This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: gcj Linpack performance
- To: gcc at gcc dot gnu dot org
- Subject: Re: gcj Linpack performance
- From: Alexandre Petit-Bianco <apbianco at fencer dot cygnus dot com>
- Date: 10 Jul 2001 09:51:44 -0700
- Newsgroups: cygnus.egcs
- Organization: Cygnus Support
- References: <20010710123524.E12110@csr-pc2.zib.de>
[Note that you could add java-project@gcc.gnu.org when debating of
java related issues, that'll let you reach our Java audience.]
Florian Schintke <schintke@zib.de> writes:
> I just tested the performace of gcj 3.0 with the Linpack benchmark
> http://www.netlib.org/benchmark/linpackjava/Linpack.java and got the
> following results:
I guess you had to modify the source code a bit.
> Why is gcj not as fast as IBMs JIT compiler? Maybe this info helps
> you to look how to improve the performance.
There are a lot of reasons (and to summarize: we haven't really done a
lot of optimization work) but given the structure of the Linpack
source code (prior to modification to let it run with gcj, I guess)
sub-optimal inlining seems to be a likely culprit.
Could you please post the code you've been working with (if it's
allowed -- it seems so)
./A