Thanks for the links, I will check them out.

<t>Hi,I need to calculate the eigenvalue of a matrice using c++. I have visual studio 2008 express. Does Boost have this function, or are there some libraries with this function available that anyone has used. Has anyone done this in c++ before.The matrices are large ~40 rows x 40 columns.Thanks! </t>

<t>1. I tried const double, the speed was basically the sameI used 3 nested loops for no particular reason, but both "atan" and the approximation function use the same nested loops. So the comparison between atan and the approximate are valid. Changing j++ and ++j just has an effect on the code call...

<t>I tried the method in the second link. I couldn't connect to the first one. The speed of the algorithm is slower then "atan()" in math.h.About 21 seconds vs 18 seconds for 100 Million Passesdouble ArcTanApprox(double);void TestArcTan(void){ int i,j,k,x3; int k_max,ss; time_t x1,x2; double y[1001]...

The code slows down to such a degree because the function is called 360 x 1000 x 5000 times.There are no other higher order math logic used, so most of the calculations are very simple. If I add an exp() function it slows from 1000 seconds to 1300 seconds.

great I will look at the paper and post the code for my example.

No, the range of values is just 0 to ~2

Hi,I started using the arctangent function in c++. Visual Studio 2008 express. atan(), in math.hIt is really slow. My loop used to take 100 minutes, and now it takes 140 minutes. Is there any approximation or other function that is faster?Thanks

<t>Hi,I am trying to predict my option value change. I have the option value by time period, which is the actual real option value. Then I have the option value that the model predicts.To test how statistically close the two results are what test should I use - R^2?, there are about 200 values of ac...

Right I was second guessing the formula because the results looked funny, but that was because of some other error with the data.Thanks

Hi,If my option value = 11,000delta = -6.14 = dV/dSgamma = 3.87What is the option value if the index moves 2%, I mean index moves from 1100 to 1100 * 1.02so dS = 0.02 * 1100does change in option value = Theta * dt + 0.5 * gamma * dS^2does dS^2 = (0.02 * 1100) ^2?

<t>We are doing monte carlo. So I guess Matlab isn't a good choice.But as far as C++, is there a need for MC? you can download the express version for free - visual studio express 2008 is what I am using.Matlab is expensive, I think a license costs ~10K per year, but C++, Vb.net, C# can be had for n...

I think the point is that a plumber would design better plumbing tools then an engineer.Hands down

<t>Thats a good point everyone has an agenda.In my case I am the end user, and most of the time the end user's agenda is most in line with what is best for the company. I am using the software, so I know how to best create it so it is most user friendly, if something goes wrong, I am the one in trou...

I have seen too many cases where a programmer or IT person has their own agenda and the project follows this, like use this code or that code.The coder / it person has an agenda is not in line with the project, it is this misalignment that leads to issues.