Am 19.03.21 um 09:26 schrieb Harry van der Wolf:
Op vr 19 mrt. 2021 om 08:41 schreef 'Kay F. Jahnke' via hugin and other
free panoramic software <[email protected]
<mailto:[email protected]>>:
Am 18.03.21 um 21:21 schrieb Harry van der Wolf:
> I can't find the question backs, but you asked me about the speed
after
> your improvements.
>
> The current version is about 10-15% faster.
Ha, that's nice to hear, Did you look at memory consumption, too?
That's
where I had hoped to see about 30% improvement , because I switched to
single-float masks.
Yes, that too. The memory usage went from 3.2GB to 2.7GB for the image
sets I used (sets of 5 brackets of 13.5MB).
Less than I hoped, but every little helps ;)
With regard to the "??--snapshot_threads=X".
That still doesn't work for me. I tested on linux on the master repo. I
have a 4-core CPU. no matter whether I set X to 4, 3 or 2. Always all 4
cores are used and it actually creates many more processes.
The command I use from my batch file is:
${LUX} -W \
?? ?? ?? ?? ?? ?? ??--blending=hdr \
?? ?? ?? ?? ?? ?? ??--snapshot_like_source=yes \
?? ?? ?? ?? ?? ?? ??--snapshot_facet=1 \
?? ?? ?? ?? ?? ?? ??--snapshot_prefix="${PREFIX}" \
?? ?? ?? ?? ?? ?? ??--snapshot_extension=$SUFFIX \
?? ?? ?? ?? ?? ?? ??--aeb_auto_brightness=yes \
?? ?? ?? ?? ?? ?? ??--next_after_fusion=yes \
?? ?? ?? ?? ?? ?? ??--build_pyramids=no \
?? ?? ?? ?? ?? ?? ??--build_raw_pyramids=no \
?? ?? ?? ?? ?? ?? ??--quality_interpolator_degree=1 \
?? ?? ?? ?? ?? ?? ??--snapshot_threads=3 \
?? ?? ?? ?? ?? ?? ??"${PTO}"
In this case you see " --snapshot_threads=3", but of course I did vary
that number.
Good thing you're so meticulous. I managed to reproduce the behaviour
you criticized, and it made me find a silly bug where, while lux waits
for a snapshot, it simply burns cycles in an unchecked loop, wehre it
should instead have slept brief naps and checked only every few
milliseconds whether the snapshots were finally ready. No surprise your
CPU was busy.
I pushed the bug fix to master.
lux will still create many threads - on a four-core there should be 18
or so - but they should *not* be active together while the exposure
fusion happens.
Please try again, and if it still doesn't work let me know.
Kay
--
A list of frequently asked questions is available at:
http://wiki.panotools.org/Hugin_FAQ
---
You received this message because you are subscribed to the Google Groups "hugin and other free panoramic software" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/hugin-ptx/d72267ea-7340-4c6c-d3db-8e1e594fd093%40yahoo.com.
