PureBytes Links
Trading Reference Links
|
On win2000 they said that one drive the secondary would say PIO when the
change was implemented. My drivers say that it is running in UDMA mode. I
just bought a UDMA ATA100 PCI card that boasts it is faster than the
resident IDE ATA100 system. I'll find out shortly.
Jim Bronke
Phoenix, AZ
----- Original Message -----
From: "Barry Burgess" <bburgess@xxxxxxxxxxxxx>
To: "Bob Scott" <rlscott@xxxxxxxx>; "Jim Bronke" <jvbronke@xxxxxxxx>; "Omega
List" <omega-list@xxxxxxxxxx>
Sent: Wednesday, May 30, 2001 9:21 PM
Subject: Re: HD access times
>
>
> > Your Welcome Bill.
> > Thanks goes to Bill Wynne for getting this thread started. It alerted me
> to
> > the weakness that had existed in Win2000. I've just received a response
to
> a
> > query to my motherboard manufacture(IWill) and find out that it has a BX
> > chipset which only supports ATA33. Seems that you have to read between
the
> > lines when you buy these boards. They talk about their components like
> they
> > are the latest and greatest but then you find out they are not. I'm
going
> to
> > try to add an add on IDE driver board that will get the job done.
>
> It does not matter much if the board is ATA 33 or 66. The main problem is
> getting your disk to be read is UDMA rather than PIO
> My original tests are below and I was at a loss to figure out why the P700
> was so slow
>
> I tested the following on a number of systems
> Read & write mode mode: 65536 blocks - Test file size: 10Mb
> CPU Op Sys ATA read write
> P500 Win2000 66 19284 13281
> P700 NTSP6 66 3385 2662
> P400 Win98SE 33 5234 5389
> P350 Win98SE Old 5130 4980
> P333 NTSP6 33 3256 2228
>
> Here are the new results on the P700 changing to UDMA
> P700 NT SP6 66 10882 9650
> This is greater than a 220% improvement
> Thanks to Bob Perry
> I found the following site that was useful
> http://www.benchtest.com/nt_udma1.html
> It is not just the improved I/O but the reduction on CPU utilization
> Thanks to all that got me started on this track - it has made a big
> improvemnt.
>
>
>
>
|