MathGroup Archive 2007

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

Search the Archive

Re: To dual-core or not to dual-core

  • To: mathgroup at smc.vnet.net
  • Subject: [mg73473] Re: To dual-core or not to dual-core
  • From: David Bailey <dave at Remove_Thisdbailey.co.uk>
  • Date: Mon, 19 Feb 2007 01:25:27 -0500 (EST)
  • References: <er9cs3$dad$1@smc.vnet.net>

terry wrote:
> Dear Mathematica Experts,
> 
> I currently run simulations in Mathematica 5.2 that take on the order
> of months to complete. These
> simulations all consist of operations like the following:
> 
>          1) {a,b,c,...} - {x,y,z,...}
>          2) a - {x,y,z,...}
>          3) a*{x,y,z,...}
>          4) Apply[Plus, Abs[{x,y,z,...}]]
> 
> I know that some operations in 5.2 will automatically take advantage
> of mulitple cores if they
> exist on a machine, splitting up the work automatically across cores
> without any custom coding
> on the users part, but does this apply to operations like the ones
> I've listed above?
> 
> Any and all insights are graciously appreciated!
> 
> Terry
> 
> 
I would look very carefully at your code before changing your hardware.

One way to make arithmetic very, very slow is to inadvertently force 
Mathematica to do exact fractional calculations when you think you are 
performing real number calculations. If that is not an issue, you might 
want to look at the use of Compile.

David Bailey
http://www.dbaileyconsultancy.co.uk


  • Prev by Date: Re: explanation need!
  • Next by Date: Re: Coaxing N[] to work
  • Previous by thread: To dual-core or not to dual-core
  • Next by thread: Re: To dual-core or not to dual-core