You can pick up the tests if Andrew Morgan still has them on his system, at
ftp://fire.physics.ucla.edu/pub/benchmarks/wes-bench.tar.gz
If they're gone, I can send you a SHAR file tonight. (system is at home)
Both he and I are very curious what the potential performance of
the Alpha is for floating point work. The best I've seen so far is
in the filter application code in the benchmark. Andrew's run was
a little over 200 MFLOPS. I would expect with a good compiler that
it should run over 500 MFLOPS. My Ppro 200MHZ does 110MFLOPS on it.
I'm curious about your comment on the binary. Do you mean that if one
builds a static linked binary, it will execute under Linux?? I assume
if you use shared object modules, that wouldn't work.
Wes