- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
the development manual says : The vector pragma indicates that the loop should be vectorized, if it is legal to do so . What does legal imply here? loop dependencies ?
Thanks
- Tags:
- CC++
- Development Tools
- Intel® C++ Compiler
- Intel® Parallel Studio XE
- Intel® System Studio
- Optimization
- Parallel Computing
- Vectorization
Link Copied
1 Reply
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
#pragma vector over-rides cost analysis, where the compiler estimates whether vectorization will be be beneficial for the target instruction set. So it will try to vectorize even if it expects to lose performance. It usually includes an assertion such as aligned or unaligned (these direct specific handling of data alignment) or nontemporal (to over-write entire blocks without reading into cache).
Unlike #pragma omp simd it doesn't assert vector independence. So it is often used together with __restrict definition of assignment through pointer.
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