Thank you very much.
It really works with your advice.
Best regards,
Kim
> 2018. 8. 23. 오전 8:55, Moises Hernandez <[log in to unmask]> 작성:
>
> Yes, a compatible problem.
> Your GPU has compute capability 6.1 (https://developer.nvidia.com/cuda-gpus <https://developer.nvidia.com/cuda-gpus>)
> and that is not supported for CUDA versions <= 8.0
>
> I recommend you to install CUDA 9.2 and download the correct files from here:
> https://users.fmrib.ox.ac.uk/~moisesf/Bedpostx_GPU/Installation.html <https://users.fmrib.ox.ac.uk/~moisesf/Bedpostx_GPU/Installation.html>
>
> Moises.
>
> On Wed, 22 Aug 2018 at 19:48, 김대현 <[log in to unmask] <mailto:[log in to unmask]>> wrote:
> Hi Hernandez.
>
> Thanks for your response.
>
> I found error log in bedpostx_gpu.o4101.1~4101.4.
>
> The error message is below.
>
> ...................Allocated GPU 0...................
> Log directory is: /home/dkim/Downloads/sub-CC110045.bedpostX/diff_parts/data_part_0000
> Number of Voxels to compute in this part: 48717
> Number of Directions: 63
> SubPart 1 of 4: processing 12800 voxels
> ERROR: fit_PVM_single_kernel: invalid device function
>
> Is that means that the CUDA or GPU did not work properly?
>
> Best regards,
>
> Kim.
>
>
>
>> 2018. 8. 23. 오전 8:32, Moises Hernandez <[log in to unmask] <mailto:[log in to unmask]>> 작성:
>>
>> Hi Kim,
>>
>> I would check that the samples from bedpostx_gpu were generated correctly.
>> You can check for instance {subject}.bedpostx/merged_f1samples.nii.gz file
>>
>> Also, check the logs files: {subject}.bedpostx/logs
>> The may contain an error message.
>>
>> Moises.
>>
>> On Wed, 22 Aug 2018 at 19:24, SUBSCRIBE FSL DH Kim <[log in to unmask] <mailto:[log in to unmask]>> wrote:
>> Hello FSL experts,
>>
>> I'm Kim from South Korea.
>>
>> I need some help.
>>
>> I recently use bedpostx_gpu in Centos 7, GeForce GTX 1060, CUDA 7.5 and FSL ver 5.0.11.
>>
>> The bedpostx_gpu work without error message.
>>
>> However, when I do PROBTRACKX with the folder of bedpostx_gpu, the fdt_paths.nii.gz showed only showed the region of interest that I drawed.
>>
>> When I do the same PROBTRACKX with the folder of bedpostx without gpu, the result showed good. I did the same patient with same ROI file.
>>
>> The difference is only the xxxx.bedpostX folder.
>>
>> Here is the fdt.log with bedpostx result without gpu.
>>
>> set tool probtrackx
>> set probtrack(bedpost_dir) /Volumes/GRAID/prob_tra/sub-CC110045.bedpostX
>> set probtrack(xfm)
>> set probtrack(mode) seedmask
>> set probtrack(exclude_yn) 0
>> set probtrack(usereference_yn) 0
>> set probtrack(verbose_yn) 0
>> set probtrack(loopcheck_yn) 1
>> set probtrack(modeuler_yn) 0
>> set probtrack(curvature) 0.2
>> set probtrack(nsteps) 2000
>> set probtrack(steplength) 0.5
>> set probtrack(nparticles) 5000
>> set probtrack(fibthresh) 0.01
>> set probtrack(distthresh) 0.0
>> set probtrack(sampvox) 0.0
>> set probtrack(oneway_yn) 1
>> set probtrack(wayorder_yn) 0
>> set probtrack(waycond) AND
>> set probtrack(reference) /Volumes/GRAID/prob_tra/sub-CC110045/ROI2.nii.gz
>> set probtrack(output) /Volumes/GRAID/prob_tra/output4
>>
>>
>> Here is fdt.log with gpu.
>>
>> set tool probtrackx
>> set probtrack(bedpost_dir) /Volumes/GRAID/prob_tra/bedpostx_gpu_s_share/sub-CC110045.bedpostX
>> set probtrack(xfm)
>> set probtrack(mode) seedmask
>> set probtrack(exclude_yn) 0
>> set probtrack(usereference_yn) 0
>> set probtrack(verbose_yn) 0
>> set probtrack(loopcheck_yn) 1
>> set probtrack(modeuler_yn) 0
>> set probtrack(curvature) 0.2
>> set probtrack(nsteps) 2000
>> set probtrack(steplength) 0.5
>> set probtrack(nparticles) 5000
>> set probtrack(fibthresh) 0.01
>> set probtrack(distthresh) 0.0
>> set probtrack(sampvox) 0.0
>> set probtrack(oneway_yn) 1
>> set probtrack(wayorder_yn) 0
>> set probtrack(waycond) AND
>> set probtrack(reference) /Volumes/GRAID/prob_tra/sub-CC110045/ROI2.nii.gz
>> set probtrack(output) /Volumes/GRAID/prob_tra/output5
>>
>> What should I check for the BEDPOSTX_gpu?
>>
>> I appreciate your concern.
>>
>> Best regards,
>>
>> KIM
>>
>> ########################################################################
>>
>> To unsubscribe from the FSL list, click the following link:
>> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1 <https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1>
>>
>> To unsubscribe from the FSL list, click the following link:
>> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1 <https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1>
>
> To unsubscribe from the FSL list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1 <https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1>
> To unsubscribe from the FSL list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1 <https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1>
########################################################################
To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
