[Openmcl-devel] Quick HW question...

[Tim Bradshaw]

On 16 Nov 2010, at 15:16, Jon Anthony wrote:

> This is some good information.  Thanks for the pointers.  But it also
> highlights an issue I've thought about from time to time: with modern
> processor architectures (especially pipelines, caches, and now cores)
> how does one _not_ write naive code for these things?  Sure, 90+% of the
> worry on this goes to the compiler writers, but it can be easy to
> accidentally write something that defeats their efforts.

The best answer I know is to profile mercilessly where performance matters.