MathGroup Archive 2008

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

Search the Archive

Re: Machine-precision Exp[] function

  • To: mathgroup at smc.vnet.net
  • Subject: [mg93598] Re: Machine-precision Exp[] function
  • From: Albert Retey <awnl at gmx-topmail.de>
  • Date: Mon, 17 Nov 2008 06:17:41 -0500 (EST)
  • References: <gfma98$g54$1@smc.vnet.net> <gfp247$4bq$1@smc.vnet.net>

Roman schrieb:
> answering my own question:
> 
> MachinePrecisionExp = Compile[{{x, _Complex}}, Exp[x]]
> 
> Any thoughts on this?


It is what I would have suggested to look at. You need to decide whether
that is a good solution for your case anyway. Some things to consider:

- With _Complex you will get spurios imaginary parts in the results
which may or may not cause problems or slowdowns later.

- Maybe there are larger parts of formulas that you want to compile,
probably even whole parts of your code. You should consider that if you
are after very fast code. Since unfortunatly Compiles aren't efficient
when beeing nested (they cause callbacks to the evaluator), you should
try to Compile as large junks of code as managable.

hth,

albert


  • Prev by Date: Re: Machine-precision Expfunction
  • Next by Date: RE: Gluing file together
  • Previous by thread: Re: Machine-precision Exp[] function
  • Next by thread: Re: Machine-precision Exp[] function