Uniforum unix-box benchmarks

[vasilatos]

I ran more of Teus Hagan's compute and disk benchmarks 
from San Diego on assorted boxes at the /Vendor/Group Circus in D.C.
last week. Times (per the table he published last year), are given 
below. I was unfortunately less than rigorous about noting particulars
of cpu model and operating frequency in each case. I think they are 
more or less uniquely determined by system model. Neither did I
look in their cabinets to see what sort of disk drives had been 
installed for the show.

					Nick Vasilatos
					(wjh12!boxer)


----------------------------------------------------------------------------------------
		usr	usr	usr	usr	usr	usr	real	real	real
		short	reg-sh	int	reg-int	long	reg-lng	cc#1	cc#2	run#2
----------------------------------------------------------------------------------------

Intel 380/286
(80286)	(0)	16.3	10.7	16.4	10.7	38.7	38.8	16.0	16.0	4.0

TI Nu
(68010)		14.9	8.0	18.8	7.8	18.8	7.8	15.0	15.0	5.0

HP 9000/500
(HP/prop.) 	9.7	9.7	8.0	8.0	8.0	8.0	9.7	11.3	2.1

Spectrix 10
(68010) (1)	13.2	8.3	14.9	8.1	14.9	8.1	18.0	18.0	10.0

Gould 1000
(80186)	(2)	27.9	18.3	27.6	18.4	na	na	18.0	19.0	14.0

Gould 2000
(68010)	(3)	13.3	8.8	13.9	7.1	13.8	7.1	15.9	15.4	32.5	

NS SYS16
(16000)	(4)	25.4	25.4	30.4	12.5	30.5	12.4	12.0	14.0	6.0	

__________________________________________________

(0) Not at the show, found laying about elsewhere. The cpu is an 8 mhz 286.

(1) I think the cpu on the Spectrix 10 is a 68010. I thought their glossy
identified it, but it doesn't.

(2) Time reporting on this system has definite problems. Na times above were
wildly out of range of real (observed/reported) times for the tests.

(3) The time for the disk benchmark is very suspicious.
If it is right, the disk is either the slowest in the east or the os is 
most brain damaged in the west.

(4) Again the glossy doesn't identify the specific cpu. There was other 
activity on the system (slight), about which there was nothing I could do.