Up to top level
AO15   AO16   AO17   AO18   AO19   Backgrounds   Calibration   Conference   Data   Docs   EPICMOS   EPICpn   Feedback   Gallery   Misc   OM   Pending   PhD_Theses   Publications   RGS   RadMonitor   SAS_Hardware   SAS_WS   SASv16.0   SASv16.0_Installation   SASv16.1   SASv16.1_Installation   SASv17.0   SASv17.0_Installation   SASv18.0   SASv18.0_Installation   SciSim   Simulators_other   Suggestions   Trash   Visibility   XMM-bouncing   XMM-news   XRPS   XSA   esas   incoming  

Logged in as guest

Viewing esas/43944
Full headers

From: gaetz@cfa.harvard.edu
Subject: SASv10.0, mos_back problem
Compose reply
Download message
Move To:
5 replies: 1 2 3 4 5
2 followups: 1 2

Private message: yes  no

Notes:

Notification:


Date: Mon, 17 May 2010 14:04:15 GMT
From: gaetz@cfa.harvard.edu
To: xmmhelp@sciops.esa.int
Cc: gaetz@cfa.harvard.edu
Subject: SASv10.0, mos_back problem
Full_Name: Terry Gaetz
Submission from: (NULL) (131.142.42.182)


In running mos_back, I consistently get an error:

  mos1S023-im1-400-2000.fits                                                    
                                                     
  RETFIT POSITION 1
  STATUS FTGIOU = 9367872
  Reading input file
  mos1S023-im1-400-2000.fits
  Input array size not equal to image

I'm confused as to what is happening here.
The file exists and has reasonable dimensions:
  XMM-SAS:lhea:geri% funhead mos1S023-im1-400-2000.fits | grep NAXIS
  NAXIS   =                    2 / number of data axes                          
 
  NAXIS1  =                  780 / length of data axis 1                        
 
  NAXIS2  =                  780 / length of data axis 2                        
 

I'm using the current HEASOFT 6.9 release, but the same thing happens

Looking at the source, the error is coming from 
"retfit" in esas/subs.f; retfit is called two
places in mos_back_sub.f, once for the fwc data,
and once for the observation image in detector coordinates.
There is an appropriate -obj-im-det-400-2000, so I'
SAS:lhea:geri% funhead mos1S023-obj-im-det-400-2000.fits | grep
NAXIS
NAXIS   =                    2 / number of data axes                           

NAXIS1  =                  780 / length of data axis 1                         

NAXIS2  =                  780 / length of data axis 2                         



XMM-SAS:lhea:geri% sasversion 
sasversion:- XMM-Newton SAS release and build information:

SAS release: xmmsas_20100423_1801-10.0.0
Compiled on: Fri Apr 23 18:17:14 GMT 2010
Compiled by: sasbuild@sasbld02.net4.lan
Platform   : RHEL5 32

SAS-related environment variables that are set:

SAS_DIR = /home/gaetz/local/pkgs/x86_64-linux_CentOS-5.4/xmmsas
SAS_PATH = /home/gaetz/local/pkgs/x86_64-linux_CentOS-5.4/xmmsas
SAS_CCFPATH = /home/gaetz/data/caldb/data/xmm/ccf
SAS_CCF = /data/geri3/gaetz/data.xmm.n132d/0129341501/odf/ccf.cif
SAS_ODF = /data/geri3/gaetz/data.xmm.n132d/0129341501/odf


I'm using the current HEASOFT 6.9-release, but the same thing happens
with HEASOFT-6.8


Reply 1

Resend
From: Pedro M. Rodriguez Pascual <xmmhelp@sciops.esa.int>
To: gaetz@cfa.harvard.edu
Subject: Re: SASv10.0, mos_back problem (PR#43944)
Date: Tue May 18 09:03:55 2010
Hello Terry, 

Your question has been forwarded to SAS developers for analysis. We will come
back with the results of their analysis as soon as it is available.

Sincerely,

Pedro



Reply 2

Resend
From: Pedro M. Rodriguez Pascual <xmmhelp@sciops.esa.int>
To: gaetz@cfa.harvard.edu
Subject: Re: SASv10.0, mos_back problem (PR#43944)
Date: Mon May 24 11:29:55 2010
Terry,

According to developers, the problem is related to parameters initialization in
calls to fitsio library; Linux is not affected, but Macs are. It will be fixed
in the next release.

Sincerely,

Pedro



Followup 1

Compose reply
Download message
Date: Sat, 29 May 2010 14:04:23 -0400
From: Terry Gaetz <gaetz@cfa.harvard.edu>
To: "Pedro M. Rodriguez Pascual" <xmmhelp@sciops.esa.int>
Cc: Terry Gaetz <gaetz@cfa.harvard.edu>
Subject: Re: SASv10.0, mos_back problem (PR#43944)
Pedro M. Rodriguez Pascual wrote:
> Terry,
> 
> According to developers, the problem is related to parameters
initialization in
> calls to fitsio library; Linux is not affected, but Macs are. It will be
fixed
> in the next release.

Hi Pedro.

NOTE: The problem *DOES* affect linux - I was in fact reporting the
problem for linux boxes (x86_64) running CentOS 5.4 (& 5.5),
(and CentOS is a repackaging of RHEL 5.4 (and 5.5)).

I haven't actually tried XMMSAS or XMMESAS on the intel
Mac (Snow Leopard, 10.6.3) yet, since my main machines are the
*linux* boxes running CentOS.

When will a patch be available?  Since all my observations
are of diffuse objects, I really need the xmmesas software running.
Now that it is integrated into SAS, I no longer have the option
of debugging and recompiling my own version...

Thanks.

Terry

> Sincerely,
> 
> Pedro
> 
> 
> 
> ================================================================================================
> This message and any attachments are intended for the use of the addressee
or addressees only.
> 
> The unauthorised disclosure, use, dissemination or copying (either in whole
or in part) of its content is prohibited.
> 
> If you received this message in error, please delete it from your system
and notify the sender.
> 
> E-mails can be altered and their integrity cannot be guaranteed.
> 
> ESA shall not be liable for any e-mail if modified.
> =================================================================================================



Reply 3

Resend
From: Pedro M. Rodriguez Pascual <xmmhelp@sciops.esa.int>
To: gaetz@cfa.harvard.edu
Subject: Re: SASv10.0, mos_back problem (PR#43944)
Date: Tue Jun  8 12:01:30 2010
Hello Terry,

My apologies for this late reply, but it has taken me a while to run esas tasks
on the 0129341501 data-set.

Unfortunately, it is no easy for me to test exactly the same configuration you
are using: SAS built on a 32 bits running on a 64 bits computer. However, from
the previous analysis, the fact that the tasks run on 32bits machines and the
status value returned by the cfitsio function ftgiou in your email, the
initialization of the fitsio library seems very likely to be in the origin of
the problem. 

SAS developers at ESAC are currently testing a patch provided by external
developers. Upon success of these tests, the patch will be made public, but I
cannot give now an estimate other than "some weeks" on when it will happen.

Sorry,

Pedro



Followup 2

Compose reply
Download message
Date: Tue, 15 Jun 2010 23:38:36 -0400
From: Terry Gaetz <gaetz@cfa.harvard.edu>
To: "Pedro M. Rodriguez Pascual" <xmmhelp@sciops.esa.int>
Cc: Terry Gaetz <gaetz@cfa.harvard.edu>
Subject: Re: SASv10.0, mos_back problem (PR#43944)
Pedro M. Rodriguez Pascual wrote:
> Hello Terry, 
> 
> Your question has been forwarded to SAS developers for analysis. We will
come
> back with the results of their analysis as soon as it is available.
> 
> Sincerely,
> 
> Pedro

Hello Pedro

I am sorry for the delay in replying -- a family matter has kept me
out of touch.

The patched version fixes the previous problem, but now I run into
another issue.  In processing the corner data, the exposure-weighted
count rate ends up with NaN's.  This propagates through the
processing, and the task dies in POLY_FIT, with a floating point error.

The problem seems to start while processing mos1-qpb.fits.gz:

   /home/gaetz/data/caldb/data/xmm/esas_ccf/mos1-qpb.fits.gz
    Rate:    3.1818010E-02  Rate error:  NaN
    Hard:    3.7560976  Hard error:  NaN

The end of the log looks like:

  Reading input file
  mos1S001-im5-400-2000.fits
  Average non-zero value of FWC image:    1.4981754
  FWC image scaling for chip:  5 NaN   3.8590000E+04 NaN
  mos1S001-im7-400-2000.fits 

  Reading input file
  mos1S001-im7-400-2000.fits
  Average non-zero value of FWC image:    1.4894192
  FWC image scaling for chip:  7 NaN   3.7945492E+04 NaN
  Reading input file
  mos1S001-obj-im-det-400-2000.fits
  Total observation counts in band    8.0534000E+04
  Total background counts in band  NaN
  DGAUSJ, singular matrix, pos 2
  MQM ==> Bad return from DGAUSJ
  Bad return from MQM
  Program stopped in POLY_FIT
  Position 01

I can send the rest of the log if you need it.  It showed up in my test
case (the observation id I sent before), and also the science case
I'm working on.  Both cases died the same way, suggesting that there
is a deeper problem.

Please forward this to the appropriate people, and let me know if
further information is needed.

Thank you.

Terry



Reply 4

Resend
From: Pedro M. Rodriguez Pascual <xmmhelp@sciops.esa.int>
To: gaetz@cfa.harvard.edu
Subject: Re: SASv10.0, mos_back problem (PR#43944)
Date: Tue Jun 22 08:05:53 2010
Hello Terry,

sorry for this long silence. 

Your question was immediately forwarded to SAS developers and I have been
waiting for their comments to reply. Unfortunately, I have not got  them yet and
this message is just to let you know that your question has not been forgotten.

with our apologies, 

Pedro


Reply 5

Resend
From: Pedro M. Rodriguez Pascual <xmmhelp@sciops.esa.int>
To: gaetz@cfa.harvard.edu
Subject: Re: SASv10.0, mos_back problem (PR#43944)
Date: Fri Jul  2 09:17:45 2010
Hello Terry,

SAS developers do not reproduce the last problem reported, on obsid 0129341501.
Would it be possible you send to us the complete log of mos-back when run on
0129341501 ? Alternatively, the complete log and obsid corresponding to your
last message (Jun 15) would be fine.

Thanks, 

Pedro

Up to top level
AO15   AO16   AO17   AO18   AO19   Backgrounds   Calibration   Conference   Data   Docs   EPICMOS   EPICpn   Feedback   Gallery   Misc   OM   Pending   PhD_Theses   Publications   RGS   RadMonitor   SAS_Hardware   SAS_WS   SASv16.0   SASv16.0_Installation   SASv16.1   SASv16.1_Installation   SASv17.0   SASv17.0_Installation   SASv18.0   SASv18.0_Installation   SciSim   Simulators_other   Suggestions   Trash   Visibility   XMM-bouncing   XMM-news   XRPS   XSA   esas   incoming  

Logged in as guest


Please make your (short) question the subject of your request!


Web interface using JitterBug ... back to the XMM home page