Re: Apple II in Scientic American

[Roger Johnstone <rojaws@orcon.net.nz>]

In <1109852926.317322.103840@g14g2000cwa.googlegroups.com> anoneds@
netscape.net wrote:
> 
> Roger Johnstone wrote:
>> In <d057mq$m05$1@nntp.itservices.ubc.ca> Ben Lai wrote:
>> > I was reading the January issue of Scientific America on the plane
> (
>> > only thing available) and came across a column about robots, and 
>> > they
>> > have a little chart comparing the "raw computing muscle" of all the
>> > top machines of their time.  You can see it here:
>> >
>> > http://www.sciam.com/article.cfm?chanID=sa006&colID=30
>> > &articleID=00078A55-0CE7-11BF-AD0683414B7F0000
>> > (Just click on the "More on this article/SIDEBAR" part)
>> >
>> > First thing I noticed was the huge gap between the Apple II and
>> > Commodore 64 - considering that they had basically the same 
>> > processor
>> > that's a bit odd.  And then the Playstation2 being barely slower 
>> > than
>> > a Dual G5, also a  bit out of touch.
>> >
>> > Oh well...
>> >
>> > Ben
>>
>> Yeah, someone screwed up the Apple II position on that graph. The C64 
>> is
>> in about the right position, but the Apple II is shown as running at 
>> one
>> tenth the number of instructions per second! A mean feat considering 
>> they
>> use the same CPU running at the same clock speed.
>>
> 
> Has anyone run any tests to generate the numbers used for this chart?
> Seems to me that the C64 is too high on the graph, not that the Apple
> is too low. Basically, the graph looks like BS to me anyway. But a C64
> really shouldn't be that close to a PC of that era, and neither one
> belongs in the same area as a VAX 11/780. Having worked on one during
> that timeframe, and having a C64 to compare to it, they are very
> different beasts.
> 
> I wonder how much money was poured into such worthless research...
> 
> Ben

The chart may be right if it's just comparing speed in instructions per 
second. That's why I think the C64 is in the right position. A 6502 
instruction takes between two and six cycles to execute. At 1MHz that 
gives a 'thoretical peak' speed of 500000 instructions per second. The 
Apple II should be in the same position, as should the original IBM PC 
since from memory the 8088 took a minimum of eight cycles for each 
instruction.

Of course measuring the speed of a CPU by the rate of instruction 
execution is pretty meaningless, since it doesn't take into account how 
many actions the instruction is performing or how much data it's moving 
around.

-- 
Roger Johnstone, Invercargill, New Zealand
http://vintageware.orcon.net.nz/
________________________________________________________________________
  No Silicon Heaven? Preposterous! Where would all the calculators go?

              Kryten, from the Red Dwarf episode "The Last Day"