From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-return-92651-listarch-gcc=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 6127 invoked by alias); 15 Mar 2004 18:22:48 -0000
Mailing-List: contact gcc-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Archive: <http://gcc.gnu.org/ml/gcc/>
List-Post: <mailto:gcc@gcc.gnu.org>
List-Help: <http://gcc.gnu.org/ml/>
Sender: gcc-owner@gcc.gnu.org
Received: (qmail 6115 invoked from network); 15 Mar 2004 18:22:47 -0000
Received: from unknown (HELO scanner2.ics.uci.edu) (128.195.1.36)
  by sources.redhat.com with SMTP; 15 Mar 2004 18:22:47 -0000
Received: from vino.ics.uci.edu (vino.ics.uci.edu [128.195.11.198])
	by scanner2.ics.uci.edu (8.12.10/8.12.10) with ESMTP id i2FILQ4J020367;
	Mon, 15 Mar 2004 10:21:27 -0800 (PST)
Message-Id: <200403151821.i2FILQ4J020367@scanner2.ics.uci.edu>
To: Stelios Xanthakis <sxanth@ceid.upatras.gr>
Cc: Roger Sayle <roger@eyesopen.com>, gcc@gcc.gnu.org
Subject: Re: GCC viciously beaten by ICC in trig test!
References: <Pine.GSO.4.21.0403151400100.29891-100000@zenon.ceid.upatras.gr>
From: Dan Nicolaescu <dann@godzilla.ics.uci.edu>
In-Reply-To: <Pine.GSO.4.21.0403151400100.29891-100000@zenon.ceid.upatras.gr>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Date: Mon, 15 Mar 2004 18:22:00 -0000
X-ICS-MailScanner: Found to be clean
X-ICS-MailScanner-SpamCheck: not spam (whitelisted),
	SpamAssassin (score=-96.69, required 5, MSGID_FROM_MTA_SHORT,
	USER_IN_WHITELIST)
X-SW-Source: 2004-03/txt/msg00732.txt.bz2

Stelios Xanthakis <sxanth@ceid.upatras.gr> writes:

  > On Sun, 14 Mar 2004, Dan Nicolaescu wrote:
  > 
  > > Roger Sayle <roger@eyesopen.com> writes:
  > >   >         fsin
  > >   >         fmul    %st(0), %st
  > > 
  > > Intel 8.0 (that was used in the original test) generates something
  > > very different:

Please be careful when snipping, the essential part that you deleted
is this: 


        call      __libm_sse2_sincos                            #7.15
                                # LOE ebp esi edi xmm0 xmm1
..B1.4:                         # Preds ..B1.1

i.e. ICC 8 generates a call to an SSE library function instead of
using the fsin instruction. Given that this changed from ICC 7 to ICC
8, the library function is probably faster. 

  > >         mulsd     %xmm1, %xmm1                                  #10.25
  > >         mulsd     %xmm0, %xmm0                                  #10.15
  > >         addsd     %xmm1, %xmm0                                  #10.25
  > >         movsd     %xmm0, (%esp)                                 #10.25
  > >         fldl      (%esp)                                        #10.25
  > > 
  > 
  > Does --fpmath=sse fix this?
  > Can the processor in question do sse for doubles?
  > 
  > In my experience, "--fpmath=sse --fsingle-precision-constants"
  > generates much faster code for a raytracer I have here.

See above.