[Date Index] [Thread Index] [Author Index]

Re: Is Mathematica on a PC Too Slow?

• To: mathgroup at smc.vnet.net
• Subject: [mg7083] Re: Is Mathematica on a PC Too Slow?
• From: karl at faepds01.tu-graz.ac.at (karl unterkofler)
• Date: Wed, 7 May 1997 01:58:10 -0400 (EDT)
• Organization: TU Graz
• Sender: owner-wri-mathgroup at wolfram.com

In article <5k9fqh$m2c at smc.vnet.net>, Dave Smith <buyguy at erols.com> wrote:

> Dear Newsgroup,
> 
> I have Mathematica 3.0 installed on a Mac IIvx with a math coprocessor 
> and System 7.5 with 20 MB RAM, but I'm interested in timing for other 
> personal computing systems as well.  
> 
> I have tried to benchmark the time required for some common tasks to 
> decide if this software in a personal computer environment can support 
> a system dynamics and controls R&D consultancy.  
> 
> Here's what I found so far:
>    NDSolve on a linear, 2nd order, time invariant equation takes 40 
> seconds to compute 30 seconds of simulated time.  Real system 
> simulations are often nonlinear and of higher order.  I have visions 
> of this taking hours on my Mac, if it can be done at all. 
>    Neither Eliminate nor Solve can handle the mild, low-order 
> nonlinear equations common in modeling gas turbines...no errors, no 
> messages, they simply don't work.  I'm also interested in robot 
> manipulator dynamics, but these are more difficult yet so I'm not at 
> all optimisitic.


I think timing[10001!][[1]] will take about 100 seconds on a IIcx

Todays machines are below 1 sec.
Power Tower Pro 225:         0.783 sec
Pentium pro 200(linux):       1.04 sec

Karl U.