Hmm... strange. The documentation is a bit confusing:
- http://www.darktable.org/usermanual/ch06s02.html.php 'number of
background threads: This controls how many parallel threads are used
to create thumbnails during import. On 32bit systems it is strongly
recommended to set this to 1. Needs a restart if changed (default 2).'
- http://www.darktable.org/usermanual/ch08.html.php#d0e15770 'number
of background threads: This parameter defines the maximum number of
threads that are allowed in parallel when importing film rolls or
doing other background stuff. For obvious reasons on 32-bit systems
you can only have one thread eating resources at a time. So you need
set this parameter to 1; anything higher will kill you. For the same
reason you also must set the number of parallel export threads to 1.'

Note the 2nd definition includes 'or doing other background stuff',
and also mentions a setting 'parallel export threads'; that seems to
refer to a setting called 'parallel_export' (default=1): 'set this
variable to num_threads if you want multithreaded export to process
multiple images at a time. be warned: every thread will need at the
very least 1GB of memory. setting this to 1 switches on per-image
parallelization'. According to the blog entry on OpenCL
(http://www.darktable.org/2012/03/darktable-and-opencl/), it's best
left at the default of 1: 'Before going into the details, the above
already makes clear that we should not process several images in
parallel with OpenCL. We already make maximum use of GPU memory by
tiling and the nature of GPU processing will already parallelize
processing to the max on a pixel by pixel basis. No room for
additional parallelization. In preferences set "export multiple images
in parallel" to 1.'

Thanks for the info,
Kofa

On 6 October 2014 14:53, Markus Jung <[email protected]> wrote:
> Because it is not just one thread ;)
>
> darktable exploits parallelization at different levels. Most, if not all
> modules distribute their computations automagically to different
> threads. Adding more concurrency by running several rendering pipes in
> parallel will most likely have a negative performance impact because of
> negative cache effects and increased overhead.
> Additionally, this option does only affect the thumbnail generation
> process. I have not observed the CPU usage during thumbnail generation,
> but i am assuming there is not much benefit from running it in parallel.
>
> Regards,
> Markus
>
> Am 06.10.2014 um 14:43 schrieb KOVÁCS István:
>> Why do you limit yourself to just 1 thread? On my system (4 GB RAM,
>> old dual-core CPU) 2 threads work fine with 8 GB of swap space.
>>
>> On 6 October 2014 11:38, Markus Jung <[email protected]> wrote:
>>> Hello Bernhard,
>>>
>>> most likley, something with your darktable configuration went wrong.
>>> Please take a look at your core preferences:
>>> http://www.darktable.org/usermanual/ch06s02.html.php
>>>
>>> I am using (with the same amount of RAM):
>>> mipmap cache: 1536
>>> background threads: 1
>>> host memory limit: 4096
>>> minimum single buffer size: 16
>>>
>>> Regards,
>>> Markus
>
>
> ------------------------------------------------------------------------------
> Slashdot TV.  Videos for Nerds.  Stuff that Matters.
> http://pubads.g.doubleclick.net/gampad/clk?id=160591471&iu=/4140/ostg.clktrk
> _______________________________________________
> Darktable-users mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/darktable-users

------------------------------------------------------------------------------
Slashdot TV.  Videos for Nerds.  Stuff that Matters.
http://pubads.g.doubleclick.net/gampad/clk?id=160591471&iu=/4140/ostg.clktrk
_______________________________________________
Darktable-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/darktable-users

Reply via email to