Uploaded image for project: 'Data Management'
  1. Data Management
  2. DM-13847

Investigate AST inefficiencies

    Details

    • Type: Story
    • Status: Done
    • Resolution: Done
    • Fix Version/s: None
    • Component/s: afw
    • Labels:
      None
    • Templates:
    • Team:
      External

      Description

      Profiling measureCoaddSources.py reveals that about 1/3 of the time is going to AST transforms, which is much larger than desirable. For example, 15% of the time is spent in malloc, which is used by GetMapping, which is called by the ast::Mapping::apply* family of functions; and ast::Mapping::_tran (32.9%) calls through to something called ValidateMapping.part.59 which uses a whopping 12.8%. The result is that when getting the PSF, we are spending more time in lsst::afw::geom::linearizeTransform than we are in warping the pixels to the coadd frame.

      We need to determine whether our use of AST is inefficient or AST is inherently inefficient. Either way, we need to devise a strategy to deal with this.

        Attachments

          Issue Links

            Activity

              People

              • Assignee:
                price Paul Price
                Reporter:
                price Paul Price
                Reviewers:
                Russell Owen
                Watchers:
                Jim Bosch, Paul Price, Russell Owen, Tim Jenness
              • Votes:
                0 Vote for this issue
                Watchers:
                4 Start watching this issue

                Dates

                • Created:
                  Updated:
                  Resolved:

                  Summary Panel