[gradsusr] How to get around gxout grid dying after X t's?

[Jennifer Adams]

Hi, Chris — 
You are using re(), so I can’t speak for or test any memory leaks that might be in that code. I can duplicate what you are doing with re() using lterp() instead, and I don’t get any seg faults. Also, you have a ‘set font’ command in there, and there is a bug that was fixed in version 2.1.a3 that was using up memory because the the font files were being re-initialized every time a character was drawn. Try using the latest version of GrADS and lterp() and see if you are still getting a seg fault. 

I created this dummy descriptor file (for a 4-degree grid) and opened is as a second data file, alongside the GFS output:
dset ^foo.bin
options template
undef -9.99e8
xdef 90 linear 2 4
ydef 45 linear -88 4
tdef 1 linear 01Jan0001 1dy
zdef 1 linear 1 1
vars 1
foo 0 99 foo
endvars

And then I tried your expression with lterp:

ga-> var=lterp(500/(60-((tmpprs(lev=500)-273.15 + tmpprs(lev=1000)-273.15 + tmp2m-273.15)/2)*-3),lat.2(t=1))

And I got a plot (grfill and grid on top) without any error messages or ballooning memory size for all time steps in the model run. 
—Jennifer

On Oct 22, 2015, at 7:38 AM, Christopher Gilroy <chris.gilroy at gmail.com> wrote:

> 'define zcomp = 500/(60-((tmpprs(lev=500)-273.15 + tmpprs(lev=1000)-273.15 + tmp2m-273.15)/2)*-3)'
> 
> 'set gxout grid'
> 'set gridln off'
> 'set dignum 1'
> 'set font 15'
> 'set digsiz 0.10'
> 'set lat 18 62'
> 'set lon -130 -58'
> 
> 'd re(zcomp,4)'
> 
> 
> Makes it to image 25/hour 138 of a loop before grads dies.
> 
> I can plot that exact same var shaded without issue, but if I wanted to draw it shaded AND gridded, you could see the problem.
> 
> 
> 
> On Thu, Oct 22, 2015 at 2:32 AM, James T. Potemra <jimp at hawaii.edu> wrote:
> How are you plotting?  Could it be you have either a corrupt value, or a really large number somehow in the mix?
> 
> On 10/21/15 6:16 PM, Christopher Gilroy wrote:
>> So I'm plotting a variable and after about 25 t's (and that's with re(var, 4) even) of a loop grads dies on me with memory allocation. It's not the server. I've tried reinit, clear (which I use at the end of the loop to obviously clear the previous displays) undef, but I can't seem to figure out why it doesn't like running like that.
>> 
>> Secondly, whenever I get around to it, the internal memory limit is changeable in the source, correct? 
>> 
>> 
>> _______________________________________________
>> gradsusr mailing list
>> gradsusr at gradsusr.org
>> http://gradsusr.org/mailman/listinfo/gradsusr
> 
> 
> _______________________________________________
> gradsusr mailing list
> gradsusr at gradsusr.org
> http://gradsusr.org/mailman/listinfo/gradsusr
> 
> 
> 
> 
> -- 
> -Chris A. Gilroy
> _______________________________________________
> gradsusr mailing list
> gradsusr at gradsusr.org
> http://gradsusr.org/mailman/listinfo/gradsusr

--
Jennifer M. Adams
Center for Ocean-Land-Atmosphere Studies (COLA)
111 Research Hall, Mail Stop 2B3
George Mason University
4400 University Drive
Fairfax, VA 22030 





-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://gradsusr.org/pipermail/gradsusr/attachments/20151022/c822d7b6/attachment.html