Messages in this thread | | | Date | Thu, 17 Dec 1998 19:37:05 +0100 (CET) | From | Rik van Riel <> | Subject | Re: mmap() is slower than read() on SCSI/IDE on 2.0 and 2.1 |
| |
On Thu, 17 Dec 1998, Stephen C. Tweedie wrote: > On Thu, 17 Dec 1998 01:27:05 +0100 (CET), Rik van Riel > <H.H.vanRiel@phys.uu.nl> said: > > > Once again, you seem to be forgetting about read-behind. > > No, I'm ignoring read-behind. If people access data backwards > then tough. Real disks don't work that way.
This is exactly why we want to do proper read-behind...
There are quite a lot of programs that do tiled backwards reading of data, just think about large matrix stuff or some image viewing or manipulation stuff.
It is relatively cheap to do read-behind, so I don't see any reason to let the backwards-reading programs slow down the rest of the system by causing 'improper' disk access patterns.
We can simply implement the read-behind in the same way we do read-ahead, but we start at address-16 instead of address itself.
cheers,
Rik -- the flu hits, the flu hits, the flu hits -- MORE +-------------------------------------------------------------------+ | Linux memory management tour guide. H.H.vanRiel@phys.uu.nl | | Scouting Vries cubscout leader. http://www.phys.uu.nl/~riel/ | +-------------------------------------------------------------------+
- 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/
| |