This is the mail archive of the
java-patches@gcc.gnu.org
mailing list for the Java project.
Re: PATCH: interpreter.cc bug patch and performance improvements
- From: Tom Tromey <tromey at redhat dot com>
- To: Chris Sears <cbsears_sf at yahoo dot com>
- Cc: java-patches at gcc dot gnu dot org
- Date: 11 Jan 2002 17:01:32 -0700
- Subject: Re: PATCH: interpreter.cc bug patch and performance improvements
- References: <20020108180050.55701.qmail@web14702.mail.yahoo.com>
- Reply-to: tromey at redhat dot com
>>>>> "Chris" == Chris Sears <cbsears_sf@yahoo.com> writes:
Chris> Run the cm3 benchmark on a stock 3.0.3. The performance will be
Chris> 644. Comment out the insn_nop block and move the insn_nop
Chris> label in front of the first NEXT_INSN. Very simple, very dumb
Chris> optimization.
Doing this isn't really that important.
If you are concerned about the minor code duplication taken by the nop
processing, you can put `insn_nop:' before any other NEXT_INSN. Try
folding it into insn_aconst_null.
Are `nop's common in Java bytecode? I think `gcj -C' never generates
them. I don't know about other compilers.
Anyway, if this gives you some kind of improvement, I'll check it in
even though it doesn't seem all that important.
Tom