On Wed, Oct 23, 2002 at 11:54:42PM +0900, Bang Jun-Young wrote:

 > Also, I found that my MMX-optimized i686_copyin() is faster than plain
 > old memcpy for data > 2~3 KB. It seems that saving/restoring FP state in/
 > from stack is quite expensive for small data copying (it needs 108 bytes
 > of memcpying from processor to memory plus some overhead).

It needs a lot more than that for the SSE/SSE2 case.

-- 
        -- Jason R. Thorpe <thorpej@wasabisystems.com>