Subject: Re: Port benchmarks
To: Manuel Bouyer <bouyer@antioche.lip6.fr>
From: Andrew Gallatin <gallatin@cs.duke.edu>
List: port-alpha
Date: 04/04/2000 16:40:16
Manuel Bouyer writes:
> >
> > I just had a peek at our bcopy - there's appears to be no loop unrolling
> > at all (look at the bcopy_samealign_lp). What's the typically loop
> > overhead on an alpha CPU?
>
> I don't know, but I would expect that a tuned bcopy would improve things ...
> unfortunably I don't know alpha assembly.
Somebody might want to take a look at Appendix B of the Compiler
Writer's Guide for the 21264
(http://ftp.digital.com/pub/Digital/info/semiconductor/literature/cmpwrgd.pdf)
They present how to optimize a checksumming routine. I imagine most
of this could be transferrable to optimizing bcopy.
Cheers,
Drew
------------------------------------------------------------------------------
Andrew Gallatin, Sr Systems Programmer http://www.cs.duke.edu/~gallatin
Duke University Email: gallatin@cs.duke.edu
Department of Computer Science Phone: (919) 660-6590