- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
An application that is dynamically linked to the Intel IPP v4.0 libraries has a faster execution time than when statically linked at AMD CPU.
Link Copied
2 Replies
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Yes I believe so. I believe that IPP is using the MMX chip on the processor to accelerate it's computations. I'm not quite sure how it maps to AMD's 3DNow! but it seems like it doesn't very well, if at all. I have been running on my AMD athlon at home and on my INtel here at work and by far, the intel is much much much faster despite having only a slight frequency difference. I hope this helps.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi,
First of all, Intel IPP uses PX code on non-Intel processors, it is the easiest way to avoid any issues with compatibility in instruction set. Second, IPP uses OMP threading in some functions in DLLs and does not use it in static libraries, so on dual-way systems IPP DLLs can be faster than static libraries
You also can be interested to lookonother topics about using IPP on non-Intel processors:
Regards,
Vladimir
Message Edited by vdudnik on 10-07-2005 10:44 AM

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