If the file you are reading in is not corrupted on the disk it might be
being corrupted in memory. You could run memtest86 on the machine to check
if your memory is working correctly.

Peace,

Matt.

-----Original Message-----
From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On Behalf
Of Lindgren, Kristen, Ann
Sent: Thursday, January 22, 2009 3:54 PM
To: [log in to unmask]
Subject: Re: [FSL] nifti_read_buffer in probtrackx

Ok, here is the error I'm getting. The second of these is repeated
numerous times down the screen.

Errors:++ WARNING
nifti_read_buffer(/Applications/fsl/subjects/739001/data.bedpostX/merged
_f2samples.nii.gz):
data bytes needed = 4784128
data bytes input = 3243742
number missing = 1540386 (set to 0)

Errors:++ WARNING
nifti_read_buffer(/Applications/fsl/subjects/739001/data.bedpostX/merged
_f2samples.nii.gz):
data bytes needed = 4784128
data bytes input = 0
number missing = 4784128 (set to 0)


I tried opening the file that seems to be causing the error
(merged_f2samples.nii.gz) in fslview and it opens without any errors.

I'm currently rerunning bedpostx on this subject to see if that fixes
the problem and I'll also try running a completely different subject to
see if I get this error on them as well. In the meantime, any thoughts?
Thanks!

Kristen

P.S. The RAM was just upgraded to 4GB but it appears to be in working
order...


-----Original Message-----
From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On
Behalf Of Matt Glasser
Sent: Thursday, January 22, 2009 12:34 PM
To: [log in to unmask]
Subject: Re: [FSL] nifti_read_buffer in probtrackx

This indicates that the file is corrupted. You can try opening the file
in
FSLView to see if you still get the error (if so, you will need to rerun
whatever processing step generated the file). If you are getting the
error
randomly, that suggests a serious hardware problem.

Peace,

Matt.

-----Original Message-----
From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On
Behalf
Of Lindgren, Kristen, Ann
Sent: Thursday, January 22, 2009 7:58 AM
To: [log in to unmask]
Subject: [FSL] nifti_read_buffer in probtrackx

Is there any way that one could get this error due to RAM issues? Or
this
kind of error always mean there's something wrong with the file? The
error
I'm getting doesn't look like the missing 4 files error that I'm finding
in
the archives. Unfortunately my copying of the error text didn't work
and I
already closed the window, but it had a series of WARNING
nifti_read_buffer
where the first one in the list had the data bytes input as maybe 2/3 of
the
bytes needed, and then the remaining ones in the list (filled the
screen)
were 0 for data bytes input.

For this dataset, eddy_correct and dtifit were either done in 3.3 or 4.0
(or
later), bedpostx was in 4.0 or later, and probtrackx was in 4.1.2. At
the
time I was running an additional subject through probtrackx in another
window. The computer has 4GB of RAM (Mac - Dual Core Intel) and nothing
else was running at the time. I had gotten a similar error the first
time I
tried to run two subjects at the same time and I think the error comes
on
the subject that I start running first, but I may have mistaken which
subject the first error was on.

I'm currently rerunning probtrackx on the subject I just got the error
on
(my 2nd attempt to run two people at once) to see if I get the error
again
when it's only one running one at a time.

Anyone have any thoughts on this? Thanks and sorry for the longwinded
e-mail!

Kristen