Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 26 Oct 2000 09:25:18 +0100
From:      Steve Coles <scoles@tripos.com>
To:        ArildV@ifi.uib.no
Cc:        freebsd-hardware@FreeBSD.ORG
Subject:   Re: Promise Ultra-ATA100 vs FastTrak 100 Speeds
Message-ID:  <39F7EA6E.A4FCEE20@tripos.com>
References:  <39F59F1C.A31756E8@rasmus.uib.no>

next in thread | previous in thread | raw e-mail | index | archive | help
Thanks for the reply, this would account for the fact that no matter how =
I
configure the drives on the Ultra-100, the speed is always the same withi=
n
the scope of the test ( bonnie results on a UFS for the tests were always
the same from one disk combination to another - only slower ).

Of course, I forgot to mention there was one other variation between the
tests - the controllers were tested in different MOBOs. Ultra in a cheap
PCChips MOBO K6-2@550, FastTrak was in a GigaByte, SIS 5591-based, K6-2@3=
50.
So the noticably slower speeds were in the faster system with the Ultra-1=
00.

The PCChips MOBO in the test has a SIS 5595/530 ATA controller that
~purports~ to be ATA/66, yet the FreeBSD code says it is only ATA/33. I
think that this could be the cause ( the MOBO, not the FreeBSD ata code )

I have tried to run this MOBO controller at ATA/66 with the FastTrak unde=
r
NT there was horrific FS corruption on the FastTrak stripe when both copy=
ing
from a disk on the MOBO controller to the stripe on the FastTrak. Hence I
swapped the FastTrak to the GigaByte MOBO out of paranoia.

I shall check the ATA cabling, and try some other drives that will arrive
soon.

S=F8ren hinted in Stable yesterday that some IBM drives such as yours ( a=
nd
coincidentally mine are the same )

<snip>
The only drives I can recommed are IBM's, but beware the new DTLA series
are so fast that some old controllers can have problems with them.


Steve

Arild Eiken=E6s Vengen wrote:

> Steve Coles wrote:
>
> > There has been a lot of press that these are the same cards, and I
> > seem to remember that FreeBSD treats the FastTrak as just an IDE
> > controller, so can anyone enlighten me on the vast speed differential
> > between the two:
> <snip>
> > Summarised Results
> > ------------------
> > 1) One drive alone (no ccd) reads and writes at 30 Mb/Sec
> > 2) Ultra 100 reads and writes at 29 MB/sec
> > 3) FastTrak 100 writes at 50 MB/sec, reads at 48 MB/sec
>
> > 1) Why is the Ultra so slow ( constructive answers only please :) ?
>
> The only thing I can think of is that the Ultra 100 isn`t utilizing the
> udma/100-channel, and falls back to udma/33, since both reads and write=
s
> are at 29MB/sec witch is just maxing out the udma/33-standard. When
> running my IBM 75GXP off of an udma/33-controller (BX-chipset) I never
> get more than 29MB/sec, never 31 or 32MB/sec.
>
> I seem to remember something about the udma/66-versions from Promise
> (both RAID and non-RAID) was the same card (with a different BIOS), but
> I thought the udma/100-versions had more differences.
>
> Arild.
>
> To Unsubscribe: send mail to majordomo@FreeBSD.org
> with "unsubscribe freebsd-hardware" in the body of the message



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




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?39F7EA6E.A4FCEE20>