Problem solved... quite embarassing.
When I changed the header files to reflect that the data was 256x256, I
didn't change the voxel sizes... So the voxels were 3.44x3.44x4, or
whatever, when the x and y values should have been a quarter that size.
That was causing some problems. Now it works fine.
Thanks,
Chris
Dennis Thompson wrote:
> Are you running a 64 bit version of matlab. The 32 bit version will
> run fine on 64bit machines but will have the same memory issues as if
> running 32 bit architecture.
>
> On Fri, Nov 7, 2008 at 11:04 AM, Chris Watson
> <[log in to unmask]> wrote:
>
>> I recompiled the mex files, and still got the same problem. I'm not sure if
>> there's a memory leak, because both 'top' and 'free' said that only 50% of
>> the system's memory was being used.
>>
>> I've attached a screenshot of one of the FA maps. I don't have a clue as to
>> why the coronal and sag views are squished like that... Perhaps the header
>> file is incorrect? I originally created the header files assuming the data
>> would be 64x64; I then just changed the x and y values to 256. I make the
>> headers in MRIcro. I ran segmentation on one of the older FA images, that
>> was 64x64, and it worked fine.
>>
>> Thanks,
>> Chris
>>
>> John Ashburner wrote:
>>
>>> It is possible that the various *.mexa64 files need to be recompiled for
>>> your platform. I don't have access to one of those nice new 64 bit
>>> computers so I can't be of much help, but if a 64 bit machine with 8Gbytes
>>> of RAM is having memory problems, then some form of memory leak seems
>>> plausible.
>>>
>>> Another thing to check is that voxel sizes, dimensions etc are all correct
>>> in your images. If your images appear to have a field of view better
>>> measured in light years than mm, then this would lead to memory problems.
>>>
>>> Best regards,
>>> -John
>>>
>>> On Friday 07 November 2008 02:24, Watson, Christopher wrote:
>>>
>>>
>>>> I tried running segmentation on a single image, and got:
>>>>
>>>> Error running job: Error using ==> horzcat
>>>> Out of memory. Type HELP MEMORY for your options.
>>>> In file "/occipital/matlab/spm5/spm_preproc.m" (v946), function
>>>> "spm_preproc" at line 475. In file
>>>> "/occipital/matlab/spm5/spm_config_preproc.m" (v1032), function "execute"
>>>> at line 472.
>>>>
>>>> My machine has 8 Gb of RAM. This has never been an issue in the past...
>>>> I'm
>>>> running RHEL4, x86_64 arch.
>>>>
>>>>
>>>> -----Original Message-----
>>>> From: John Ashburner [mailto:[log in to unmask]]
>>>> Sent: Thu 11/6/2008 3:44 PM
>>>> To: Watson, Christopher; [log in to unmask]
>>>> Subject: Re: [SPM] Seg Violation; mex file spm_krutil
>>>>
>>>> Which SPM version, and which updates were installed?
>>>>
>>>> Best regards,
>>>> -John
>>>>
>>>> On Thursday 06 November 2008 18:28, Chris Watson wrote:
>>>>
>>>>
>>>>> I've been trying to run segmentation on a bunch of images. However, at
>>>>> some point during the first subject, a segmentation violation occurs.
>>>>> I've recreated the same exact problem several times now. I've attached
>>>>> the message that pops up in the window when it occurs. It has to do with
>>>>> spm_krutil.mexa64. Any suggestions for fixing this?
>>>>>
>>>>> Thanks,
>>>>> Chris
>>>>>
>>>>>
>>>
>>>
>>
>
>
|