This is the mail archive of the
java@gcc.gnu.org
mailing list for the Java project.
RE: status of gcj's boehm collector?
- From: Jeff Sturm <jsturm at one-point dot com>
- To: "Boehm, Hans" <hans_boehm at hp dot com>
- Cc: "'Adam Megacz'" <gcj at lists dot megacz dot com>, java at gcc dot gnu dot org
- Date: Tue, 4 Dec 2001 15:48:44 -0500 (EST)
- Subject: RE: status of gcj's boehm collector?
On Tue, 4 Dec 2001, Boehm, Hans wrote:
> For single-threaded applications that allocate lots of short-lived
> wrappers like yours, it would be interesting to try the generational
> collector.
Like Adam, we run an interpreter that creates a lot of garbage for each
collection. I experimented a bit with incremental mode on Solaris. In
brief, there was really no benefit other than shorter pause times;
however I didn't test with more than about 10MB of long-lived objects.
(That wasn't totally unexpected since marking is by far the most expensive
phase of garbage collection.)
Larger data sets certainly could have made the outcome different, but
frankly that was all the real data our customer had at the time, and I
haven't bothered to generate a synthetic data set.
In conclusion, the GC performance from boehm-gc was more than adequate for
our interpreted application, and we don't require incremental mode at all.
I'm not disappointed at all; that was good news for us.
Jeff