Messages in this thread | | | Date | Wed, 29 Mar 2000 19:25:07 +0200 (CEST) | From | (Arjan van de Ven) | Subject | Re: [patch] preemptive kernel, preemptive-2.3.52-A7 |
| |
In article <20000329093127.C27237@hq.fsmlabs.com> you wrote: > If read/write is too slow, figuring out some way to force it into > mmap or do something else smart is an alternative to checking needs > resched.
Or "just" do the same thing the mm system does: work in chunks with the size as specified in /proc/sys/vm/page-cluster. This is a relatively small size (typically 32Kb to 64Kb) and is supposed to be set to a value the harddisk is comfortable with (read: no more than half the cache-size on the disk). I remember a discussion around the 2.1.125 time that stated that clustering in blocks of 32Kb almost doubled the throughtput of hdd's in that time, but that larger blocks decreased performance.. write() could schedule (or attempt to if needed) after such a block. I hope this will be good for both hdd throughput and latency...
Greetings, Arjan van de Ven
- 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/
| |