User Tools

Site Tools


difx:difx-todolist-2019

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
Next revision
Previous revision
difx:difx-todolist-2019 [2019/11/15 12:28]
adamdeller
difx:difx-todolist-2019 [2020/02/24 16:49] (current)
adamdeller
Line 3: Line 3:
 == To be prioritised == == To be prioritised ==
  
-  * Why is complex interlaced VDIF data not fringing in DiFX-2.6.1? 
-  * Test outputbands on ASKAP oversampled/​overlapping data after changing autobands logic to select channels closest to the centre of recorded band in the case that multiple options are available. 
-  * PolConvert single dish mode and document 
-  * DiFX segfaulting on OSX 
-  * difx2mark4 pcal support for multiple streams 
-  * Tbd Shanghai correlation issue 
-  * Adding cpumon to report current cpu speed and current power dissipation 
-  * Mark6 playback speed (locally high, but via network speeds are slower) 
-  * Python3: audit of current changes, what is the strategy going forward. 
-  * Libraries not appearing when running DiFX on ozstar under slurm 
-  * MPI parallel tracing of DiFX to look for bottlenecks 
-  * Generate some one page EOIs for the next call for SSEs on ADACS 
-  * See whether ADACS would co-supervise DiFX student 
-  * How to allocate nodes in a specific order under slurm 
-  * Absolute paths in difx input files - is there a way to deal with them better? 
-  * Virtual trunk missing components 
  
 == Standing Items == == Standing Items ==
Line 26: Line 10:
     * Reconsider the approach...     * Reconsider the approach...
   * Update the pdf documentation for currently undocumented tools, and keep updated (Adam, WFB)   * Update the pdf documentation for currently undocumented tools, and keep updated (Adam, WFB)
- 
-== Before the end of the week == 
- 
  
 == High Priority == == High Priority ==
 +  * Determine why complex interlaced VDIF data is not fringing in DiFX-2.6.1 (CP, AD, JW)
 +  * Change outputbands logic to choose appropriate spectral regions in overlap cases (JW)
   * DiFX pcals with multiple datastreams (fix in difx2mark4). In progress. (HR, John)   * DiFX pcals with multiple datastreams (fix in difx2mark4). In progress. (HR, John)
   * Mark6 group handing plan (WFB, JJ, MW, JW, HR)   * Mark6 group handing plan (WFB, JJ, MW, JW, HR)
Line 50: Line 33:
     * Support for $BITSTREAMS largely complete in vex2difx     * Support for $BITSTREAMS largely complete in vex2difx
     * Plan: finalise, test     * Plan: finalise, test
 +  * Mark6 playback speed (locally high, but via network speeds are slower)
  
 == Medium Priority == == Medium Priority ==
 +  * Update vdifio to make the utilities compatible with Legacy VDIF (AD)
 +  * Test outputbands on ASKAP oversampled/​overlapping data after changing autobands logic is updated for the overlap regions (AD)
   * Ocean loading: index by position rather than antenna name/code (Walter? + David Gordon)   * Ocean loading: index by position rather than antenna name/code (Walter? + David Gordon)
     * Want to circumvent 2-letter name overload, and ability to have two codes for one station     * Want to circumvent 2-letter name overload, and ability to have two codes for one station
Line 57: Line 43:
   * Documentation of polconvert on the wiki (Cormac)   * Documentation of polconvert on the wiki (Cormac)
     * Awaiting test data from Parkes     * Awaiting test data from Parkes
 +  * PolConvert single dish mode and document (GC)
   * Update vex2xml for vex2 (GC)   * Update vex2xml for vex2 (GC)
   * difx2fits: add option to split files at a given size (WB, CR)   * difx2fits: add option to split files at a given size (WB, CR)
Line 72: Line 59:
   * Cleanly separate data format and data source (DiFX 3)   * Cleanly separate data format and data source (DiFX 3)
   * Option for zero-padded FFT   * Option for zero-padded FFT
 +  * Extending cpumon to report current cpu speed and current power dissipation (WB)
 +  * Generate some one page EOIs for the next call for SSEs on ADACS (AD)
 +  * How to allocate nodes in a specific order under slurm (AD,CR,HR)
  
 == Low Priority == == Low Priority ==
 +  * Virtual trunk missing components (find out what is missing and update)
   * Use extracted pulse cal to predict correlated tones and then remove from visibilities   * Use extracted pulse cal to predict correlated tones and then remove from visibilities
   * Investigate whether merging pcal plotting tools is possible (Jan)   * Investigate whether merging pcal plotting tools is possible (Jan)
Line 97: Line 88:
     * Fully complete now.  Needs testing.     * Fully complete now.  Needs testing.
   * Add DiFX diagnostics   * Add DiFX diagnostics
 +  * MPI parallel tracing of DiFX to look for bottlenecks
  
 == Fantasy wishlist == == Fantasy wishlist ==
difx/difx-todolist-2019.1573781283.txt.gz · Last modified: 2019/11/15 12:28 by adamdeller