blf@wookumz.ai.mit.edu (Bill Frandsen) (03/19/91)
For Sale:
---------
CMI (Creative Microsystems Inc.) Processor Accelerator
Consists of a PC board with a 16MHz 68000 chip, and a 16MHz 68881 math
coprocessor with crystal. To install, you simply remove the 7MHz 68000
from your Amiga 2000 (only) and instal this board in it's place. The
board runs at either the regular 7.14MHz for full compatibility, or at
14.28MHz for improved performance. Software is included to switch
between both speeds at will. A jumper is also provided to force the
board to boot at either speed. (I usually ran the board at 7MHz at
boot time, and ran the PA program in the startup sequence to install
the math chip and run at 14MHz). A disk full of software (useful and
demo) is included.
Note: The CMI company is now out of business, but I have been told
thru USENET that the company DigiF/X or something like that is now
handling the support of the board. Also, due to a problem with certain
DMA devices, the board will not work at 14MHz with the Bridgeboard. The
speed must be set at 7MHz before it will work with the Bridgeboard.
(Supposedely this can be fixed by cutting a trace on the PA board, call
this DigiF/X company to find out how)
The real speedup is about 10% for most programs like LHARC/ZOO that can
be measured. Speed up for specific programs that can use the ffp.library
are in the range of 20 times! (A demo included runs at 50 seconds with
7Mhz, and 2 seconds with the 14MHz math coprocessor installed). However,
most programs do not access this library.
Below is the information provided by Kramden SI 2.0b:
System Information v2.0b
Kramden Utilities (c) 1988,1989 Pregnant Badger Software
Main Processor: Motorola 68000
Math Peripheral: Motorola 68881
Computing Performance relative to:
A1000 IBM/XT
Integer: 1.8 5.9
Floating Point: 3.2 24.2
Reason for selling: Getting a much faster (more expensive) accelerator!
Price: $100 US Dollars, which I will include shipping to anywhere in the
continental United States. (US Postal Service Priority Mail).
Terms: Check or money order. I will ship upon receipt of payment.
Contact: blf@churchy.ai.mit.edujdickson@jato.jpl.nasa.gov (Jeff Dickson) (03/20/91)
>CMI (Creative Microsystems Inc.) Processor Accelerator > >Computing Performance relative to: > > A1000 IBM/XT > Integer: 1.8 5.9 > Floating Point: 3.2 24.2 > >Contact: blf@churchy.ai.mit.edu The Amiga's computing performance cannot be compared to that of the IBM PC. The whole time the Amiga is cranking out these numbers, it is sharing the CPU amoung several processes, tasks, devices, etc. You could only produce accurate results if the Amiga were "dry" (not running anything else). -jeff
swarren@convex.com (Steve Warren) (03/20/91)
In article <1991Mar19.185203.3096@jato.jpl.nasa.gov> jdickson@jato.Jpl.Nasa.Gov (Jeff Dickson) writes:
##CMI (Creative Microsystems Inc.) Processor Accelerator
##
##Computing Performance relative to:
##
## A1000 IBM/XT
## Integer: 1.8 5.9
## Floating Point: 3.2 24.2
##
##Contact: blf@churchy.ai.mit.edu
#
# The Amiga's computing performance cannot be compared to that of the
#IBM PC. The whole time the Amiga is cranking out these numbers, it is sharing
#the CPU amoung several processes, tasks, devices, etc. You could only produce
#accurate results if the Amiga were "dry" (not running anything else).
#
# -jeff
You missed the point, Jeff. These are not benchmark results showing
performance results on the 1000 vs a PC/XT. These are the performance
results of an accelerator, showing how much faster it is than the given
machines.
ie it is 1.8 times faster than a stock 1000, and it is 5.9 times faster
than a PC/XT. So calm down. If you will just do the math, this is
indicating that the 1000 is 3 times faster than the XT.
_.
--Steve ._||__ DISCLAIMER: All opinions are my own.
Warren v\ *| ----------------------------------------------
V {uunet,sun}!convex!swarren; swarren@convex.com
--