Error in fxcor
Emiliano Jofre wrote on Nov 08, 2009
Hi,
I have a problem using fxcor. if I try to perform a cross-correlation with the same or two different spectra, in some of them works fine but in others I get the message:
PANIC in `/iraf/iraf/noao/bin.linux/x_rv.e': Salloc underflow
ERROR: Abnormal termination of child process 'noaobin$x_rv.e'
I did a flpr, but doesn't work and When I run fxcor in a new iraf directory such as /tmp works fine. I was running a large application at the same time, so maybe is a memory problem, but now, if I do not run anything else the problem persist.
Anyone can tell me what can I do?
Thanks
Emiliano.
I have a problem using fxcor. if I try to perform a cross-correlation with the same or two different spectra, in some of them works fine but in others I get the message:
PANIC in `/iraf/iraf/noao/bin.linux/x_rv.e': Salloc underflow
ERROR: Abnormal termination of child process 'noaobin$x_rv.e'
I did a flpr, but doesn't work and When I run fxcor in a new iraf directory such as /tmp works fine. I was running a large application at the same time, so maybe is a memory problem, but now, if I do not run anything else the problem persist.
Anyone can tell me what can I do?
Thanks
Emiliano.
Mike Fitzpatrick wrote on Nov 08, 2009
Something like a 'salloc underflow' is usually a coding problem, but it can be caused by a specific comination of parameters or in some cases the data characteristics. There was an old problem (v2.10 or so) that sounded similar, but nothing currently I'm aware of. Could you post the images and 'dpar' listings of your fxcor/pset parameters to ftp://iraf.noao.edu/pub so I can try to reproduce it?
It would also help to know your IRAF and OS version.
-Mike
It would also help to know your IRAF and OS version.
-Mike
Emiliano Jofre wrote on Nov 08, 2009
Hi Mike,
I am using IRAF V2.13 in ubuntu 7.10. My fxcor parameters are:
objects = "doprntcdfbhr8232.fits" List of object spectra
templates = "doprntcdfbhr8232.fits" List of template spectra
(apertures = "*") Apertures to be used
(cursor = "") Graphics input cursor\n
(continuum = "both") Continuum subtract spectra?
(filter = "none") Fourier filter the spectra?
(rebin = "smallest") Rebin to which dispersion?
(pixcorr = no) Do a pixel-only correlation?
(osample = "4520-4800,4900-5100,5200-5250,4129.27-4329.65,4366-4515") Obje
(rsample = "4520-4800,4900-5100,5200-5250,4129.27-4329.65,4366-4515") Temp
(apodize = 0.1) Apodize end percentage\n
(function = "gaussian") Function to fit correlation
(width = INDEF) Width of fitting region in pixels
(height = 0.) Starting height of fit
(peak = no) Is height relative to ccf peak?
(minwidth = 10.) Minimum width for fit
(maxwidth = 100.) Maximum width for fit
(weights = 1.) Power defining fitting weights
(background = 0.) Background level for fit
(window = 400.) Size of window in the correlation plot
(wincenter = INDEF) Center of peak search window\n
(output = "pruu") Root spool filename for output
(verbose = "long") Verbose output to spool file?
(imupdate = no) Update the image header?
(graphics = "stdgraph") Graphics output device\n
(interactive = yes) Interactive graphics?
(autowrite = yes) Automatically record results?
(autodraw = yes) Automatically redraw fit results?
(ccftype = "image") Output type of ccf\n
(observatory = "casleo") Observation location database
(continpars = "") Continuum processing parameters
(filtpars = "") Filter parameters pset
(keywpars = "") Header keyword translation pset\n
(mode = "ql")
I think the problem are not the images or units, because the same images worked ok last week, but now fxcor does not work fine in my default login.cl directory but it does in a new iraf directory such as /tmp.
Let me know if you can figure out the problem.
Thanks again
Emiliano.
I am using IRAF V2.13 in ubuntu 7.10. My fxcor parameters are:
objects = "doprntcdfbhr8232.fits" List of object spectra
templates = "doprntcdfbhr8232.fits" List of template spectra
(apertures = "*") Apertures to be used
(cursor = "") Graphics input cursor\n
(continuum = "both") Continuum subtract spectra?
(filter = "none") Fourier filter the spectra?
(rebin = "smallest") Rebin to which dispersion?
(pixcorr = no) Do a pixel-only correlation?
(osample = "4520-4800,4900-5100,5200-5250,4129.27-4329.65,4366-4515") Obje
(rsample = "4520-4800,4900-5100,5200-5250,4129.27-4329.65,4366-4515") Temp
(apodize = 0.1) Apodize end percentage\n
(function = "gaussian") Function to fit correlation
(width = INDEF) Width of fitting region in pixels
(height = 0.) Starting height of fit
(peak = no) Is height relative to ccf peak?
(minwidth = 10.) Minimum width for fit
(maxwidth = 100.) Maximum width for fit
(weights = 1.) Power defining fitting weights
(background = 0.) Background level for fit
(window = 400.) Size of window in the correlation plot
(wincenter = INDEF) Center of peak search window\n
(output = "pruu") Root spool filename for output
(verbose = "long") Verbose output to spool file?
(imupdate = no) Update the image header?
(graphics = "stdgraph") Graphics output device\n
(interactive = yes) Interactive graphics?
(autowrite = yes) Automatically record results?
(autodraw = yes) Automatically redraw fit results?
(ccftype = "image") Output type of ccf\n
(observatory = "casleo") Observation location database
(continpars = "") Continuum processing parameters
(filtpars = "") Filter parameters pset
(keywpars = "") Header keyword translation pset\n
(mode = "ql")
I think the problem are not the images or units, because the same images worked ok last week, but now fxcor does not work fine in my default login.cl directory but it does in a new iraf directory such as /tmp.
Let me know if you can figure out the problem.
Thanks again
Emiliano.
Mike Fitzpatrick wrote on Nov 08, 2009
I can't figure it out in detail, but if it was working before and you haven't changed fxcor, then something else changed in the last week. You can do a MKIRAF to reinitialize your parameters, or else do an "unlearn" on each pset and task in turn and try resetting the parameters until you narrow down which parameters might be causing problems.
I would first try unlearning the CONTINPARS pset or disabling the continuum fitting in the task, otherwise have a look at whether specifying the osample/rsample is causing a problem.
-Mike
I would first try unlearning the CONTINPARS pset or disabling the continuum fitting in the task, otherwise have a look at whether specifying the osample/rsample is causing a problem.
-Mike
Last post on Nov 08, 2009