Hi Dex,
You'll be glad to hear that I'm working on the 64-bit Windows build of
enblend/enfuse. I've sorted out 64-bit builds of most of the
dependencies (and their dependencies), so I'm quite close to the actual
build. The following text from the v4.0 readme might be of interest to
you - I'll endeavour to produce builds with a variety of permutations of
these options.
***** ENABLE_IMAGECACHE=ON/off
Activate the Image-Cache feature, which allows for processing of large
images. The Image Cache handles swapping parts of the images to disk.
The default is "on", i.e., enable the image cache.
The image cache is recommended on systems where memory is scarce, but
the images to blend or fuse are large.
The image-cache feature and OpenMP support must not be activated
together.
***** ENABLE_OPENMP=on/OFF
Parallelize parts of Enblend and Enfuse with OpenMP. See
http://www.openmp.org/
Note that OpenMP support and the image cache must not be activated
both. The default is to configure with image cache and without
OpenMP.
If OpenMP support has been enabled, the utilization of special
features of the actual, underlying OpenMP implementation can be
controlled as usual with the environment variables OMP_NUM_THREADS,
OMP_NESTED and OMP_DYNAMIC. See the OpenMP specification for details
on the usage of these variables.
Tom
On 20/05/2010 9:13 AM, dex Otaku wrote:
On May 18, 6:20 am, Tom Glastonbury<[email protected]> wrote:
Note that enblend/enfuse version 4.0 is what you'll get from my Windows
64-bit hugin builds. As you rightly say, it's not the 64-bit version -
I'll work on that. As for your question - from reading some comments
elsewhere, it looks like enblend/enfuse use some hand-crafted SSE code,
and will use the 64-bit versions for 64-bit builds (note this is a
somewhat speculative comment). I'd imagine that this could make a
meaningful difference to performance. Anyhow, I'll get the 64-bit
version going shortly then you can compare them side-by-side.
I'm not even that worried about potential performance hits between
32/64-bit .. What concerns me more is that rendering even modestly-
sized panos [16k*8k equirectangular] from more than a few images [37
or 38 in my case] with --fine-mask turned on [as mentioned above]
causes enblend to exit with out of memory errors [only after doing
most of the job, frustratingly] apparently regardless of the amount of
actual RAM or swap available.. on Windows, at least.
Some digging through the Mac Hugin fora uncovers the same issue which
can apparently be fixed on that platform by using the openmp
version[s] of enblend. On Windows, this is not the case,
unfortunately [at least, not with the openmp version of enblend 3.2
which exits with the same error].
Your work is very much appreciated, Tom, as is Yuval's.
--
You received this message because you are subscribed to the Google Groups "hugin and
other free panoramic software" group.
A list of frequently asked questions is available at:
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to
[email protected]
For more options, visit this group at http://groups.google.com/group/hugin-ptx
