Use parallel processing to speed up obs processing #733
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
CoryMartin-NOAA
commented
Nov 15, 2023
|
|
||
| # Check if the converter was successful | ||
| # if os.path.exists(yaml_output_file): | ||
| # rm_p(yaml_output_file) | ||
|
|
||
| # run all bufr2ioda yamls in parallel | ||
| with mp.Pool(num_cores) as pool: |
There was a problem hiding this comment.
Note, the easiest way to do this was to split them into python and YAML+executable groups. If they are roughly equal in size, that probably is okay, but we may want to combine them all into the same pool?
RussTreadon-NOAA
approved these changes
Nov 16, 2023
There was a problem hiding this comment.
Not familiar with python multiprocessing. Is running in parallel from python the most efficient approach?
Copied ush/ioda/bufr2ioda/run_bufr2ioda.py to a working copy of gdas-validation. Ran gdasprepatmiodaobs. Log file indicates jobs ran in parallel.
^[[38;21m2023-11-16 00:54:20,329 - INFO - run_bufr2ioda.py: Executing /work2/noaa/da/rtreadon/gdas-validation/global-workflow/sorc/gdas.cd/ush/ioda/bufr2ioda/bufr2ioda_gpsro_bufr_combined.py -c /work2/noaa/da/rtreadon/gdas-validation/comrot/gdas_eval_satwind_JEDI/gdas.20210801/00/obs/gpsro_bufr_combined_2021080100.json^[[0m
^[[38;21m2023-11-16 00:54:20,329 - INFO - run_bufr2ioda.py: Executing /work2/noaa/da/rtreadon/gdas-validation/global-workflow/sorc/gdas.cd/ush/ioda/bufr2ioda/bufr2ioda_satwind_scat.py -c /work2/noaa/da/rtreadon/gdas-validation/comrot/gdas_eval_satwind_JEDI/gdas.20210801/00/obs/satwind_scat_2021080100.json^[[0m
^[[38;21m2023-11-16 00:54:20,330 - INFO - run_bufr2ioda.py: Executing /work2/noaa/da/rtreadon/gdas-validation/global-workflow/sorc/gdas.cd/ush/ioda/bufr2ioda/bufr2ioda_adpsfc_prepbufr.py -c /work2/noaa/da/rtreadon/gdas-validation/comrot/gdas_eval_satwind_JEDI/gdas.20210801/00/obs/adpsfc_prepbufr_2021080100.json^[[0m
^[[38;21m2023-11-16 00:54:20,330 - INFO - run_bufr2ioda.py: Executing /work2/noaa/da/rtreadon/gdas-validation/global-workflow/sorc/gdas.cd/ush/ioda/bufr2ioda/bufr2ioda_adpupa_prepbufr.py -c /work2/noaa/da/rtreadon/gdas-validation/comrot/gdas_eval_satwind_JEDI/gdas.20210801/00/obs/adpupa_prepbufr_2021080100.json^[[0m
^[[38;21m2023-11-16 00:54:20,330 - INFO - run_bufr2ioda.py: Executing /work2/noaa/da/rtreadon/gdas-validation/global-workflow/sorc/gdas.cd/ush/ioda/bufr2ioda/bufr2ioda_conventional_prepbufr_ps.py -c /work2/noaa/da/rtreadon/gdas-validation/comrot/gdas_eval_satwind_JEDI/gdas.20210801/00/obs/conventional_prepbufr_ps_2021080100.json^[[0m
^[[38;21m2023-11-16 00:54:20,330 - INFO - run_bufr2ioda.py: Executing /work2/noaa/da/rtreadon/gdas-validation/global-workflow/sorc/gdas.cd/ush/ioda/bufr2ioda/bufr2ioda_sfcshp_prepbufr.py -c /work2/noaa/da/rtreadon/gdas-validation/comrot/gdas_eval_satwind_JEDI/gdas.20210801/00/obs/sfcshp_prepbufr_2021080100.json^[[0m
^[[38;21m2023-11-16 00:54:20,330 - INFO - run_bufr2ioda.py: Executing /work2/noaa/da/rtreadon/gdas-validation/global-workflow/sorc/gdas.cd/ush/ioda/bufr2ioda/bufr2ioda_satwind_amv_goes.py -c /work2/noaa/da/rtreadon/gdas-validation/comrot/gdas_eval_satwind_JEDI/gdas.20210801/00/obs/satwind_amv_goes_2021080100.json^[[0m
^[[38;21m2023-11-16 00:54:20,331 - INFO - run_bufr2ioda.py: Executing /work2/noaa/da/rtreadon/gdas-validation/global-workflow/sorc/gdas.cd/ush/ioda/bufr2ioda/bufr2ioda_acft_profiles_prepbufr.py -c /work2/noaa/da/rtreadon/gdas-validation/comrot/gdas_eval_satwind_JEDI/gdas.20210801/00/obs/acft_profiles_prepbufr_2021080100.json^[[0m
The total run time for the parallel prepatmiodaobs was 06:28 (mm:ss)
The previous serial job took 11:06.
Nice reduction!
|
@RussTreadon-NOAA I'm not sure if it's the most efficient, but I think this will work fine provided we don't need to run on multiple nodes. This was the fastest way to speed everything up. Next, we may wish to combine the pools so that the |
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.
Using python multiprocessing to generate obs in parallel. The current list of obs goes from 10+ minutes to completing in ~5.5 minutes.