.. okay, it has been removed. iirc this is the loop that used to be parallel (needs some added sync inside):
https://github.com/darktable-org/darktable/blob/master/src/control/jobs/control_jobs.c#L1236 -jo On Mon, Sep 10, 2018 at 3:26 PM johannes hanika <hana...@gmail.com> wrote: > > heya, > > to tell you the truth i don't remember the state of the exporter. we > did have this option precisely for what you describe. one pixel > pipeline one GPU, and several for the export. i think it did lead to a > certain amount of race conditions for sequence numbers etc, and for > CPU based export, it didn't actually speed things up (in fact it > slowed it down). > > so the option was removed but it's possible that the backend still > supports it to some extent. even for multi-GPU setups our experience > back then was that export speed was much more io bound (read big RAW, > write full res high quality JPG..) than compute bound, so the idea > didn't receive much attention. today we arguably have quite a few very > expensive operations so we might want to re-evaluate this. > > cheers, > jo > > On Mon, Sep 3, 2018 at 8:09 PM Lutz Labusch <lutz.labu...@gmail.com> wrote: > > > > Hi all, > > > > i want to say thank you for darktable, which is a pretty nice program, > > first. > > > > As a owner of two (n) powerful GPU devices it will be very interesting to > > have the possibility to have two (n) exporting threads for larger jobs. Is > > there somewhere an hidden option for this? > > > > Thanks and best regards > > > > ___________________________________________________________________________ > > darktable developer mailing list to unsubscribe send a mail to > > darktable-dev+unsubscr...@lists.darktable.org ___________________________________________________________________________ darktable developer mailing list to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org
