BLAS -- low level basic linear algebra. Along with LAPACK, you can
buy a book on these industry standard APIs in stores.
LAPACK -- higher level linear algebra stuff
These two ROCK for one of our apps. Java front end, way-fast matrix
multiply and invert at the back, and out come solved redundant
simultaneous equations. Currently, Mac-only, but we have some code
in place to decide between a java translation of the c translation of
the original FORTRAN, or a native lib. When last we tested it, it
ran on everything, but screamed on a dual G5.
Of course, this was a near-ideal data set for JNI - a fairly small
matrix and vector thrown across the wall, then inverted, multiplied,
and perturbed many times until convergence, with just a small vector
coming back.
With a few exceptions, its all vectorized (except double
precision), its all 64-bit savvy (except vDSP) and its all free to
use and comes preinstalled on users' machines.
Pegging two cpus at once with no extra effort is always fun.
Scott
_______________________________________________
Do not post admin requests to the list. They will be ignored.
PerfOptimization-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden
