Is there a high precision library in Intel Math Kernel Library?

[losemind]

Hi all,

In my project, I need high precision and numbers with precision to be as high as possible.

At the same time, speed is also my major concern.

I am using MPFR multiple precision library, which does not have platform dependent optimization with respect to Intel platform

I am wondering if Intel Math Kernel library has high precision library? Or MPFR has platform dependent optimization?

Thanks a lot

[TimP]

According to the web page, branches of gmp other than mpfr offer best performance by some measures. Certainly, there's been plenty of platform specific optimization.

[losemind]

-->>branches of gmp other than mpfr offer best performance by some measures.

** What other branches of GMP offer better performance for high precision floating point computations? I thought the MPFR is the only branch of GMP that offers high precision floating point computations...

-->>Certainly, there's been plenty of platform specific optimization.

** Where? That's exactly I am looking for. So you say that MKL does not have the high precision libraries but there are others? Please give me a few pointers... Thanks