) * ENH: Putting the new API together on a base workflow (#143) * ENH: Autogenerate ``B0FieldIdentifiers`` from ``IntendedFor`` (#142) * ENH: Finalizing the API overhaul (#132) * ENH: Keep a registry of already-used identifiers (and auto-generate new) (#130) * ENH: New objects for better representation of fieldmap estimation (#114) * ENH: Show FieldmapReportlet oriented aligned with cardinal axes (#120) * ENH: New estimation API (featuring a TOPUP implementation!) (#115) * DOC: Minor improvements to the literate workflows descriptions. (#162) * DOC: Fix typo in docstring (#155) * DOC: Enable NiPype's sphinx-extension to better render Interfaces (#131) * MAINT: Drop Python 3.6 (#160) * MAINT: Enable Git-archive protocol with setuptools-scm-archive (#153) * MAINT: Migrate TravisCI -> GH Actions (completion) (#138) * MAINT: Migrate TravisCI -> GH Actions (#137) * MAINT: Minimal unit test and refactor of pepolar workflow node (#133) * MAINT: Collect code coverage from tests on Circle (#122) * MAINT: Test minimum dependencies with TravisCI (#96) * MAINT: Add FLIRT config files to prepare for TOPUP integration (#116)
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.
We're still having memory issues when interpolating with TOPUP-generated coefficients.