[R-SIG-Mac]timing R and R and R and R

Thomas Lumley tlumley@u.washington.edu
Sat, 17 Nov 2001 11:42:23 -0800 (PST)

On Sat, 17 Nov 2001, Jan de Leeuw wrote:

> I guess this implies that my ATLAS, which is the one used in the
> CRAN Darwin distribution, may not be optimal for your machine.

Yes. It's going to be better than the default BLAS routines, partly
because any sensible loop unrolling helps and partly because Macs aren't
*that* different, but people who compile their own ATLAS should see even
more improvement on large matrix operations.