This is the mail archive of the
java@gcc.gnu.org
mailing list for the Java project.
RE: Correction to last message, GC status
- To: "Boehm, Hans" <hans_boehm at hp dot com>
- Subject: RE: Correction to last message, GC status
- From: Jeff Sturm <jsturm at one-point dot com>
- Date: Tue, 26 Jun 2001 15:20:37 -0400 (EDT)
- cc: "'tromey at redhat dot com'" <tromey at redhat dot com>, java at gcc dot gnu dot org
On Tue, 26 Jun 2001, Boehm, Hans wrote:
> I suspect that we'll also see a measurable performance increase from linking
> statically linking the GC library against libgcj. Allocation calls from
> inside libgcj (e.g. for strings) will no longer cross a dynamic library
> boundary.
That's true. I'm wondering though why that should have any effect on
performance. For PIC code, non-static procedures will be called through
the PLT in either case. I don't see anything special about the dynamic
library boundary besides sparse virtual addresses.
OTOH, systems that prelink dynamic libraries (e.g. win32) should benefit
from combining the libraries.
Jeff