Messages in this thread | | | Date | Sat, 15 May 1999 22:37:31 -0400 | From | "Alex K." <> | Subject | Memcopy by 8? |
| |
OK, here's my ideas for an optimization: in string-486.h in 2.2.8, the memcpy function is either implemented into 3 separate functions: memcopy by 4, 2, and 1 bytes. For this purpose, the number of bytes to be copied is %ed by 4 and 2. I have two optimizations in mind: 1: Implement a memcpy by 8 for MMX machines that would use the MOVQ command and an MMX register, and an analogous command on non-MMX machines (for consistency) that would use a doubled long move. 2: Since the MOD is takes anyways, why not implement memcpy(to,from,n) as this: block copy (memcpy by 4 or 8) until n-n%{8,4 (depending on the implementation)} and then memcpy by bytes for the rest of the array. Wouldn't this be more efficient, esp. with optimization 1? Can someone give me their opinion?
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |