Re: Precision and rounding
- To: mathgroup at smc.vnet.net
- Subject: [mg33231] Re: Precision and rounding
- From: "Allan Hayes" <hay at haystack.demon.co.uk>
- Date: Sat, 9 Mar 2002 18:05:15 -0500 (EST)
- References: <a675kf$d14$1@smc.vnet.net> <a6ci84$ef6$1@smc.vnet.net>
- Sender: owner-wri-mathgroup at wolfram.com
Numerics is not my area but the following may be of some help:
NIntegrate[Sin[x+y],{x,0,7},{y,0,7}]//Timing
{46.46 Second,0.323366}
NIntegrate[Sin[x+y],{x,0,7},{y,0,7},
PrecisionGoal->3]//Timing
{0.61 Second,0.323366}
In some cases we can restrict the number of significant digits used in
internal calculations.
Default setting of $MaxPrecision is 10^6
Block[{$MaxPrecision=3},
NIntegrate[Sin[x+y],{x,0,7},{y,0,7}]
]//Timing
NIntegrate::ploss: Numerical integration stopping due to loss of
precision. \
Achieved neither the requested PrecisionGoal nor AccuracyGoal; suspect
one of \
the following: highly oscillatory integrand or the true value of the
integral \
is 0. If your integrand is oscillatory try using the option \
Method->Oscillatory in NIntegrate.
{0.11 Second,0.323368}
Block[{$MaxPrecision=6},
NIntegrate[Sin[x+y],{x,0,7},{y,0,7},
PrecisionGoal->3]
]//Timing
{0.99 Second,0.323366}
--
Allan
---------------------
Allan Hayes
Mathematica Training and Consulting
Leicester UK
www.haystack.demon.co.uk
hay at haystack.demon.co.uk
Voice: +44 (0)116 271 4198
Fax: +44 (0)870 164 0565
<bradamanteus at yahoo.com> wrote in message news:a6ci84$ef6$1 at smc.vnet.net...
> There must be a simple answer to this:
>
> I'm producing a large output file that takes a long time to compute, and I
> think this is partly becuase Mathematica always has my output as 14
decimal
> places, and I don't need such accuracy. Is there a way to reduce
calculation
> time by reducing the precision of all calculations in my model. Rounding
> with N doesn't work.
>
> Also, I'm told that I can reduce the time of my program by looking at a
> debugging tool that tells me how long the program spends on certain parts
of
> my code. What is this?
>
> thanks
>
> Leanne Ussher
>
>
>