Profiling work running on Xeon Phi from host?

Profiling work running on Xeon Phi from host?

I'm using OpenCL with Xeon Phi on Linux. The host code is executed on host operating system and the kernel code is executed on a Xeon Phi card. I wonder if there is any way to profile (cache misses, instructions, etc.) of the kernel code on Xeon Phi? I would expect something like 

//host code:

read_counter();

kernel_code_on_Xeon_Phi();

stop_counter();

Can I do this with VTune™ Amplifier XE?

Thanks and regards,

Tuan

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

Yes, you can do this with VTune Amplifier - just as for host code.

thank you!

Leave a Comment

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