On Thu, 26 Aug 2004, IOhannes m zmoelnig wrote:
guenter geiger wrote:
I had done that 4 years ago with the (original?) sig2pix external that I wrote. It was fast. It was not very general though ...
the sig2pix hasn't changed much since (at least i don't think that performance has decreased)
Ah, sorry then. I have to admit that I haven't used that for a long time, and I could vaguely remember that in order to get a waterfall display with the new pix2sig you had to use a large blocksize with overlap, which in theory should reduce the performance "overlap" times. But maybe I am talking nonsense here.
Guenter
i haven't done any profiling, but i think a lot of CPU-power is wasted while casting from float to int and back again, and when getting the luminance of the image to map to vertices. performance isn't *that* bad, bit i thought it could be better.
i don't think that conversion to pixes and then to vertices is optimal.
Just found it amusing ... back to lurker state.
mfg.asr IOhannes
GEM-dev mailing list GEM-dev@iem.at http://iem.at/cgi-bin/mailman/listinfo/gem-dev