Code

Richardson Lucy Parallelization V2#274

Code

Open

avalluvan wants to merge 50 commits intocositools:developcositools/cosipy:developfrom avalluvan:feature/RLparallelavalluvan/cosipy:feature/RLparallelCopy head branch name to clipboard

+607-63

Commits

Commits on Aug 14, 2024

• Added new option "RLparallel" to ImageDeconvolution.deconvolution_algorithm_classes

  avalluvan

  committed
  79109f3

  Copy full SHA for 79109f3

• Renamed RLparallel.py to RichardsonLucyParallel.py and adapted code skeleton from RichardsonLucy.py

  avalluvan

  committed
  c918ab5

  Copy full SHA for c918ab5

• Installed mpi4py to cosipy venv and added it to skeleton

  avalluvan

  committed
  015f648

  Copy full SHA for 015f648

• Created subprocess call. Ported MPI to separate script RLparallelscript.py

  avalluvan

  committed
  f259d36

  Copy full SHA for f259d36

• Added code to register results if save_results_flag is set to True

  avalluvan

  committed
  4cde4dd

  Copy full SHA for 4cde4dd

• End-to-end working script version commit

  avalluvan

  committed
  e98fb79

  Copy full SHA for e98fb79

• Unclear what files are causing conflicts

  avalluvan

  committed
  cd45e44

  Copy full SHA for cd45e44

Commits on Aug 27, 2024

• Add modified files and new RL parallel script

  Show description for ef279f4
  avalluvan

  authored
  ef279f4

  Copy full SHA for ef279f4

• Another tiny change

  Show description for 871c9a8
  avalluvan

  authored
  871c9a8

  Copy full SHA for 871c9a8

Commits on Aug 29, 2024

• Modified configuration file for RLparallel

  avalluvan

  authored
  27c8afe

  Copy full SHA for 27c8afe

Commits on Oct 18, 2024

• Merge branch 'cositools:develop' into develop

  avalluvan

  authored
  f079988

  Copy full SHA for f079988

• git fetch changes unclear. Performing a commit to avoid data loss.

  avalluvan

  committed
  f044589

  Copy full SHA for f044589

• This merge may result in data losses. Copies of files have been created to subsequently overwrite remote.

  Show description for 3036a7a
  avalluvan

  committed
  3036a7a

  Copy full SHA for 3036a7a

Commits on Oct 24, 2024

• Create LMDR.ipynb and add response creation attempt with h5py.

  Show description for 35b8293
  avalluvan

  committed
  35b8293

  Copy full SHA for 35b8293

Commits on Oct 25, 2024

• Created general code for multidimensional interpolation

  avalluvan

  committed
  d04f3c3

  Copy full SHA for d04f3c3

Commits on Oct 30, 2024

• Attempted listmode response on the fly generation

  avalluvan

  committed
  ac47f23

  Copy full SHA for ac47f23

Commits on Nov 3, 2024

• Add ListModeResponse.py

  avalluvan

  committed
  c257676

  Copy full SHA for c257676

Commits on Nov 6, 2024

• Used histpy functions to dramatically simplify the codebase

  avalluvan

  committed
  7de27fd

  Copy full SHA for 7de27fd

Commits on Nov 7, 2024

• Modified code to accept len(dr.axes) or len(dr.axes)-1 number of inputs.

  Show description for 4c23188
  avalluvan

  committed
  4c23188

  Copy full SHA for 4c23188

Commits on Nov 9, 2024

• Added comments to get_point_source_response()

  avalluvan

  committed
  71085fd

  Copy full SHA for 71085fd

Commits on Nov 11, 2024

• Tested new interpolation scheme on LMDR.ipynb

  avalluvan

  committed
  dc35034

  Copy full SHA for dc35034

• Adhoc fix for interpolated psr calculation with deepcopy

  avalluvan

  committed
  f8bbaa4

  Copy full SHA for f8bbaa4

• Simplified __getitem__ and get_point_source_response()

  Show description for cdcb9b0
  avalluvan

  committed
  cdcb9b0

  Copy full SHA for cdcb9b0

• Ported ListModeResponse interpolation functionalities to DetectorResponse

  avalluvan

  committed
  aaff648

  Copy full SHA for aaff648

• Removed ListModeResponse.py. All developed features incorporated in DetectorResponse.

  avalluvan

  committed
  caa2d44

  Copy full SHA for caa2d44

• Merge pull request #1 from avalluvan/feature/general_response

  Show description for 2c80caa
  avalluvan

  authored
  2c80caa

  Copy full SHA for 2c80caa

Commits on Nov 21, 2024

• Added code to compress empty axes, i.e., those with axis shape = (1,)

  avalluvan

  committed
  dafd4f8

  Copy full SHA for dafd4f8

Commits on Dec 10, 2024

• Create new file RichardsonLucyWithParallelSupport.py. Stash changes to other files

  avalluvan

  committed
  de31c0a

  Copy full SHA for de31c0a

• Skeleton for RLWithParellelSupport.py. Need to update MPI.DOUBLE with custom data types

  avalluvan

  committed
  ebdf1ca

  Copy full SHA for ebdf1ca

• Looks like RLWithParallelSupport.py is complete. Need to create data interface and main script to test the implementation

  avalluvan

  committed
  9aeba42

  Copy full SHA for 9aeba42

Commits on Dec 11, 2024

• Everything up to run_deconvolution works with new data interface WPS

  avalluvan

  committed
  b1d61a4

  Copy full SHA for b1d61a4

• Dry run with single MPI process works with RichardsonLucy and dataIFWithParallelSupport

  avalluvan

  committed
  61afc80

  Copy full SHA for 61afc80

• RichardsonLucyWPS.py works with both dataIF_DC2 and dataIFWPS for single node execution

  avalluvan

  committed
  75b0b81

  Copy full SHA for 75b0b81

Commits on Dec 12, 2024

• Working set up for small numproc MPI runs

  avalluvan

  committed
  5d5eb3e

  Copy full SHA for 5d5eb3e

• Syntactically correct. Algorithm has some discrepancies

  avalluvan

  committed
  59f9ce8

  Copy full SHA for 59f9ce8

• Archive old version of RLparallelscript.py and RichardsonLucyParallel.py

  Show description for 406cf9a
  avalluvan

  committed
  406cf9a

  Copy full SHA for 406cf9a

Commits on Dec 13, 2024

• Single and multi process produce same, exact outputs!

  Show description for 2eb7e25
  avalluvan

  committed
  2eb7e25

  Copy full SHA for 2eb7e25

• Replaced old RichardsonLucy.py with new RichardsonLucy.py that now supports

  Show description for 1cd0d2c
  avalluvan

  committed
  1cd0d2c

  Copy full SHA for 1cd0d2c

• Minor bug in image_deconvolution now that RichardsonLucyParallel.py has been removed. Bug fixed.

  avalluvan

  committed
  1f4eeed

  Copy full SHA for 1f4eeed

• Removed unnecessary imports in new RichardsonLucy.py

  avalluvan

  committed
  14c34fd

  Copy full SHA for 14c34fd

• Added comments. Removed stubs of RLparallel. Ready to rebase and create pull request

  avalluvan

  committed
  860b5ce

  Copy full SHA for 860b5ce

Commits on Dec 15, 2024

• Migrated RichardsonLucy slicing and message passing to DataIFWithParallel

  Show description for 63527db
  avalluvan

  committed
  63527db

  Copy full SHA for 63527db

• Merge branch 'develop' into feature/RLparallel

  avalluvan

  authored
  82fc4f8

  Copy full SHA for 82fc4f8

• Bug fix in reshaping and concatenating epsilon and C slices

  avalluvan

  committed
  c085c17

  Copy full SHA for c085c17

Commits on Dec 16, 2024

• Replacing superfluous edits with files from cositools/develop

  avalluvan

  committed
  279f6f2

  Copy full SHA for 279f6f2

• Undo changes to dataIF_COSI_DC2.py and allskyimage.py

  avalluvan

  committed
  04eb674

  Copy full SHA for 04eb674

• Undoing changes to RichardsonLucy.py Estep structure

  avalluvan

  committed
  b99ff6a

  Copy full SHA for b99ff6a

• Incorporated review suggestions about placement of parallel_computation and master_node flags

  avalluvan

  committed
  cddaa24

  Copy full SHA for cddaa24

• Rename dataIFWithParallel.py to dataIF_Parallel.py. Undo changes to image deconvolution config file.

  avalluvan

  committed
  7a2f65c

  Copy full SHA for 7a2f65c

• Calculate parallel and master_node flags in RLparallelscript.py

  avalluvan

  committed
  17aae47

  Copy full SHA for 17aae47