>> The hardware was one of these PCI/ISA PC-style Alphas, with an
>> EB164 motherboard - the compiler flags that I explicitely
>> specified were -O3 for the Alpha and -O for the P6 (my codes
>> generally seem most happy with just -O on the P6) - the test that
>> failed definitely does
Bill> I'd try -O on the Alpha, gcc has been known to be over
Bill> exuberent with inlines that causes alot of cache misses, and
Bill> decreases performance.
Actually, in my experience, gcc -O2 is usually best. -O doesn't do
instruction scheduling and -O3 does too much automatic inlining (I did
say: in my experience, YMMV).
Another question: does your application do sqrt() a lot, by chance?
The old sqrt implementation was a real dog (it computed the result a
bit at a time!). I haven't kept track of what the latest libc uses
from Red Hat, glibc will use the fast version. If you want to try
this version, give me a holler.
Other than that: Bill's mail says it all. :-)
--david