[daip] [!10531]: AIPS - IMAGR returns only NaN values in the image&beam/fails completely after upgrading OS to Fedora 25
Blazej Nikiel-Wroczynski
nraohelp at nrao.edu
Wed Jun 14 13:41:18 EDT 2017
Blazej Nikiel-Wroczynski updated #10531
---------------------------------------
IMAGR returns only NaN values in the image&beam/fails completely after upgrading OS to Fedora 25
------------------------------------------------------------------------------------------------
Ticket ID: 10531
URL: https://help.nrao.edu/staff/index.php?/Tickets/Ticket/View/10531
Full Name: Blazej Nikiel-Wroczynski
Email: iwan at oa.uj.edu.pl
Creator: User
Department: AIPS Data Reduction
Staff (Owner): -- Unassigned --
Type: Issue
Status: Open
Priority: Default
SLA: NRAO E2E
Template Group: Default
Created: 14 June 2017 05:41 PM
Updated: 14 June 2017 05:41 PM
Due: 16 June 2017 05:41 PM (2d 0h 0m)
Resolution Due: 22 June 2017 05:41 PM (8d 0h 0m)
Dear NRAO,
I have recently updated my OS installation, and since on, my AIPS (DEC16 "ready-to-use" version acquired from the NRAO AIPS website) is behaving oddly when trying to deconvolve UV sets with IMAGR. There are three behaviours I have seen:
1) Beam & Image are NaN only, but IMAGR proceeds with moreover reasonable flux&beam size values, claiming successful ending;
2) Beam is NaN only, and IMAGR stops stating that:
LOCALH> IMAGR1: Task IMAGR (release of 31DEC14) begins
LOCALH> IMAGR1: UVGET: Using flag table version 1 to edit data
LOCALH> IMAGR1: Create HCG92-LO .IMAGR . 1 (UV) on disk 4 cno 548
LOCALH> IMAGR1: Beginning channel 3 through 58 with 8 IFs
LOCALH> IMAGR1: IMACPY: found 17570224. of 17592064. possible channels
LOCALH> IMAGR1: IMACPY: average frequency is then 1714.159 MHz
LOCALH> IMAGR1: IMACPY: Copied 39268 visibilities to be imaged
LOCALH> IMAGR1: QINIT: did a GET of 5156 Kwords, OFF 17565988608583
LOCALH> IMAGR1: UVWAIT: begin finding uniform weights
LOCALH> IMAGR1: UVWAIT: Average grid weight 2.325E-02
LOCALH> IMAGR1: UVWAIT: Adding temperance S 4.650E+00
LOCALH> IMAGR1: UVWAIT: begin applying uniform or other weights
LOCALH> IMAGR1: UVWAIT: Sum of weights in 3.080E+02 and out 3.080E+02
LOCALH> IMAGR1: UVWAIT: Noise is increased by a factor 1.000 due to weighting
LOCALH> IMAGR1: UVWAIT: Average summed weight 2.325E-02 over 17570224. vis
LOCALH> IMAGR1: Create HCG92-LO .IBM001. 2 (MA) on disk 4 cno 551
LOCALH> IMAGR1: Create HCG92-LO .ICL001. 16 (MA) on disk 4 cno 552
LOCALH> IMAGR1: QINIT: did a FREE of 5156 Kwords, OFF 17565988608583
LOCALH> IMAGR1: QINIT: did a GET of 5310 Kwords, OFF 17565988450887
LOCALH> IMAGR1: GRDFLT: X and Y convolution type = SPHEROIDAL
LOCALH> IMAGR1: GRDFLT: X and Y parms = 3.0000 1.0000
LOCALH> IMAGR1: GRDFLT: convolution function sampled every 1/100 of a cell
LOCALH> IMAGR1: GRDMEM: Ave 448 Channels; 1.659688E+09 to 1.768500E+09 Hz
LOCALH> IMAGR1: Field 1 Sum of gridding weights = 3.26943E+03
LOCALH> IMAGR1: Field 1 Beam min = NaN Jy, max = NaN Jy
LOCALH> IMAGR1: Fit Gaussian FWHM = 21.551 x 11.962 arcsec, PA= -51.7
LOCALH> IMAGR1: CLBHIS: minimum component 0.500 of current peak
LOCALH> IMAGR1: QINIT: did a FREE of 5310 Kwords, OFF 17565988450887
LOCALH> IMAGR1: QINIT: did a GET of 34806 Kwords, OFF 17565958246983
LOCALH> IMAGR1: Field 1 min = -5.0 MilliJy,max = 81.1 MilliJy
LOCALH> IMAGR1: BGC Clean: using 2035 cell beam + residuals > 81.06 MilliJy
LOCALH> IMAGR1: 0 Residual map points loaded
LOCALH> IMAGR1: CLMPAC: NO RESIDUAL POINTS LOADED INTO AP, I QUIT
LOCALH> IMAGR1: CLMPAC: ERROR LOADING RESIDUALS FOR CLEAN field number 0001
LOCALH> IMAGR1: CLNCYC: ERROR CLEANING CLEAN process object
LOCALH> IMAGR1: CLNUV1 : ERROR CLEANING CLEAN process object
LOCALH> IMAGR1: CLNUV : ERROR CLEANING CLEAN process object
LOCALH> IMAGR1: Deleting UV work file:
LOCALH> IMAGR1: Destroyed 1 extension files of type AN
LOCALH> IMAGR1: Destroyed 1 extension files of type FQ
LOCALH> IMAGR1: Destroyed UV image file: catno= 548 disk= 4
LOCALH> IMAGR1: Purports to die of UNNATURAL causes
LOCALH> IMAGR1: localhos 31DEC14 TST: Cpu= 11.4 Real= 13 IO= 2761
3) Similar to (2), but a malformed beam image is produced (eg. there are multiple ripples arranged in a vertical line, like it tried to put beams for all of the IFs/channels into a straight line).
If I set NITER=0, then a non-cleaned image looks ok, but the beam image is still wrong. The outcome of cleaning is always the same for a given set i.e. re-running IMAGR yields same results.
Also, I tried cleaning different sets (WSRT, AIPS, GMRT) to see if this is related just to a certain one telescope - unfortunately not.
Testing on different machines and AIPS versions shows that the problem happens only on my system. It was not present before the upgrade.
Memory banks seem to be ok, judging on the load test, and the hard drive should also be.
Please let me know, if you need more specific information, and what kind of diagnostic input would be neccessary.
Looking forward to hearing from you!
Błażej
------------------------------------------------------
Staff CP: https://help.nrao.edu/staff
More information about the Daip
mailing list