From owner-freebsd-current  Wed Mar 18 18:00:49 1998
Return-Path: <owner-freebsd-current@FreeBSD.ORG>
Received: (from majordom@localhost)
          by hub.freebsd.org (8.8.8/8.8.8) id SAA16274
          for freebsd-current-outgoing; Wed, 18 Mar 1998 18:00:49 -0800 (PST)
          (envelope-from owner-freebsd-current@FreeBSD.ORG)
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.19])
          by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id SAA16243;
          Wed, 18 Mar 1998 18:00:26 -0800 (PST)
          (envelope-from bde@godzilla.zeta.org.au)
Received: (from bde@localhost)
	by godzilla.zeta.org.au (8.8.7/8.8.7) id MAA04545;
	Thu, 19 Mar 1998 12:59:00 +1100
Date: Thu, 19 Mar 1998 12:59:00 +1100
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199803190159.MAA04545@godzilla.zeta.org.au>
To: mrcpu@cdsnet.net, sos@FreeBSD.ORG
Subject: Re: Stream_d benchmark... Wow, there really are differences in hardware
Cc: current@FreeBSD.ORG
Sender: owner-freebsd-current@FreeBSD.ORG
Precedence: bulk
X-Loop: FreeBSD.ORG

>Hmm, Then I should be proud of my noname system (p6/200/128MB 72pEDO):
>
>Function      Rate (MB/s)   RMS time     Min time     Max time
>Copy:         117.0286       0.2758       0.2734       0.2812
>Scale:        117.0286       0.2782       0.2734       0.2812
>Add:          130.7234       0.3784       0.3672       0.4141
>Triad:        125.3878       0.3917       0.3828       0.4219
>
>So what ??

So someone forgot to ensure that the doubles are aligned.  This requires
using -malign-double and fixing ld so that it actually works.
-malign-double -static gives perfect misalignment here:

0000b070 D _edata
			<- padding not expected by gcc
0000b074 b _a
00191a74 b _b
00318474 b _c
0049f5d4 B _errno

I fudged alignment by adding `static int x;' before the declaration of
a[], and NOT using -malign-double.
 
>> I figured that within reason, most mb's would have similar performance,
>> but I was wrong.
>> 
>> All boxes are P6-200's, 256MB RAM (all RAM is 60ns FP as far as I know).
>> 
>> Box 1 is a SuperMicro P6DNE:
>> Function      Rate (MB/s)   RMS time     Min time     Max time
>> Copy:          60.7395       0.2704       0.2634       0.2832
>> Scale:         60.8640       0.2719       0.2629       0.3097
>> Add:           70.9688       0.3508       0.3382       0.3602
>> Triad:         71.1647       0.3494       0.3372       0.3565

This seems a bit slow, even with misaligned doubles.

Bruce

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message