Convolution and LAT updated for OpenMP

The latest GraphicsMagick 1.3 snapshot (1.3.020080906) is updated so 
that convolution and local area thresholding are accelerated on 
multicore systems using OpenMP.

Example of LAT (-lat 10x10-5%) speed improvement:

Results: 10 iter 18.87s user 18.88s total 0.53 iter/s (0.53 iter/s cpu)
Results: 10 iter 18.97s user 10.13s total 0.99 iter/s (0.53 iter/s cpu)
Results: 10 iter 19.02s user 7.20s total 1.39 iter/s (0.53 iter/s cpu)
Results: 10 iter 19.17s user 5.75s total 1.74 iter/s (0.52 iter/s cpu)

Example of convolve (-sharpen 0x1.0) speed improvement:

Results: 10 iter 17.02s user 17.03s total 0.59 iter/s (0.59 iter/s cpu)
Results: 10 iter 16.99s user 9.12s total 1.10 iter/s (0.59 iter/s cpu)
Results: 10 iter 16.96s user 6.46s total 1.55 iter/s (0.59 iter/s cpu)
Results: 10 iter 16.95s user 5.18s total 1.93 iter/s (0.59 iter/s cpu)

Since these algorithms take a lot of CPU, they speed up linearly as 
cores are added.

I have yet to hear about anyone else's experience with OpenMP-enabled 
GraphicsMagick 1.3.

Thus far I am finding that Solaris and Linux are very good at OpenMP 
and FreeBSD and OS-X are rather poor with OpenMP.

Bob
======================================
Bob Friesenhahn
[email protected], http://www.simplesystems.org/users/bfriesen/
GraphicsMagick Maintainer,    http://www.GraphicsMagick.org/


-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/