Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

- Intel Community
- Software Development SDKs and Libraries
- Intel® oneAPI Math Kernel Library & Intel® Math Kernel Library
- TGAMMA approximation method.

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Mute
- Printer Friendly Page

Bernard

Black Belt

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

07-04-2012
11:21 AM

32 Views

TGAMMA approximation method.

Isthis possible to discuss on this forum the approximation method of VML Tgamma function?

I'am interested in what an approximation method is used to calculate TGamma values.

Link Copied

2 Replies

Sergey_M_Intel2

Employee

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

07-04-2012
09:13 PM

32 Views

We do not disclose algorithmic details of math library functions implemented in our products.

Widely known approximations of gamma functions are Stirling's and Lanczos's ones. Good overview of the function along with its approximations can be found here http://mathworld.wolfram.com/GammaFunction.html. You might also want to look at numerical analysis classic book by Abramowitz and Stegun "Handbook of Mathematical Functions".

I hope that helps,

Regards,

Sergey

Bernard

Black Belt

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

07-04-2012
10:37 PM

32 Views

Widely known approximations of gamma functions are Stirling's and Lanczos's ones

Thank you Sergey for your answer.

I have already implemented gamma function with the help of stirling approximation.Code and run-time tests can be found here http://software.intel.com/en-us/forums/showthread.php?t=106032

I had a problem with an accuracy in the range [0.001,1.0] and was curious how it was solved by Intel programmers(my problem was solved with the help of minimax approximation in problematic range).

I have a Stegun and Avramovitz classic book which serves me well as source ofvariousformulas for my mathematical library.

I do understand that such a discussion will not be possible here.

For more complete information about compiler optimizations, see our Optimization Notice.