Date: Mon, 18 Oct 1999 21:52:03 +0200
From: Marc Lehmann <pcg AT opengroup DOT org>
To: pgcc AT delorie DOT com
Subject: Re: using cmov* not that great
Message-ID: <19991018215203.B612@cerebro.laendle>
Mail-Followup-To: pgcc AT delorie DOT com
References: <3 DOT 0 DOT 32 DOT 19990926231125 DOT 00c757a0 AT pop DOT xs4all DOT nl> <380A4183 DOT B3794352 AT lycosmail DOT com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <380A4183.B3794352@lycosmail.com>; from Adam Schrotenboer on Sun, Oct 17, 1999 at 05:37:08PM -0400
X-Operating-System: Linux version 2.2.12 (root AT cerebro) (gcc version 2.95.1 19990816 (release)) 
Sender: Marc Lehmann <pcg AT goof DOT com>
Reply-To: pgcc AT delorie DOT com
X-Mailing-List: pgcc AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk

On Sun, Oct 17, 1999 at 05:37:08PM -0400, Adam Schrotenboer <ajschrotenboer AT lycosmail DOT com> wrote:
> I can not thing of any (intel x86 at least) cpu that can use such.

Sure it can. One can reorder code that it suits static prediction.

> Itanium (stupid name, Merced was better) however I do believe will be

even better, it can just execute the if branch speculatively and then use
the equivalent of cmov/cadd. calculating two much in many cases hurts less
than a single jump.

-- 
      -----==-                                             |
      ----==-- _                                           |
      ---==---(_)__  __ ____  __       Marc Lehmann      +--
      --==---/ / _ \/ // /\ \/ /       pcg AT opengroup DOT org |e|
      -=====/_/_//_/\_,_/ /_/\_\       XX11-RIPE         --+
    The choice of a GNU generation                       |
                                                         |