Hi Ikko

There is no  "- - g" option, but a “-g" one. Retry using:

bedpostx <datadir> -n 3 -model 2 -g

Cheers
Stam

On 2 Jul 2014, at 14:41, 木村一皓 <[log in to unmask]> wrote:

> Dear all,
>
> I'm now using NVIDA CUDA 5.5, Grid Engine 2011.11, Centos 6.5, fsl 5.06.
> When I try to run bedpostx_gpu from HCP data like this,
>
> bedpostx_gpu <subject directory> --n=3 --model=2 --g
>
> then came
>
> subjectdir is <subject directory>
> subjectdir is <subject directory>
> Making bedpostx directory structure
> Copying files to bedpost directory
> Queuing parallel processing stage
> Queuing post processing stage
> Type <subject directory>.bedpostX/monitor to show progress.
> Type <subject directory>.bedpostX/cancel to terminate all the queued tasks.
> You will get an email at the end of the post-processing stage.
>
> When I typed <subject directory>.bedpostX/monitor, there came the problem as follows.
>
> 0 parts processed of 67
> All parts processed
>
>  And, when I tried bedpostx instead of bedpostx_gpu, neither was successful, with " 0 parts processed of 67  All parts processed."
> How could it occur?  Am I missing some steps?
>
> Best,
> Ikko
>
>