>I'll dig it out, and give it a test.  Could someone send me the stats for
>the kernel's use of bcopy?  All 3 parameters ideally, or at least the bottom
>2 bits of the src and dest addresses.

Erm I'll have to calculate those for you. I only have the size stats at the
moment. I need to add a few patches to the existing bcopy to generate the
approriate stats.
What stats are needed to produce the best code ? I can count alignment for each
call, size of each call, aligment per call per size ?
I suspect what will be needed is a general purpose optimised bcopy for variable
length blocks / alignment (using ldm/std etc.) and tests and branches to
specialised routines for the most frequent lengths.

>Oh, and what assembler format?
gas would be best

Cheers,
				Mark

-- 
Mark Brinicombe				amb@physig.ph.kcl.ac.uk
Research Associate			http://www.ph.kcl.ac.uk/~amb/
Department of Physics			tel: 0171 873 2894
King's College London			fax: 0171 873 2716