Am 05.08.2017 um 21:57 schrieb Allin Cottrell: > Details (including details on how to get this working on Linux) can be > found at > > http://ricardo.ecn.wfu.edu/pub/gretl/svm/gretl-svm.pdf First, this is great and might also be something for Artur's Twitter feed. Of course clearly indicating this is alpha stage and testers are needed.

> I'll be glad to hear from testers when the holidays are over. So far I can only report (with the latest snapshot) that your first example runs OK here, too.

But when we specificy a "ranges_outfile" parameter and that is a file name as per the doc, where would that file be saved on my disk? I looked for it but didn't find it (at least not quickly).

Am 17.08.2017 um 21:07 schrieb Allin Cottrell: > On Thu, 17 Aug 2017, Sven Schreiber wrote: >> >> First, this is great and might also be something for Artur's Twitter feed. >> Of course clearly indicating this is alpha stage and testers are needed. > > Agreed, but before advertising this some more I'd like to catch the > documentation up with the code -- which now supports MPI for > cross-validation, among other refinements. I see. >> But when we specificy a "ranges_outfile" parameter and that is a file name >> as per the doc, where would that file be saved on my disk? I looked for it >> but didn't find it (at least not quickly). > > It should get written into your workdir (which is now, of course, displayed > in the GUI unless you choose to turn that off). If that didn't happen, it's > a bug. I don't think it did. I explicitly set the workdir to the newly created SVM dir and ran the example listing 1 (slightly modified, leaving out the file open line since the file was already loaded). No new files have appeared there.