C # and NMath for Computational Finance and Econometrics

Question

Currently, I (mainly) use C ++ for computationally intensive econometrics in my research. I wanted to move to a more productive environment. D sounds promising, but maybe not mature enough for me to be fully productive right now. I recently met C # and the NMath library. It seems very impressive. Does anyone use them for such an end? What are your impressions? I am ready to give up some speed at runtime if this improves my overall performance (ownership clock is ticking). You appreciated the thoughts and suggestions!

Answer 1

I assume that you are referring to the Centerspace NMath product?

We recently started using them in a production application; although it’s very easy so far (performing linear regression to determine the relationship between GJ of gas in a gas turbine versus MW, technically this is a feature of NMath Stats). So far, it has been a pleasure to take advantage, and their support has been very quick in response to the expired evaluation license, waiting for our financial department to process their account.

In terms of performance, it uses the Intel Math Kernel Library where possible, so performance should be good; this did not negatively impact our limited use.

I look forward to using the library elsewhere in our application, as appropriate!

Answer 2

(EMSI), D . , , C, , ++ (IMHO). BLAS, Intel MKL .., - . D, , - - .

Answer 3

​​ ILNumerics. , . , , , .

Matlab . ILNumerics ( ) , , Matlab. , n , . .

, : . ( std, cov, var, mean, abs ..), HILO .