Skip to content
Snippets Groups Projects

Chunks for corrected JUNGFRAU data should contain entire images

Merged Thomas Kluyver requested to merge jf-output-chunks into master

While investigating something else, I noticed that calibrated JUNGFRAU data is chunked with a chunk shape of 1 × 1 × 1 × 512. I believe this is a mistake - it means that each single module image is stored in 1024 tiny chunks. This PR makes it 1 × 1 × 512 × 1024, or one image per chunk, like we do for AGIPD.

This makes the files about 5% smaller in my test, and makes writing much quicker (~20 seconds to ~5 seconds per file). Writing took longer than processing before, so this is a significant speedup. Reading will also be quicker, assuming people usually read complete images (it will be slower if you read a single row of pixels).

cc @ahmedk

Merge request reports

Loading
Loading

Activity

Filter activity
  • Approvals
  • Assignees & reviewers
  • Comments (from bots)
  • Comments (from users)
  • Commits & branches
  • Edits
  • Labels
  • Lock status
  • Mentions
  • Merge request status
  • Tracking
Please register or sign in to reply
Loading