[comp.windows.ms] Norton SI differences

dve@mace.cc.purdue.edu (zhou) (01/19/91)

	Several mesgs on the net lately use Norton SI index to compare
	speed loss in enchanced mode. This brings some doubt in my mind.

	I have Norton 5.0 and a 16-SX with 4Meg (70nsec) Ram on board
	and a 15msec Hard Drive. Yet the SI index of my machine is about
	10.5 in DOS and 8.5 in Windows (Enchanced mode) while other people
	have 15 plus in dos on SX laptop and 22 on IBM 386-20. Surely my
	machine can't be that slow. It's Gateway 2000 machine which should
	be an average perfomer. So I suspect the Norton 4.5 and 5.0 give
	different numbers.

	Can somepople confirm this?

	Joe

daly@ecs.umass.edu (Bryon Daly, ECE dept, UMass, Amherst) (01/20/91)

In article <1991Jan20.064216.10926@nmt.edu>, wasnsr@nmt.edu (T.O.R.S.O.) writes:
> In article <6627@mace.cc.purdue.edu> dve@mace.cc.purdue.edu (zhou) writes:
>>
>>	Several mesgs on the net lately use Norton SI index to compare
>>	speed loss in enchanced mode. This brings some doubt in my mind.
>>
>>	I have Norton 5.0 and a 16-SX with 4Meg (70nsec) Ram on board
>>	and a 15msec Hard Drive. Yet the SI index of my machine is about
>>	10.5 in DOS and 8.5 in Windows (Enchanced mode) while other people
>>	have 15 plus in dos on SX laptop and 22 on IBM 386-20. Surely my
>>	machine can't be that slow. It's Gateway 2000 machine which should
>>	be an average perfomer. So I suspect the Norton 4.5 and 5.0 give
>>	different numbers.
>>
>>	Can somepople confirm this?
> 
> 	I am working on a 386-20 DX and the SI from 4.5 gives me an index
>    of 21 without windows and a 16.5 with windows, while norton 5.0 SI gives
>    me 12.3 without windows and 9.0 inside windows.
> 
> 	Have the Norton people changed their rating system??
> 
> 				wayne,
> 				 wasnsr@titan.nmt.edu

According to the Norton 5.0 manual, they have made changes in their benchmark,
with the newer benchmark generally giving smaller values than the old one.
(I think they were trying to legitimize their infamous benchmark by making it
more accurate.)
The CPU index I get on my 386-33 are as follows:
               Ver 4.5                 Ver 5.0
               _______                 _______
Plain DOS:      40.5                    34.7

386 Enhanced:   34.3                    23.6 *

Std mode:       40.3                    34.7

*Interestingly enough, the Ver 5.0 CPU index is a continuous sampling one (i.e,
if you move the mouse, you can see the number drop, then rise when you stop
movement) and the index varies as you watch, in 386 Enhanced mode (with no 
user action) (Fun to watch! :-).  But I have seen the number vary from as low
as 4.1 to an incredible 260.1 (I'm not lying!  It went off the scale a few
times!  Of course, something about windows is screwing up the benchmark)
The number above of about what is shown most of the time.

Note: the programs were not run from PIF's, but as straight executables from
the file manager.  In the background were no really "active programs"; eyes,
4dos shell, winpost were iconized, winclock was running, and so was fileman.

-Bryon Daly
daly@ecs.umass.edu