Re: gcj performance on Solaris 2.6

[Tom Tromey <tromey at redhat dot com>]

>>>>> "Van" == Van Nguyen <v_b_nguyen@hotmail.com> writes:

Van> My problem was I didn't realize I can use the same optimization
Van> options on gcj.  So it was run without any optimizations.  I
Van> tried the optimizers you used, and got 25400 iterations out of
Van> Sieve. And it beats Java Hotspot.

Be warned that gcc without -O generates really bad code.
If you are running performance tests, always use at least -O1.
-O2 is probably better.  Often, -O3 can give worse results than -O2,
though, because (I hear) it inlines too aggressively (perhaps not a
problem for Java, where inlining is still not quite there).

Tom