[gate-users] Maximum file size exceeded problem

Giovanni Santin giovanni.santin at cern.ch
Sun Sep 19 09:38:40 CEST 2004


Dear Amarnath

I was not aware of any solution to this problem via compilation flags.
It sounds interesting. 

Nevertheless, in the last days I wrote a few modifications to Gate to have 
an automatic file swap when the file size exceeds a user defined limit. 

The old file fileCoincidences.dat is then closed and a new one 
fileCoincidences_1.dat is opened.

This should work for both Coincidences and Singles.

The file size limit is set to 2GB by default, and the user can set it to a 
different value, or to zero (in this case no file swap is done).

In addition, there will be the possibility to choose the variables to 
include in Coincidences and Singles ASCII files, with a mask.

I hope the new additions will come out soon in the next release, but we
still need to wait for some validation feedback as usual before releasing 
the code.

If the release will wait, I'll make the single modified files available 
for testing.

Best regards,
-- 
Giovanni
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Giovanni Santin
Space Environments and Effects Analysis
ESA - ESTEC - TEC-EES
Keplerlaan 1, 2200 AG Noordwijk, Netherlands
e-mail: giovanni.santin at esa.int
Tel: +31 71 565 6267
Fax: +31 71 565 4999


On Thu, 16 Sep 2004, Amarnath Srinivasan wrote:

> Hello Zhang and Others,
> I am using Gate in Redhat 9  for simulating a PET scanner and produce output
> in Ascii Format. When the output file exceeds 2Gb I get an error saying
> "FileSize Limit Exceeded" and the program terminates. I am hitting the 2**31
> bit limit of the datatype off_t.
> 
> 1.   I am looking for ways to fix this problem.Based on Mr.Zhang's
> provious mail I looked for large file support in Redhat Linux. RH9 supports
> large files. But to enable that for gate,  gate should be compiled in gcc
> with the "large file support turned on". I would like to know if gate
> supports
> large files? What is the procedure to compile gate with large file support
> turned on? Has anyone done this before?
> 
> 2.  Will doing "gcc -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64  program.c"
> by editing the make file for gate help? If yes what are the files that
> should be edited ?
>  Does this have to be done for geant too?
> 
> 3. I would also like to know if someone has produced files larger than 2GB
> in gate in "any" platform?
> 
> Thankyou
> Amarnath Srinivasan
> Research Assistant
> Center for Advanced Imaging
> PO Box 9235
> Dept of Radiology, HSC, WVU
> Morgantown, WV 26506
> asrinivasan at hsc.wvu.edu
> Phone Office: 304 293 6925 /Lab: 304 293 5586
> ----- Original Message ----- 
> From: "Long ZHANG" <zhanglong99 at tsinghua.org.cn>
> To: <gate-users at lphe1pet1.epfl.ch>
> Sent: Friday, August 27, 2004 2:46 AM
> Subject: Re: [gate-users] Maximum file size exceeded problem
> 
> 
> > Dear Srinivasan,
> >
> > There are several limits on Linux Filesystem, the one you met is defined
> by VFS
> > module, that is to say, all filesystemson your kerenl have a maximus size
> of 2GB.
> >
> >
> > So the solution is to update your kernel and use another filesystem. A
> journal
> > filesystem can resolve this problem (EXT2/EXT3 can't). You'd better query
> Linux
> > Journal Filesystem on GOOGLE, then you can get many good solution.
> >
> > Hope this helps!
> >
> > Best regards,
> >
> > Long
> >
> > In your mail:
> > >From:"Amarnath Srinivasan" <asrinivasan at hsc.wvu.edu>
> > >Reply-To: GATE feedback and helpline for Users
> <gate-users at lphe1pet1.epfl.ch>
> > >To: "GATE feedback and helpline for Users" <gate-users at lphe1pet1.epfl.ch>
> > >Subject: Re: [gate-users] Maximum file size exceeded problem
> > >
> > >
> > >Hello,
> > >
> > >We were running gate on RedHat 9 . When the output ascii file
> > "gate_coincidence.dat" size  exceeded 2GB I got an error saying maximum
> file size
> > exceeded and the program terminated.
> > >Does anybody have idea on how to fix this problem.
> > >
> > >
> > >
> > >thankyou
> > >Amarnath Srinivasan
> > >Research Assistant
> > >Center for Advanced Imaging
> > >PO Box 9235
> > >Dept of Radiology, HSC, WVU
> > >Morgantown, WV 26506
> > >asrinivasan at hsc.wvu.edu
> > >  ----- Original Message ----- 
> > >  From: yuxuan.zhang at di.mdacc.tmc.edu
> > >  To: gate-users at lphe1pet1.epfl.ch
> > >  Sent: Tuesday, August 24, 2004 5:08 PM
> > >  Subject: [gate-users] about random coincidences
> > >
> > >
> > >
> > >  hi all,
> > >
> > >  In my simulation I noticed a strange result that I can not figure it
> out.
> > >  My setup is that: using cylindricalPET system with 30 rsetors, one
> point
> > >  source at the center. The radiation activity is very high, thus there
> are
> > >  lots of random coincidences (accidental coincidences) in the totoal
> > >  coincidences. Now I do the statistics on these random coincidences
> according
> > >  to the sector index, then I found that the coincidences between 2
> opposite
> > rsectors
> > >  (e.g. No.0 and 15, or No.5 and 20) are much more than the coincidences
> between
> > other
> > >  2 sectors ( No. 0 and 14, or No. 0 and 16 ...) Usually the difference
> can be
> > 50% or 100% depend on the thickness of the detector. Is anyone else
> noticed that?
> >
> > >  Can you have a try in your computer to find out whether this is a bug
> of GATE
> > or there
> > >  is somthing wrong with my simulation? Thank you very much.
> > >
> > >
> > >  Best regards,
> > >
> > >  Yuxuan Zhang
> > >
> > >
> > >  Dr. Yuxuan ZHANG
> > >  Dept. Experimental Diagnostic Imaging
> > >  Univ. Texas, MD Anderson Cancer Center
> > >  1515 Holcombe Blvd, Unit 217
> > >  Houston, TX 77030-4095
> > >
> > >  Tel: +1-713-745-1671
> > >  Fax: +1-713-745-1672
> > >
> > >
> >
> >---------------------------------------------------------------------------
> ---
> > >
> > >
> > >  _______________________________________________
> > >  gate-users mailing list
> > >  gate-users at lphe1pet1.epfl.ch
> > >  http://lphe1pet1.epfl.ch/mailman/listinfo/gate-users
> > >_______________________________________________
> > >gate-users mailing list
> > >gate-users at lphe1pet1.epfl.ch
> > >http://lphe1pet1.epfl.ch/mailman/listinfo/gate-users
> > >
> >
> >
> > _______________________________________________
> > gate-users mailing list
> > gate-users at lphe1pet1.epfl.ch
> > http://lphe1pet1.epfl.ch/mailman/listinfo/gate-users
> 
> _______________________________________________
> gate-users mailing list
> gate-users at lphe1pet1.epfl.ch
> http://lphe1pet1.epfl.ch/mailman/listinfo/gate-users
> 


More information about the Gate-users mailing list