Dag Sverre Seljebotn skrev:
>It seemed to me then that the auto-vectorizer only tackles parallel arithmetic >>and not reductions... >(the strides/power of 2 shouldn't matter if it is declared "c" ?) I know... with mode "c" or "fortran", the last or first stride is known at compile time. Still the message from the vectorizer seem to indicate that this is the problem. I tried this with hand-written C as well. The inner loop is not vectorized, so you are probably right about reductions I did not get "data ref analysis failed" message from the auto-vectorizer. I only get this message with Cython-generated C. Sturla _______________________________________________ Cython-dev mailing list [email protected] http://codespeak.net/mailman/listinfo/cython-dev
