Sather performance--I'm impressed.

[Johann Hibschman]

Hi all,

Yes, I'm once more filling this mostly-silent newsgroup with frothing
rants. Well, nice frothing rants this time.

In a desperate attempt to not do any work, I converted a C++ computational
program that I've been working on over to Sather. I ended up with a
final program that was cleaner than the C++ one, that was easier to
extend, and that ran about twice as fast.

Like I said, I'm impressed.

The precise problem was an exercise in numeric integration in which I
had integrals of integrals of bessel functions, where I had
approximated the various functions at each step with heavily tweaked
Chebyshev polynomials to ease the computational load. So, in any case,
it's not vector/matrix math, nor is it complex hydrodynamic simulation,
but it still works, and works well.

On the downside, now I have to think really hard to figure out what all
these nice numerical results are telling me.

Cheers,

- Johann

--
Johann Hibschman
joh...@physics.berkeley.edu