[daip] Are you still available for AIPS support?

[Eric Greisen]

On 01/05/2016 09:02 PM, Ross Burns wrote:
> Dear Eric
>
> Its great to know that AIPS is still in good hands! Are you working at
> Socorro now? It'd be great to meet you in person sometime - though i
> have no immediate plans to go to the US, i'm likely to attend some
> conference/collaboration at some point.
>
> I have an issue in AIPS that I'd like to ask you about.
> Yesterday, I installed 31DEC15 (I usually install the most latest
> 'frozen' version), as an upgrade from 31DEC12 which I had been using for
> a while. Im running on mac.
>
> The first thing I did after installing was to run midnight_job, just to
> make sure i got all of any patches (im not sure if these come with the
> download tarball).
>
> Im reducing 16xIF VLBI data from the Korean KVN.
> I was running UVCOP, to split a selection of IFs when the task failed.
> Here are my inputs and the error message:
>
> *************************** Inputs:
> *******
> *UVCOP
> *******
> tget uvcop
> default
> indisk 3; outdisk 3; getn 1;
>
> Bif 1; Eif 8;
> outclass 'K-band'
> go; wait; pcat
>
>
> *************************** Message:
>
> LOCALH> UVCOP1: Task UVCOP  (release of 31DEC15) begins
> LOCALH> UVCOP1: Create KVN         .K-BAND.   2 (UV)  on disk  3  cno    3
> LOCALH> UVCOP1: No flag table is applied this time
> LOCALH> UVCOP1: Copied from the beginning
> LOCALH> UVCOP1:         to  the end
> LOCALH> UVCOP1: Copied    78255. vis records
> LOCALH> UVCOP1: Dropped       0. flagged vis records
> LOCALH> UVCOP1: Copied AT file from vol/cno/vers  3    1   1 to  3    3   1
> LOCALH> UVCOP1: Copied CT file from vol/cno/vers  3    1   1 to  3    3   1
> LOCALH> UVCOP1: Updating tables for IF/FREQID/channel selection
> LOCALH> UVCOP1: Reformatted AN file from vol/cno/vers  3    1   1 to  3
>     3   1
> LOCALH> UVCOP1: Reformatted CL file from vol/cno/vers  3    1   1 to  3
>     3   1
> LOCALH> UVCOP1: Reformatted CL file from vol/cno/vers  3    1   2 to  3
>     3   2
> LOCALH> UVCOP1: Reformatted CL file from vol/cno/vers  3    1   3 to  3
>     3   3
> LOCALH> UVCOP1: Reformatted CL file from vol/cno/vers  3    1   4 to  3
>     3   4
> LOCALH> UVCOP1: Reformatted CL file from vol/cno/vers  3    1   5 to  3
>     3   5
> LOCALH> UVCOP1: Reformatted CL file from vol/cno/vers  3    1   6 to  3
>     3   6
> LOCALH> UVCOP1: Reformatted CL file from vol/cno/vers  3    1   7 to  3
>     3   7
> LOCALH> UVCOP1: Reformatted CQ file from vol/cno/vers  3    1   1 to  3
>     3   1
> LOCALH> UVCOP1: GCINI: NTABGC=   200 EXCEEDS MAXIMUM ALLOWED
> LOCALH> UVCOP1: GCINI: ERROR INITIALIZING GAIN CURVE TABLE FOR WRIT
> LOCALH> UVCOP1: GCSEL: ERROR   1 INITING NEW TABLE
> LOCALH> UVCOP1: Destroyed  1 extension files of type CQ
> LOCALH> UVCOP1: Destroyed  7 extension files of type CL
> LOCALH> UVCOP1: Destroyed  1 extension files of type AN
> LOCALH> UVCOP1: Destroyed  1 extension files of type CT
> LOCALH> UVCOP1: Destroyed  1 extension files of type AT
> LOCALH> UVCOP1: Destroyed  1 extension files of type HI
> LOCALH> UVCOP1: Destroyed  1 extension files of type NX
> LOCALH> UVCOP1: Destroyed  1 extension files of type FQ
> LOCALH> UVCOP1: Destroyed UV image file: catno=      3 disk= 3
> LOCALH> UVCOP1: Purports to die of UNNATURAL causes
> LOCALH> UVCOP1: localhos 31DEC15 TST: Cpu=     31.5  Real=     71  IO=
>     1888
>
>
>
>
>
>
>
> To give a bit of background, this task is part of a script which had
> been working fine on my 31DEC12.
> Furthermore, i tried to circumvent the problem by using SPLAT instead of
> UVCOP, but had the same error message about GCINI and GCSEL.
>
> To isolate the issue I tried running UVCOP on some fresh data and also
> on some data which i had been working on previously using 31DEC12 (saved
> on an AIPS disk held on an external hard drive [indisk 3 in my inputs]
> ). Both gave the same message, which suggests that the problem arises in
> UVCOP (and SPLAT) rather than in the way the data was FTLDed or MSORTed
> or any calibration done before UVCOP.
>
>
>
>
> Do you have any idea what might be causing this issue with UVCOP (and
> SPLAT)? In all honesty, i dont know what GCINI and GCSEL are - i guess
> they are parameters found under the bonnet which most users never
> encounter - unless theres a problem.

We changed the processing of GC (gain curve) tables when I found what 
appeared to be an absurd parametrization.  The new limit on the gain and 
sensitivity values is 200 per IF which still seems to me very excessive. 
  It appears that the GC tables from your data exceed this.

Can you run PRTAB on one of your existing GC tables and send me the 
header and the output for the first couple of rows?

Eric Greisen