Message-ID:	<19990521130302.45592@atrey.karlin.mff.cuni.cz>
Date:	Fri, 21 May 1999 13:03:02 +0200
From:	Jan Hubicka <hubicka AT atrey DOT karlin DOT mff DOT cuni DOT cz>
To:	pgcc AT delorie DOT com
Subject:	Re: Benchmark PGCC vs EGCS on a K6-2
References:	<373F3AA2 DOT A446D611 AT informatik DOT hu-berlin DOT de> <Pine DOT LNX DOT 4 DOT 10 DOT 9905181826020 DOT 1284-100000 AT data DOT mandrakesoft DOT com> <19990519105631 DOT 40676 AT atrey DOT karlin DOT mff DOT cuni DOT cz> <3743ADE8 DOT C938ADBB AT informatik DOT hu-berlin DOT de> <19990520104723 DOT 18724 AT atrey DOT karlin DOT mff DOT cuni DOT cz> <374515C3 DOT DD91DD72 AT informatik DOT hu-berlin DOT de>
Mime-Version:	1.0
X-Mailer:	Mutt 0.84
In-Reply-To:	<374515C3.DD91DD72@informatik.hu-berlin.de>; from Jens-Uwe Rumstich on Fri, May 21, 1999 at 08:13:55AM +0000
Reply-To:	pgcc AT delorie DOT com
X-Mailing-List:	pgcc AT delorie DOT com
X-Unsubscribes-To:	listserv AT delorie DOT com

> Hi!
> 
> > > ooops... The mov is not faster?
> > >
> > Only advantage of mov is reduced dependency on flags. But this advantage is not
> > high enought to mask code size increase and decoder slowdown caused by very
> > large opcode.
> 
> But a MOV REG,Immediate is a short instruction and gets executed during
> the
> decode phase. So except the larger opcode it should be faster...
> Strange, I
> think I should try it out...
According to my tests, execution delay of xor reg,reg is not problem, because
K6 decodes much slower than executes the GCC code.
For some purpose mov reg,0 seems to decode slowly. I've done some benchmarks
and at least at my friend's computer it don't behave like short decodable
instruction, but as long decodable. This is pretty strange.
Maybe this depends on K6 chip version. I've seen quite large differences
between various K6...

Honza

- Raw text -