nbench-byte

[Igor Schein <igor txc com>]

Hi,

I've been playing with nbench-byte on my 21164 Alpha, using ccc with
different flags to compare optimization.  Here're a few observations:

1)  -O4 is better than -O3
2)  using -inline all improves speed of all tests except BITFIELD
test, which loses about 15% of performance.  It'd be interesting if
some1 could take a look into it. 	
3) when both -O4 and -inline all are used, the program gets stuck forever
in FP SIMULATION test.  Can any1 reproduce it?  I already submitted a bug
report on it.

Any comments?

Thanks

Igor