🚀 go-pugleaf

In article <46ai5g$2e2@news-e1a.megaweb.com>, BrianChen@GNN.COM (Brian K. Chen) writes:
|> triumph@ix.netcom.com (Gregory Patten) wrote:
|> 
|> >This is all true, but due to decreased overhead, Windows runs faster
|> >with IDE drives. So it depends on what you're doing. BTW, it shouldn't
|> >cosst him $400 extra, if he's willing to use and NCR SCSI controller.
|> >Perhaps $100 for the controller and $100 extra for a 1 gig HD. Of
|> >course if you need 2 gig or more of storage in a single drive, you have
|> >to go SCSI.
|> 
|> Are you talking about the SCSI Command Overhead?  If you get the
|> Adaptect SCSI Adepter that came with a RISC processor on board, it
|> will even speed every things up.

Not more than the NCR, which also got a fast 
processor integrated into the controller chip.

I've got several benchmark results from fast
SCSI drives (Atlas and Barracudas) on both 
controllers. 
They are within a few percent of each other:

              -------Sequential Output-------- ---Sequential Input-- --Random--
              -Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks---
Machine    MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU  /sec %CPU
486/66    100  1372 97.7  5856 91.4  2420 53.8  1434 98.5  6152 82.8  85.9 15.8
P90       100  3077 96.9  6193 56.2  2445 31.7  3374 96.7  6377 52.4  78.0  7.1
P100NCR   140  1543 98.6  5150 40.3  1711 16.5  1808 98.6  5211 26.8  98.1  7.1
P100AHC   140  1584 98.9  3138 20.5  1707 19.3  1850 98.8  5209 25.3  97.6  6.1

All except the last are NCR 53c810 based 
systems. The last 4th is identical to the 
3rd, except for the 2940 card ...

Regards, STefan
-- 
 Stefan Esser, Zentrum fuer Paralleles Rechnen		Tel:	+49 221 4706021
 Universitaet zu Koeln, Weyertal 80, 50931 Koeln	FAX:	+49 221 4705160
 ==============================================================================
 http://www.zpr.uni-koeln.de/staff/esser/esser.html	  <se@ZPR.Uni-Koeln.DE>