From owner-freebsd-current  Fri Nov 15 09:03:26 1996
Return-Path: owner-current
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.5/8.7.3) id JAA29793
          for current-outgoing; Fri, 15 Nov 1996 09:03:26 -0800 (PST)
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.19])
          by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id JAA29767
          for <freebsd-current@FreeBSD.org>; Fri, 15 Nov 1996 09:03:15 -0800 (PST)
Received: (from bde@localhost) by godzilla.zeta.org.au (8.7.6/8.6.9) id DAA19190; Sat, 16 Nov 1996 03:55:55 +1100
Date: Sat, 16 Nov 1996 03:55:55 +1100
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199611151655.DAA19190@godzilla.zeta.org.au>
To: cracauer@wavehh.hanse.de, freebsd-current@FreeBSD.org
Subject: Re: EDO vs. other (Re: GigaByte GA-586DX-512 Motherboard)
Sender: owner-current@FreeBSD.org
X-Loop: FreeBSD.org
Precedence: bulk

>I just ran that famous RAM benchmark that came down -hackers some time
>ago:

Infamous? :-)

>normal parity RAM (8x 32 MB installed)
>49005fb0   0.160 uS/op 6.24e+06 op/S 23.805 Mb/S
>8938c0df   0.369 uS/op 2.71e+06 op/S 10.338 Mb/S
>49005fb0   0.160 uS/op 6.24e+06 op/S 23.806 Mb/S
>8938c0df   0.369 uS/op 2.71e+06 op/S 10.337 Mb/S
>
>EDO w/o Parity (2x 16 MB installed)
>49005fb0   0.159 uS/op 6.29e+06 op/S 23.999 Mb/S
>8938c0df   0.366 uS/op 2.73e+06 op/S 10.421 Mb/S
>49005fb0   0.159 uS/op 6.29e+06 op/S 23.999 Mb/S
>8938c0df   0.366 uS/op 2.73e+06 op/S 10.422 Mb/S

I forget exactly what this does.  I guess it does a lot of random
accesses.  Random access is not the strongest point of RAM :-).

>I don't beleive one can speed up any real application using EDO.

It speeds up copying of large amounts of data in the kernel by
5-10% (from 75MB/sec to > 80MB/sec for uncached memory on ASUS
Triton 1 P133 systems).  This might show up in real applications
that read a lot of data from fast disks.

Bruce