This is the mail archive of the
java@gcc.gnu.org
mailing list for the Java project.
Re: change of libgcj soname for gcc-3.2?
Jeff Sturm <jsturm@one-point.com> writes:
> Anyone besides me think --indirect-dispatch should be on by default? (Once
> the interpreter is fixed, that is.)
I recall Bryce's earlier benchmarks showed zero or near-zero impact on
performance. If this is the case, then definately.
It just doesn't seem to make sense to me though -- doesn't indirect
dispatch add an extra LOAD before each CALL? Java code tends to have
tons of CALLs, since just about everything you do is a result of
invoking some method on some object...
I dunno, I believe the benchmarks, I just can't figure out why it
performs so well. Can somebody do some hand-waving for me?
- a
--
Sick of HTML user interfaces?
www.xwt.org
Some people don't care if the pie is smaller, so long as they still
get all of it.