On Sun, Dec 17, 2017 at 10:33 PM, Pascal Obry wrote:
>
> István,
>
> > I'm using darktable 2.2.5 from Pascal's PPA on Mint 18.2. I upgraded
> > from Nvidia 340 to 384, now opencl fails. It does compile the first
> > module at least, so I'm not sure if I really have a missing
Unfortunately the OpenCL compiler fails to give a reasonable build log.
What you could do is try to isolate the offending part of atrous.cl that
leads to the issues. Start by commenting out the whole code section of
atrous.cl with "#if 0 ... #endif" and narrow it down from there.
Ulrich
Am
Thanks for the quick reply. My issue seems to be different: the first
module gets compiled.
I've tried the workaround, anyway, but it failed:
==
eagle ~ # cd /usr/lib/x86_64-linux-gnu
eagle x86_64-linux-gnu # ls
nvidia/current/libnvidia-ptxjitcompiler.so.1<-- this was the
original file linked
István,
> I'm using darktable 2.2.5 from Pascal's PPA on Mint 18.2. I upgraded
> from Nvidia 340 to 384, now opencl fails. It does compile the first
> module at least, so I'm not sure if I really have a missing package.
Looks like a known issue fixed in 384.98-3. See
Done!
And thank you Stefan! :)
And Pete, yes, I had found your blog at https://landscrapes.wordpress.
com/2016/05/10/darktable/ regarding sensor profiling.
Oops, now that I look at it, I take that back, I hadn't seen yours.
Thanks, I'll review it. I'd really like to work through this process
hi,
i just played a bit with dt-2.4.0rc1 and noticed that a style (created
by history from picture A) does not apply the contained spot removals to
another picture B - although it is listed in the history of picture B
(the number of strokes is '0')
gerd
Hi
please create a ticket at
https://redmine.darktable.org/projects/darktable/issues/new
and add your link there. I have done the noise profiling and I will attach it
to your ticket. Result look good and can be used.
Stefan
> Hi -
>
> I've been muddling around attempting to create a noise