Auto-vectorization on the Intel Xeon Phi Coprocessor.

Auto-vectorization on the Intel Xeon Phi Coprocessor.


Can the Intel SDK for OpenCL applications vectorize only the dimension 0 of the workgroup or can it even vectorize loops present inside the work-items (loops in the kernel itself). I am not sure if I have made the question clear enough. Let me know if you have any questions. 



3 posts / 0 new
Last post
For more complete information about compiler optimizations, see our Optimization Notice.

Hi Sumedh,

Currently, the Intel OpenCL compiler only vectorizes along the dimension zero work-items loop. If you have a real life example in which you think that vectorizing other loop would yield better performance, then please share it here (or privately). We would like to analyze such cases.



Thanks! :) 

Leave a Comment

Please sign in to add a comment. Not a member? Join today