NRAO
 

Astronomical Image Processing System
31DEC00 Version

[AIPS]

The 31DEC00 version of AIPS is no longer available. See 31DEC02 for the current frozen release of AIPS. You might even consider 31DEC03 for the current development release of AIPS. It changes but there is a midnight job that no longer requires ssh.

AIPS and GNU G77 under Linux

Since 15APR98, AIPS has used the EGCS version of the GNU compilers with great success, mainly as it gives a considerable performance boost over the older gcc/f2c package. We now require use of the EGCS compilers for compiling AIPS with g77. Version 1.1.2 is used in-house; we have not yet tested the newly merged 2.95 version of gcc/g77 at this time, though we do not expect any surprises (and initial reports from the field indicate that the new g77 is in fact fine). AIPS assumes the glibc based C library.


WHATSNEW in 31DEC00 AIPS


See 31DEC00 AIPSLetter dated 31 December 2000 (PS, 200 Kbytes) for information on changes before about 15 December 2000. It is also available gzipped, 78 Kbytes).

See 31DEC00 AIPSLetter dated 1 April 2000 (PS, 156 Kbytes) for information on changes before about 1 April 2000. It is also available gzipped, 65 Kbytes).


  1. IMAGR offers the option to remake all fields every n'th cycle in the "OVERLAP 2" mode by setting OVERLAP to n (> 2). You are allowed to have no Clean boxes in a field. The task failed to taper data when natural weighting was requested.
  2. A bug in the Clean Component model division/subtraction routines had a small error in the frequency scaling of U and V. This affected the ability to model and remove distant interfering sources. Gridded subtraction was made faster in large cases and a bug preventing the SMODEL option from functioning was corrected.
  3. FACES is a new task to make a data model from the NVSS or WENSS catalogs, complete with images and CC files. This task, plus, SETFC, PBCOR and PATGN have new beam pattern parameters found by Rick Perley using the new fitting task PBEAM.
  4. HGEOM and OHGEO have changed inputs to allow more control of the output reference pixel. Read the help files carefully.
  5. FILLM now has the DOWEIGHT adverb to control which weighting scheme is used. FILLM will stop on end of tape. Errors recognizing when sources and modes have changed were fixed as was an error in correcting for nominal sensitivities in solar mode.
  6. CALIB and FRING could make errors in handling fast switching data, leading to noisy solutions interspersed among good ones.
  7. There is a suite of procedures intended to simplify VLBA data reduction. Type HELP VLBAUTIL for information.
  8. MATCH is a new task to renumber sources, antennas, and FQids so that one data set will match another.
  9. FXPOL was re-written to handle much more complicated cases. It now writes a new output data set in all cases.
  10. FLGIT handles the cutoff levels improperly, causing it to delete more data than expected.
  11. A new verb - ~ (the tilde sign) - was added to allow setting a scalar adverb (e.g. RASHIFT(8)) and beyond to list of values. This solves the line-length limit problem in AIPS when setting array values.
  12. XAS, the TV display, now works on 24-bit as well as 32-bit true color displays.
  13. UVPLT can plot weights, elevation, hour angle, etc and uses REFANT to control which antenna they refer to.
  14. TVLABEL now supports the GRCHAN adverb.
  15. Y2K is a new test suite for aips tasks and larger computers.

Distribution Statistics

Courtesy of Ernie Allen.

As of October 30, 358 sites had downloaded the 31DEC00 version of AIPS (22 of these appear to be "freshening" their installation at times, a number also run the mnj).


Modified on $Date: 2007/10/29 20:59:40 $
[Eric W. Greisen]