Search All of the Math Forum:

Views expressed in these public forums are not endorsed by NCTM or The Math Forum.

Topic: Ask for help: how to let mathematica output Fortran code with fewest
float operations?

Replies: 9   Last Post: Jan 13, 2013 7:27 PM

 Messages: [ Previous | Next ]
 Tang Laoya Posts: 23 Registered: 11/21/12
Re: Ask for help: how to let mathematica output Fortran code with
fewest float operations?

Posted: Jan 13, 2013 7:27 PM

On Monday, January 14, 2013 4:00:36 AM UTC+8, Axel Vogt wrote:
> On 13.01.2013 19:12, Tang Laoya wrote:
>

> > On Sunday, January 13, 2013 11:28:14 PM UTC+8, Axel Vogt wrote:
>
> >> On 13.01.2013 15:29, Tang Laoya wrote:
>
> ...
>

> >>> solutions := solve({x*y*z = b, x+y+z = a, x^2+y^2+z^2 = 1}, {x, y, z})
>
> ...
>

> >> Are you aware, what Maple finds are 6 generic solutions,
>
> >> parametric in a and b, and that they are complex valued
>
> >> in general?
>
> >
>
> > Hi Axel Vogt,
>
> >
>
> > Thank you very much for your kindly reply. I think that the complex solutions should not effect the output to Fortran code, isn't it?
>
> >
>
> > On the other hand, what I posted is just an example and I want to know how to let it simplify the express or output to Fortran code. The real problem is much more complex than this one. As Prof. Fateman pointed out, it is meanless to solve so complex equations since it is much better to solve the equations by numerical method.
>
>
>
> Yes+No:
>
>
>
> If you have a sound numerical solver covering several variables
>
> and providing complex results then it *may* be better ...
>
>
>
> But then the question should not be at 'symbolic' :-)
>
>
>
> The other thing is: there may be more than 1 solution, like for
>
> your task intersection of a plane and a cubic on the unit sphere

Hi Axel Vogt,

Thank you very much for your kindly reply. For the multiple solutions problem, I have a decide to select which one in the Fortran code.

My final problem has numerical method to do it, but it is iterative method and I am trying to find analygical method for both speed and accurate.

Thanks,
Tang Laoya

Date Subject Author
1/13/13 Tang Laoya
1/13/13 Tang Laoya
1/13/13 Richard Fateman
1/13/13 Tang Laoya
1/13/13 Nasser Abbasi
1/13/13 Tang Laoya
1/13/13 Axel Vogt
1/13/13 Tang Laoya
1/13/13 Axel Vogt
1/13/13 Tang Laoya