Thanks, Daniel! Sorry for not having found the information about cmdline
parsing and data 'precision' earlier, hope you understand... :-)
On 08/09/2013 05:01 PM, Daniel Roe wrote:
> Hi,
>
> On Fri, Aug 9, 2013 at 4:44 AM, Jan-Philip Gehrcke
> <jgehrcke.googlemail.com> wrote:
>> would it be easy for you not to cut dataset names? For example,
>
> This only happens with the 'standard' data file output format (legends
> are not truncated in grace/gnuplot output) and the reason for it goes
> way back to the beginning of cpptraj. At that time cpptraj was really
> just a side project and I was playing around with the best way to
> parallelize things, and one of the ideas I played with was having data
> split across different MPI threads and then having each thread write
> it's data to a single file, rather than have all data sent to one node
> and then written. In order to do this correctly of course you need to
> know the output file size in advance so the file pointer for each node
> can be correctly positioned. In order to make it as simple as possible
> to do this I decided everything needed to be a fixed width, and
> arbitrarily decided to just base it on the width of the data to be
> written. This is why for example default data set names have a 5-digit
> fixed-width extension. Of course, I ended up deciding to focus on
> parallelizing things with OpenMP (although in AmberTools14 cpptraj
> there will also be MPI-enabled stuff) so currently the fixed-width
> truncation only serves to play havoc with large legend names. It's
> something I've thought about changing a lot but
>
> However, you can currently work around this by manually setting column
> widths via the 'precision' command: just set the width of each data
> set large enough so your legends fit. For example, to ensure your
> legend is going to be 20 chars wide:
>
> radgyr out radgyr.mass.dat :1-13 mass
> precision radgyr.mass.dat 20 4
>
> Hope this helps,
>
> -Dan
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Aug 09 2013 - 08:30:04 PDT
