X-pop3-spooler: POP3MAIL 2.1.0 b 4 980420 -bs-
From: Dinesh Somasekhar <somasekh AT ecn DOT purdue DOT edu>
Date: Fri, 10 Jul 1998 03:08:05 -0500 (EST)
To: beastium <beastium-list AT Desk DOT nl>
Subject: Re: please benchmark / MMX #2
In-Reply-To: <19980708183649.14799@cerebro.laendle>
References: <19980708045815 DOT 20996 AT cerebro DOT laendle>
	<13731 DOT 20100 DOT 256252 DOT 407869 AT ppp-x9-13.>
	<19980708183649 DOT 14799 AT cerebro DOT laendle>
X-Mailer: VM 6.43 under 20.4 "Emerald" XEmacs  Lucid
Message-ID: <13733.51867.888722.918209@ppp-x8-34.>
Reply-To: somasekh AT ecn DOT purdue DOT edu
X-URL: http://albrecht.ecn.purdue.edu/~somasekh
X-Organization: Purdue University
X-face:  57,u^#nsk>2N2INb(Afe~^~B}j[t7+B)>u]%DYS:\XA8yL6?[*<0"yBBb@*u/e AT K#F6?Fi5
 WB?GMJvJPt$5W0bRNZ{dH%ybQO%$]M)b~_G[!*0"[82&2t[3o{EQB87/0J AT AtN6@cz1zM_FD1~{maV
 ;qd6T/s(+&j&QJ"zv5'GQU%10P\R)}aW1xq3(;U<M}8W^~Y:4M9cnj"M+R9_FKM%E[HEyle7['_!|i
 (
Mime-Version: 1.0 (generated by tm-edit 7.108)
Content-Type: text/plain; charset=US-ASCII
Sender: Marc Lehmann <pcg AT goof DOT com>
Status: RO
X-Status: A
Content-Length: 1523
Lines: 34

Hi,

  Marc, can you tell the list what exactly does the mmx patch do ?  Does
  it just use the mmx registers as temp space (at least thats how it looks)
  or does it use mmx compute instructions. 
  
"Marc" == Marc Lehmann <pcg AT goof DOT com> writes:
 > On Wed, Jul 08, 1998 at 06:04:38AM -0500, Dinesh Somasekhar wrote:
 >> 
 >> Cyrix 6x86mx  200MHz ( 100MHz x 2 ), 64M time(user)
 >> 
 >> bzip2-* -9 < /bin/bash > /dev/null

 > According to the cyrix databook (that cyrix was so kind to send to me, and
 > actually convinced me to write the mmx patch, compared with intel and amd,
 > who both said "check the web!", well...), mmx->reg transfers have a 5 cycle
 > latency, which might be one reason why the cyrix performance is so much
 > worse. at leats, emms seems to be fast ;()

  Peformance degradation seems pretty bad.  Almost 20%.  It is surprising
  that the 6x86mx 200MHz ran the test on bash in 1.54sec  as opposed 
  to a K6 266 (which is quite a bit quicker)  running it in 1.55 sec.
  Maybe the bash size is different. Have it here as 279176 bytes.

 >       -----==-                                              |
 >       ----==-- _                                            |
 >       ---==---(_)__  __ ____  __       Marc Lehmann       +--
 >       --==---/ / _ \/ // /\ \/ /       pcg AT goof DOT com       |e|
 >       -=====/_/_//_/\_,_/ /_/\_\                          --+
 >     The choice of a GNU generation                        |
 >                                                           |

--
Dinesh