This is the mail archive of the
java@gcc.gnu.org
mailing list for the Java project.
Re: Question on GCJ/Boehm Memory Utilization
- From: Rutger Ovidius <r_ovidius at eml dot cc>
- To: "Craig A. Vanderborgh" <craigv at voxware dot com>
- Cc: java at gcc dot gnu dot org
- Date: Sun, 19 Feb 2006 11:36:00 -0800
- Subject: Re: Question on GCJ/Boehm Memory Utilization
- References: <43F8107B.9060209@voxware.com>
- Reply-to: Rutger Ovidius <r_ovidius at eml dot cc>
Saturday, February 18, 2006, 10:30:19 PM, you wrote:
CAV> Hello!
CAV> The character arrays created by GCJ/Boehm for storing the
CAV> java.lang.String tokens that result from XML processing are where I
CAV> believe there is a problem. These are short strings, typically around
CAV> 10 or so characters in length. GCJ/Boehm seems to be allocating
CAV> 2048-byte heap blocks, and in many cases - even 12932-byte heap blocks,
CAV> and then putting only one or two of the tokens resulting from XML
CAV> parsing into them. Just a few characters. In some cases, 2 of these
CAV> short tokens share a 2048-byte or 12932-byte block, but very often they
CAV> don't.
You might try playing with the HBLKSIZE define in gc_priv.h.
I recall trying 512 (instead of 2048) and finding total ram use to go
down, but having more blocks overall. (On an app with lots of small strings)