I take it that #1 is the only step that's reading input images from disk? Are the threads each reading different images, or the same one? How big is the image, and is it tiled? Do you know how large your IC cache is set to, or just the default?
I'm trying to build a hypothesis of what might be going on to hit that assertion. On Sep 12, 2014, at 3:02 PM, Dave Lajoie <[email protected]> wrote: > I have 12 logical cores ( 6 physical cores ) and I am running 11 concurrent > threads which are doing the equivalent to a slate generation > > 1) Loading image in frame buffer > 2) Resize frame buffer > 3) Copy/paste specific channels between buffers > 4) apply gamma operation > 5) fill regions > 6) over operation > 7) write frame buffer to disk > > > Dave Lajoie > R&D Director | Directeur R&D > > 5605 Avenue de Gaspé, Suite 408 | Montréal, QC H2T 2A4 > Tel (514) 360-3253 ext. 130 > > On Fri, Sep 12, 2014 at 5:57 PM, Larry Gritz <[email protected]> wrote: > How many threads do you have going? And what tasks are they doing? > > > On Sep 12, 2014, at 2:54 PM, Dave Lajoie <[email protected]> > wrote: > >> Hello Larry, I have tested the python threading and I am getting the >> following message and it core dumps. >> >> /root/oiio-master/src/libtexture/imagecache.cpp:2017: failed assertion >> 'sweep != m_tilecache.end() && "no way m_tilecache can be empty and use too >> much memory"' >> Aborted (core dumped) >> >> Dave. >> >> Dave Lajoie >> R&D Director | Directeur R&D >> >> 5605 Avenue de Gaspé, Suite 408 | Montréal, QC H2T 2A4 >> Tel (514) 360-3253 ext. 130 >> >> On Sat, Sep 6, 2014 at 2:34 PM, Dave Lajoie >> <[email protected]> wrote: >> Sure no problem. >> >> > > -- > Larry Gritz > [email protected] > > > > -- Larry Gritz [email protected]
_______________________________________________ Oiio-dev mailing list [email protected] http://lists.openimageio.org/listinfo.cgi/oiio-dev-openimageio.org
