Subject: Re: 1.4.2 Observations
To: None <port-i386@netbsd.org,>
From: Thomas Michael Wanka <tm_wanka@earthling.net>
List: port-i386
Date: 03/27/2000 12:46:31
On 27 Mar 2000, at 0:30, Jonathan Stone wrote:

> that aside, a Promise UDMA?66 controller is a whople lot cheaper
> than a SCSI-160 controller, but with a string of IBM LZW glass
> drives tha latter can (I beleive) sustain more. 
> Again, corrections welcome.
> 
I contacted Promise and HPT this weekend for more information. 
For up to the built in ide controller up to Intels BX chipset that would 
depend, (I had no chance to try a 160SCSI combo) for the ide, if 
you run a background tasks that demands 50% CPU time its 
transferrate was slower than that of a scsi chain, for a standard 
"wintel" benchmark where such environmental influences are simply 
not included the ide drives could be faster. I allways read about ide 
drives with transferrates of >20MB/s but none of these drives that I 
could test were able to transfer 1.2GB in one minute, in fact I had 
one or two of these drives in identical machines (running windows) 
and the older and according to the published "benchmarks" slower 
SCSI drives outperformed the ide drives at least to my perception.

> 10,000 RPM high-capacity  IDE drives are, currently, Unobtanium.
> But so are are 1in 40gig SCSI drives.

What I wanted was to get faster drives not larger ones, I currently 
have the old 2GB Conner wide and a 4GB IBM UW drives with three 
operating  systems and they are half full (or half empty). I would pay 
twice the price of a 40GB drive if it was only 4GB but 10 times as 
fast.

mike