Difference between revisions of "ITK Registration Optimization/2007-05-01-tcon"

From NAMIC Wiki
Jump to: navigation, search
 
 
Line 32: Line 32:
 
*** Work with Julien to push script to BM engine directly (get feedback on success/fail)
 
*** Work with Julien to push script to BM engine directly (get feedback on success/fail)
 
*** Keep in touch with Steve P and Jim M.
 
*** Keep in touch with Steve P and Jim M.
 +
** Work on 3D image review over the web
 +
*** Get code from Christian
  
 
== Stephen ==
 
== Stephen ==

Latest revision as of 19:35, 8 May 2007

Home < ITK Registration Optimization < 2007-05-01-tcon

Brad

  • Waiting to hear back from Kilian
    • Kilian has the data
    • Open questions:
      • Time
      • Access to binaries (no)
      • Parameters being used (e.g., # of nodes / level)
    • Unknown schedule for this task - Kilian is on vacation
  • New images committed to test directory
  • To-do:
    • BSpline pipeline

Julien

  • Updated to new graphing software
  • Created new graphs
    • Summarize by machine, test, and day
  • Working with Seb on SSHFS
    • Will investigate hard links on same partition (Worked!)
  • To-do:
    • Continue to improve graphs and reports
    • Work with Seb on BatchMake modules for slicer
      • Will adapt BatchMake XML to support Slicer XML
      • Will extract BatchMake components for use by slicer module

Seb

  • Tuned tests to not timeout or use too much memory
  • Investigating SSHFS with Julien
  • Establishing a VS Express dashboard
  • To-do:
    • Investigate BatchMake module for Slicer
      • Design GUI
      • Work with Julien to push script to BM engine directly (get feedback on success/fail)
      • Keep in touch with Steve P and Jim M.
    • Work on 3D image review over the web
      • Get code from Christian

Stephen

  • Optimizing tests using valgrind and vtune
  • Investigating memory hogging bspline transforms and interpolations
  • Investigated memset for 0'ing arrays etc

Overall

  • Need to define a date for fixing tests. Right now the over-time graphs are confounded by the changing tests
  • Need to work on IJ articles
    • One by Seb/Julien on cmake/batchmake
    • One by Stephen on optimizations
    • One by Brad on EM (?)