I've been tossing this over for a week, as I'd like to get at least some comparison of performance between various processors in this range, as well as being able to compare them to later processors from the G3/G4 series.
I'd decided that Mathematica 3.0 would be a good choice. It's a math intensive program(since that's essentially all it does). I chose 3.0 for two reasons-1. It should run on all the processors I want t compare, and 2. More importantly, I have a retail boxed copy of it. Wolfram-the company that makes Mathematica-and Apple actually have a bit of a long relationship, as Steve Jobs actually suggested the name Mathematica to Steve Wolfram.
In any case, although I though this would be a decent benchmarking program, I struggled with coming up with something to throw at it that would take long enough to calculate that I could actually reliably measure it, without taking an excessive amount of time.
I think I've hit on a good combination with the below. Basically, it's calculating Leibniz's series to n=10,000. Leibniz's series is an aproximation of Pi. The below took about 10-12 minutes(I didn't time it for the first try) on my 9600/200MP.
I'm sure whether or not this version of Mathematica is multiprocessor aware. If it is, it should take roughly twice as long 200mhz 8600-I intend to test that.
As I said, I also want to look at G3 and G4 series processors. I suspect that sheer clock speed will help them out, although I don't think a G4 will give much(if any) benefit over a similarly clocked G3 since this version of Mathematica most certainly can not use Altivec(it was released in 1996).
In any case, below is a screenshot of the series I used. I'm going to try it out a couple of times-with the timer running-on several different computers and will report here.
I'd decided that Mathematica 3.0 would be a good choice. It's a math intensive program(since that's essentially all it does). I chose 3.0 for two reasons-1. It should run on all the processors I want t compare, and 2. More importantly, I have a retail boxed copy of it. Wolfram-the company that makes Mathematica-and Apple actually have a bit of a long relationship, as Steve Jobs actually suggested the name Mathematica to Steve Wolfram.
In any case, although I though this would be a decent benchmarking program, I struggled with coming up with something to throw at it that would take long enough to calculate that I could actually reliably measure it, without taking an excessive amount of time.
I think I've hit on a good combination with the below. Basically, it's calculating Leibniz's series to n=10,000. Leibniz's series is an aproximation of Pi. The below took about 10-12 minutes(I didn't time it for the first try) on my 9600/200MP.
I'm sure whether or not this version of Mathematica is multiprocessor aware. If it is, it should take roughly twice as long 200mhz 8600-I intend to test that.
As I said, I also want to look at G3 and G4 series processors. I suspect that sheer clock speed will help them out, although I don't think a G4 will give much(if any) benefit over a similarly clocked G3 since this version of Mathematica most certainly can not use Altivec(it was released in 1996).
In any case, below is a screenshot of the series I used. I'm going to try it out a couple of times-with the timer running-on several different computers and will report here.
