[daip] Problems with ANTAB/NX table

Mon Aug 3 10:39:48 EDT 2015

Hi, 

I was advised by Arnaud Collioud to address this AIPS question to you as he encountered similar problem before and received good help. 

I am attempting to image a large geodetic experiment (CONT14), and running into a problem while trying to import a hand-made ANTAB table. I cannot proceed with this table and even check if it is ok because every time I try to import it I run into the following error: 
>
> ANTAB1: Task ANTAB  (release of 31DEC14) begins
> ANTAB1: NXREAD: INCREASE PARAMETER MAXNX
> ANTAB1: Purports to die of UNNATURAL causes
> ANTAB1: localhost    31DEC14 TST: Cpu=       0.0  Real=       0
> AIPS 1: Resumes
>
This happens on both 31DEC14 and 31DEC15 installations (I use two different isolated binary installations on different machines). This is a reasonably large chunk of data, first 24h of a 14-days experiment, 16 IFs, 17 antennas. Arnaud encountered similar problem while trying to image some other geodetic session. He suggested that this might occur because my NX table is too large, as maximum length of it is limited. Indeed NX table was over 50000 lines, and I doubt it is possible to reduce it. Idea of Arnaud is that the parameter that defines the maximum length of the index table can be increased, but that is something that I cannot do myself. Could you please help me with this issue? I wasn't able yet to make MNJ working on either of my machines, but can try to set it up if it's the only way to patch it. 

Thanks in advance!

Best regards, 
Liza

Dr Elizaveta Rastorgueva-Foi
  AuScope VLBI Project
  School of Maths & Physics
  University of Tasmania, Australia
  Hobart TAS 7001
  efoi at utas.edu.au
   P: +61 3 6226 2445

________________________________________
From: Elizaveta Rastorgueva-Foi
Sent: Friday, July 31, 2015 7:58 PM
To: Arnaud Collioud
Cc: liza.foi at gmail.com
Subject: Re: CONT14 imaging?

Thanks, Arnaud, for taking care of this in your last day before vacation! :) You are right, NX table is WAY longer, almost 56000 lines. And it is just one day of observations. Don't know what to blame, subarrys, too short average time? But good to know there are wways, I will contact NRAO people about it and will figure out how to implement the solution.

Enjoy your holidays!

When you come back, I won't be in UTas anymore, but I am determined to finish this project. My other email is liza.foi at gmail.com, please add it to your address book not to put it in spam.

Cheers
Liza

Dr Elizaveta Rastorgueva-Foi
  AuScope VLBI Project
  School of Maths & Physics
  University of Tasmania, Australia
  Hobart TAS 7001
  efoi at utas.edu.au
   P: +61 3 6226 2445
________________________________________
From: Arnaud Collioud <Arnaud.Collioud at obs.u-bordeaux1.fr>
Sent: Friday, July 31, 2015 6:25 PM
To: Elizaveta Rastorgueva-Foi
Subject: Re: CONT14 imaging?

Hi Liza,

I am glad to see that you made some progresses on CONT14!

Yes, I encountered this error before. The « problematic » parameter is in the ANTAB.FOR file (see 31DEC14/APL/PGM/NOTST/ANTAB.FOR), which is (apparently) defining a buffer size for the index table (NX).

The current value is:
PARAMETER (MAXNX = 16384)

Probably (just a guess), your NX is too big (more than 16384 lines? You can have a look at it with PRTAB...) to be processed by ANTAB. Did you load all the session at once? Only just one session?

The only solution ( if you do not want to loose data) is to increase the buffer size. For that, 2 solutions:
* you compiled AIPS yourself (which is very unlikely). But if you did, just change the value and recompile the ANTAB.FOR file.
* you got the binary installation (safe and easy):
Ask Eric Greisen and Amy Mioduszewski @ NRAO (daip at nrao.edu) to raise the limit. Their are very reactive (they will tell you if it is possible or not, and they provide an adapted MNJ). If you adopt this solution, working MNJ are required (I do not know exactly how to configure them because this was done by our system administrators at my institute; see http://www.aips.nrao.edu/mnj.shtml or ask Eric & Amy or people at your institute).

Cheers,
Arnaud

PS:  I am in vacation tonight for 4 weeks (yes!). So I will not be reachable during that period.
I hope your problem will be resolved long before I return to the office...

Le 31 juil. 2015 à 05:03, Elizaveta Rastorgueva-Foi <Elizaveta.RastorguevaFoi at utas.edu.au> a écrit :

> Hi Arnaud,
>
> I've been working on imaging CONT14 (remember, it is my first time imaging GEO data, so took a lot of time to learn!), and run into a problem, now need serious help.
>
> CONT14 FITS files turned out to be in a funny frequency arrangement: all 16 IFs/channels together. So to test the ANTAB for this arrangement I tried to input dummy system temperatures just for one antenna, just to see how it would work with 16 IFs. However, every time I try it, I run in the following error:
>
> ANTAB1: Task ANTAB  (release of 31DEC14) begins
> ANTAB1: NXREAD: INCREASE PARAMETER MAXNX
> ANTAB1: Purports to die of UNNATURAL causes
> ANTAB1: localhost    31DEC14 TST: Cpu=       0.0  Real=       0
> AIPS 1: Resumes
>
> Same thing on another machine with 31DEC15 installation, same thing on someone's else Mac installation, same thing with dummy Tsys for all 17 antennas. Google tells that this problem occurred once before, and you was in support for that: https://listmgr.nrao.edu/pipermail/daip/2014-September/020654.html  Allegedly, the problem was cured by a patch in 31DEC14.
>
> My question: How can I get past this error? Do I repeat a mistake with my Tsys table? Or is it AIPS installation problem? I was never able to make mnj running on my laptop and desktop.
>
> Cheers,
> Liza
>
>
>
> Dr Elizaveta Rastorgueva-Foi
>  AuScope VLBI Project
>  School of Maths & Physics
>  University of Tasmania, Australia
>  Hobart TAS 7001
>  efoi at utas.edu.au
>   P: +61 3 6226 2445
> ________________________________________
> From: Arnaud Collioud <Arnaud.Collioud at obs.u-bordeaux1.fr>
> Sent: Thursday, June 11, 2015 10:15 PM
> To: Elizaveta Rastorgueva-Foi
> Subject: Re: CONT14 imaging?
>
> Hi Liza,
>
> I will be very happy to answer (if I can) any question you might have.
> However, I will probably be not very reactive (as you can see with this late email), so you should be sometimes very patient… ;)
>
> Cheers,
> Arnaud
>
>
>
> Le 5 juin 2015 à 10:50, Elizaveta Rastorgueva-Foi <Elizaveta.RastorguevaFoi at utas.edu.au> a écrit :
>
>>
>>
>> Hi Arnaud,
>>
>>
>>    You seem to be very enthusiastic ; this is great!
>>
>> Well, yes, it's a challenge, I have to nail it after all. Having said that, I would be happy to join forces with someone who is more experienced. If you are interested to be involved and have time to answer my questions, I'd appreciate advice and help.
>>
>> Cheers,
>> Liza
>>
>>
>>
>> Dr Elizaveta Rastorgueva-Foi
>>  AuScope VLBI Project
>>  School of Maths & Physics
>>  University of Tasmania, Australia
>>  Hobart TAS 7001
>>  efoi at utas.edu.au
>>   P: +61 3 6226 2445
>> ________________________________________
>> From: Arnaud Collioud <Arnaud.Collioud at obs.u-bordeaux1.fr>
>> Sent: Friday, May 29, 2015 11:14 PM
>> To: Elizaveta Rastorgueva-Foi
>> Subject: Re: CONT14 imaging?
>>
>> Hi again,
>>
>> Thank you for your report. You seem to be very enthusiastic ; this is great!
>>
>> Well, I have no defined plan concerning the imaging. I suppose everything is possible: one image per day (day-to-day comparison), combining days (better u-v coverage), all days (real interest?).
>> Unfortunately, I will be too busy in the next few months with some delayed projects to have a look at CONT14. I will probably recontact you when some of them are completed. Meantime, I hope you will make major progresses…
>>
>> Good luck!
>> Cheers,
>>
>> Arnaud
>>
>>
>> Le 28 mai 2015 à 12:54, Elizaveta Rastorgueva-Foi <Elizaveta.RastorguevaFoi at utas.edu.au> a écrit :
>>
>>> Hi Arnaud,
>>>
>>> I indeed made an attempt to image CONT14 data last year but at that time failed to calibrate it properly since I had no prior experience with calibration of geodetic data, and as well failed to obtain any good advice (even though had sought it). So I had to leave the data for some period, thus there are no any images yet. However recently one of colleagues given me a masterclass in calibration, so I am trying to redo it. Are you interested in imaging all the session, or only one of them? Anyway, I am determined to proceed with imaging since we are running a project employing real source structure for multiple sources, and I want to play with combining several days together as well.
>>>
>>> Cheers,
>>> Liza
>>>
>>>
>>>
>>>
>>> Dr Elizaveta Rastorgueva-Foi
>>> AuScope VLBI Project
>>> School of Maths & Physics
>>> University of Tasmania, Australia
>>> Hobart TAS 7001
>>> efoi at utas.edu.au
>>>  P: +61 3 6226 2445
>>> ________________________________________
>>> From: Arnaud Collioud <Arnaud.Collioud at obs.u-bordeaux1.fr>
>>> Sent: Thursday, May 28, 2015 6:11 PM
>>> To: Elizaveta Rastorgueva-Foi
>>> Subject: CONT14 imaging?
>>>
>>> Hi Liza,
>>>
>>> During the EVGA meeting, I talk with some people about the CONT14 session data, and how useful it would be to image this session. Jim told me that you started to look at it. Could you please tell me what you did so far because I am also interested in imaging it?
>>> I am currently downloading the data from GSFC (it is in progress…), and that’s all I have done.
>>>
>>> Cheers,
>>> Arnaud
>>>
>>
>>
>>
>> University of Tasmania Electronic Communications Policy (December, 2014).
>> This email is confidential, and is for the intended recipient only. Access, disclosure, copying, distribution, or reliance on any of it by anyone outside the intended recipient organisation is prohibited and may be a criminal offence. Please delete if obtained in error and email confirmation to the sender. The views expressed in this email are not necessarily the views of the University of Tasmania, unless clearly intended otherwise.
>
>