heya, worker_threads=8 does sound high. i'm using 2 here. note that all image processing is parallelised to all cores independently of this setting. this is just for slight overscheduling of different tasks, to hide the few bits of scalar code between the parallel sections. i.e. on a 16-thread machine (maybe hyperthreaded 8-core), you'd get something like 2x16=32 threads with two worker threads, plus the gui thread. i'd try reducing this setting.
cheers, jo On Sat, Apr 21, 2018 at 3:26 AM, <dt-l...@stefan-klinger.de> wrote: > johannes hanika (2018-Apr-21, excerpt): >> did you at some point adjust the number of worker threads by any >> chance? > > Not that I'd be aware of. I have > > $ grep work ~/.config/darktable/darktablerc > worker_threads=8 > > Anyplace else? > >> there is a surprising amount of different stuff going on that >> all allocates memory.. 8GB isn't all that much in the end. > > Well I never see darktable using more than 2GB, so theres like 6GB > left unused. > >> among others, the nvidia opencl backend tries to malloc something, too. so >> i'd >> >> * look through settings for super large number of worker threads somewhere >> * try with `darktable --disable-opencl` to see whether the driver >> needs all the memory > > Also segfaults. I'll attach a backtrace as before: > > $ gdb --args darktable --disable-opencl > (gdb) set logging on > (gdb) run > (gdb) thread apply all backtrace > > The picture is not entirely clear: After trying many times, it > suddenly works. Then I can sometimes recreate the error condition by > selecting one image, copy the history stack (copy all), select 30 > images, and paste the history stack (paste all, overwrite). > “Sometimes” as in “this triggers the error, but not always”. > > The following is the output from Nvidia's monitoring tool, started > before, and terminated after a crash. To me it does not look like > running out of memory. > > $ nvidia-smi dmon > # gpu pwr temp sm mem enc dec mclk pclk > # Idx W C % % % % MHz MHz > 0 12 45 0 3 0 0 405 139 > 0 12 45 0 3 0 0 405 139 > 0 12 45 0 3 0 0 405 139 > 0 12 45 0 3 0 0 405 139 > 0 12 45 0 3 0 0 405 139 > 0 12 45 0 3 0 0 405 139 > 0 12 45 0 3 0 0 405 139 > 0 12 45 0 3 0 0 405 139 > 0 12 45 0 3 0 0 405 139 > 0 12 45 0 3 0 0 405 139 > 0 12 45 0 3 0 0 405 139 > 0 12 45 0 3 0 0 405 139 > 0 12 45 0 3 0 0 405 139 > 0 12 45 0 3 0 0 405 139 > 0 12 45 0 3 0 0 405 139 > 0 12 45 0 3 0 0 405 139 > 0 12 45 1 3 0 0 405 139 > 0 12 45 2 3 0 0 405 139 > 0 12 45 0 3 0 0 405 139 > 0 12 45 0 3 0 0 405 139 > 0 12 45 1 3 0 0 405 139 > 0 12 45 3 4 0 0 405 139 > 0 12 45 0 3 0 0 405 139 > 0 12 45 0 3 0 0 405 139 > 0 12 45 0 3 0 0 405 139 > 0 12 45 0 3 0 0 405 139 > 0 12 45 0 3 0 0 405 139 > 0 12 45 0 3 0 0 405 139 > 0 12 45 0 3 0 0 405 139 > 0 12 45 0 3 0 0 405 139 > 0 12 45 0 3 0 0 405 139 > 0 12 43 0 3 0 0 405 139 > 0 12 45 0 3 0 0 405 139 > 0 12 45 0 3 0 0 405 139 > 0 12 45 0 3 0 0 405 139 > 0 29 45 0 1 0 0 3802 1544 > 0 29 46 0 0 0 0 3802 1544 > 0 29 46 10 3 0 0 3802 1544 > 0 29 46 2 1 0 0 3802 1544 > 0 29 46 0 0 0 0 3802 1544 > 0 29 46 0 0 0 0 3802 1544 > 0 29 46 0 0 0 0 3802 1544 > 0 29 45 0 0 0 0 3802 1544 > 0 29 46 0 0 0 0 3802 1544 > # gpu pwr temp sm mem enc dec mclk pclk > # Idx W C % % % % MHz MHz > 0 29 46 0 0 0 0 3802 1544 > 0 30 46 0 0 0 0 4006 1544 > 0 30 46 0 0 0 0 4006 1544 > 0 26 46 0 0 0 0 4006 911 > 0 26 46 0 0 0 0 4006 911 > 0 26 46 0 0 0 0 4006 911 > > Thanks > Stefan > > > -- > http://stefan-klinger.de o/X > Send plain text messages only, not exceeding 32kB. /\/ > \ > > ____________________________________________________________________________ > darktable user mailing list > to unsubscribe send a mail to darktable-user+unsubscr...@lists.darktable.org ____________________________________________________________________________ darktable user mailing list to unsubscribe send a mail to darktable-user+unsubscr...@lists.darktable.org
