Fix numba parallel issues #2447
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
Codecov Report
@@ Coverage Diff @@
## master #2447 +/- ##
=======================================
Coverage 66.91% 66.91%
=======================================
Files 153 153
Lines 13636 13634 -2
=======================================
- Hits 9124 9123 -1
+ Misses 4512 4511 -1
📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more |
|
This is great to see fixed. Do you have examples showing the improvement to reproducibility? |
|
Yes, here is a plot showing the comparison between a serial and a parallel run (10 threads) of the Before the fix this would have been the result: I didn't do more extensive testing, but I think this is such a trivial fix/ clear case that more extensive testing isn't necessary |
|
Great, thank you! Next up we need to actually make the monte carlo loop parallel with numba, but that is a LOT more work. |
|
Check out this pull request on See visual diffs & provide feedback on Jupyter Notebooks. Powered by ReviewNB |
|
*beep* *bop* Hi, human. The Click here to see your results. |
|
@Rodot- signed off. |
wkerzendorf
approved these changes
Oct 17, 2023
atharva-2001
added a commit
to atharva-2001/tardis
that referenced
this pull request
Nov 28, 2023
* Allow Simulation Tests to Run Independently (tardis-sn#2407) * Modify test_packet_source.py so that sim and tardis/tests can run * remove BasePacketSource * format * Add docstrings * Pre-release 2023.09.17 (tardis-sn#2411) Automated changes for pre-release 2023.09.17 * upgrading env in 2022 (tardis-sn#2410) * fix several bugs related to upgrade * fix the get function in pandas * fixed several incompatibilities for new libraries * fixed up the changes in pandas testing and other reindexing issues * fix other bugs with new environment * blackify tardis * updated the conda-lock file * add lock files * update cache number * fix simulation tests * remove units for testing * blackify * Update tardis/visualization/tools/tests/test_sdec_plot.py good catch Co-authored-by: Atharva Arya <[email protected]> * add assert * reset caches --------- Co-authored-by: Atharva Arya <[email protected]> * In gamma ray code : Changed times to np.geomspace (tardis-sn#2415) * Changed times to np.geomspace * Added myself to mailmap * rename RadialModel1D to SimulationState (tardis-sn#2417) * rename RadialModel1D to SimulationState * black files * rename simulation_state2 (tardis-sn#2421) * fixup of many files to make them pass the tests * change model for typo * Ignore tests in docstr coverage (tardis-sn#2424) ignore tests in docstr coverage * Disabled formal integral for continuum interaction (tardis-sn#2426) Disabled formal integral Co-authored-by: Alexander Holas <alexander.holas@h-its> * add linelist exposure to atom data and reader, built on current (tardis-sn#2428) * add linelist exposure to atom data and reader, built on current * add josh to mailmap * Modified how setup input energy takes isotope name. Earlier it was ga… (tardis-sn#2425) Modified how setup input energy takes isotope name. Earlier it was gamma_ray_lines.Isotope * Added a function to calculate shell masses (tardis-sn#2434) * Added a function to calculate shell masses * Added a function to calculate shell masses * changed shell masses with the new function * Fix duplicate entries in continuum line list (tardis-sn#2443) * Workaround list access * Formatting * Added error if duplicate * Remove duplicates * Fix typo, add comment * Remove bandaid fix * Cleanup * More cleanup * Fix numba parallel issues (tardis-sn#2447) * Enabled parallel * Fix the mistake * Update docs * Change glob patterns in `setup.cfg` (tardis-sn#2441) * Fix package_data in setup.cfg, mainly io and viz * more folders * Test pip installation * Remove test step * Fix more folders * restructure of geometry (tardis-sn#2422) * restructure of geometry * add radial1d boundary logic * black format * several fixes * fix epsilon * add testing of boundaries * change the r_inner_active * first integration with `from_config` working * hunting down density indexing bug * all model tests (without csvy) pass * more fixes * fix of model to simulation_state * fix inner boundary packet error * fix some leftovers * final fix for csvy * blackify * restructure to readers and remove some leftover code * further cleanup * first start of the restructure * add comment about removing quantitiness * add velocity check * add new abundance functions * remove default units * Add FAQ section to documentation (tardis-sn#2450) * Add faq * Added Overview * Fix typo * Remove lock file creation from pre release workflow (tardis-sn#2432) * Install latest version of mamba * Unpin conda lock installatiion * Split lockfilee creation to separate workflow * change docs installation- put a warning for conda forge installation and comment out install from package * fix * delete create lockfiles * download llock file * Pre-release 2023.10.20 (tardis-sn#2452) Automated changes for pre-release 2023.10.20 * Add verbose assert to NLTE rate equation (tardis-sn#2457) * add verbose assert * formatting * Fix bug in relativistic packet source (tardis-sn#2453) * Fix bug in relativistic packet source * Add initial test for BlackBodySimpleSourceRelativistic * Remove incorrect docstring * Update tardis/montecarlo/tests/test_packet_source.py Co-authored-by: Wolfgang Kerzendorf <[email protected]> * Update tardis/montecarlo/tests/test_packet_source.py Co-authored-by: Wolfgang Kerzendorf <[email protected]> * Rename all occurences of blackbodysimplesourcerelativistic * Do not hardcore blackbody_simplesource_relativistic.beta --------- Co-authored-by: Wolfgang Kerzendorf <[email protected]> * Enable nlte ionization as plasma component (tardis-sn#2458) * Rename nlte ion and electron * Enable NLTE * Fix typo * Add docs * Fix transitionprob dtype * dtype conversion * dtype for sparse matrix * Typing in numba * Fix test variable name * Add missing electron density case * Fix typo * Update docstr-cov.yml to fix failing action (tardis-sn#2462) * Update docstr-cov.yml to fix failing action * Explicitly install setuptools * NLTE Ionization solver polish (tardis-sn#2461) * assertion for first guess * Add checks to decay * Fix greater equal * Fix deprecated series index * Pre-release 2023.11.05 (tardis-sn#2466) Automated changes for pre-release 2023.11.05 * Tests Plasma Using Syrupy (tardis-sn#2413) * Add syrupy classes to conftest * Add syrupy to env file * Another folder for syrupy data * Save as npy instead of txt * Use syrupy fixtures * Different assertion functions for pandas dataframes and series * Zeta data fix * Blackify * Blackify conftest * test_hdf_plasmas using syrupy * Install syrupy with conda * Remove additional line adding syrupy as a plugin * Set update snapshots config option when generate reference is selected * Remove old refdata code from test plasmas complete * Create plasma conftest file * Text based snapshot fixture with custom location * Snapshot path * tardis regressions my fork * Refactor test_hdf_plasmas.py * Format using black * test nlte excitation using syrupy * nlte solver * plasma continuum * test tardis model density config * Lock files * Increase cache no * format using black * cleanup- sort dependencies, delete comments * Documentation to run these tests * remove commented out fixtures which are not used anywhere * docs code render fix * tests final cleanup * renaming files, yet to deal with ambr snapshot extension and curly brackets * renaming snapshots * Add comments * Add astropy import guard (tardis-sn#2470) * Add astropy import guard * mailmap * Reading in decay radiation data in atom data (tardis-sn#2471) * Reading in decay radiation data in atom data * Added docstring * Fixes the checksum * Pre-release 2023.11.26 (tardis-sn#2477) Automated changes for pre-release 2023.11.26 * Update plasma configuration documentation (tardis-sn#2459) * update plasma docs * Update mailmap * Fix typo * Add ruff rules and docs (tardis-sn#2478) * Add ruff rules and docs * Update pyproject.toml Co-authored-by: Wolfgang Kerzendorf <[email protected]> * rebuild docs --------- Co-authored-by: Wolfgang Kerzendorf <[email protected]> * Decay energy chain (tardis-sn#2448) * Added a function to calculate shell masses * Added a function to calculate shell masses * changed shell masses with the new function * Changed mass fraction to masses in to_inventories() * Co-authored-by: Wolfgang Kerzendorf <[email protected]> * Added a function to calculate total decays * Added a function to calculate energies from gamma rays and positrons. * added a function to calculate average energies of gamma rays and positrons * Added a fucntion to calculate each decay chain energies * Added dictionaries to handle multiple isotopes * Changed value to values * added tests for gamma_ray_transport * Added tests for calculating activity * Added test for activity * Added tests for two isotope * Changed Ni_isotope_mass * Added pytest paramterize * Added test for calculating shell masses * Ran test for checking activity of parent nuclide with analytical solution. * The function test_activity matches with the radioactivedecay output upto 14 decimal places. The Avogadro's number used in radioactivedecay is not from astropy. * Added tests for checking if iso_dict is returning the right key. * Added test for inventories dictionary. * Added a test to check if the calculate_average_energy function passes the right list ahead. * Added new function for testing energy budget from each decay chain. * Added a new function for energy per mass * Reading in decay radiation data in atom data * Add * Added tests for gamma ray transport. * Added tests for all functions for gamma_ray_transport. Added docstrings. * Changing decay energy chain * Added a function to get taus * Added tests for multiple isotopes * Fixes the test calculate shell masses with hand calculated values * Use Mamba instead of conda * Run asv for last 5 commits * Add pull request trigger temporarily * Reset --------- Co-authored-by: tardis-bot <[email protected]> Co-authored-by: Wolfgang Kerzendorf <[email protected]> Co-authored-by: Anirban <[email protected]> Co-authored-by: Alexander Holas <[email protected]> Co-authored-by: Alexander Holas <alexander.holas@h-its> Co-authored-by: Joshua Shields <[email protected]> Co-authored-by: Anirban <[email protected]> Co-authored-by: Christian Vogl <[email protected]> Co-authored-by: Andrew <[email protected]>
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.
📝 Description
Type: 🪲
bugfixFixes the numba parallel issues mentioned in #2021 (at least for normal packets)
📌 Resources
None.
🚦 Testing
How did you test these changes?
☑️ Checklist
build_docslabel