mwm@eris.BERKELEY.EDU (Mike (My watch has windows) Meyer) (09/09/87)
In article <3625@zen.berkeley.edu> woan@cory.Berkeley.EDU.UUCP (Ronald S. Woan) writes:
<Also if you have the bucks, why not go for '386 system, i.e. one with no
<wait states and a '387 coprocessor; that way you can knock the socks off
<of just about anything. Just look at the benchmarks in Byte.
I don't normally read Byte, but I did take a look at the benchmarks.
They were seriously flawed, having ignored things like the support
hardware, software, data types on the hardware, etc., etc., etc.
I was later told by someone I mentioned this too that Byte pointed
this out in a later issue.
Benchmarking is hard. Expecting to gain a good comparison of two
computers from a single magazine article is hopeless. Even comparing
two systems (not processors, but complete systems, including compilers
and OS) for a single application in a single article is stretching
things.
To reiterate, systems need to be evaluated for proposed use, budget
and envirnoment. Ronald's points about buying what the school in
question is liable to use is good if you're a student. This usually
means either Macs or IBM-PCoid machines. That doesn't necessarily make
them the machine of choice, though.
<mike
--
The handbrake penetrates your thigh. Mike Meyer
A tear of petrol is in your eye. mwm@berkeley.edu
Quick, let's make love before we die. ucbvax!mwm
On warm leatherette. mwm@ucbjade.BITNET