Re: stochastic fair queueing in the elevator [Re: [BENCHMARK]

Rik van Riel (riel@conectiva.com.br)
Mon, 10 Feb 2003 19:43:32 -0200 (BRST)


On Mon, 10 Feb 2003, Kurt Garloff wrote:

> Readahead kills seeks and command overhead at the expense of maybe
> transfering data needlessly over the bus and consuming RAM.
>
> AS kills seeks. (At the expense of delaying some IO a tiny bit.)
>
> If unecessary seeks are the main problem, with AS smaller READA is
> possible. If command overhead is a problem, READA needs to be large.

Think about reading inode blocks or a bunch of related .h
files. We can't nicely do readahead in either of these
scenarios, but AS should take care of them nicely...

Rik

-- 
Bravely reimplemented by the knights who say "NIH".
http://www.surriel.com/		http://guru.conectiva.com/
Current spamtrap:  <a href=mailto:"october@surriel.com">october@surriel.com</a>
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/