Re: IO degradation in 2.4.17-pre2 vs. 2.4.16

Bill Davidsen (davidsen@tmr.com)
Tue, 11 Dec 2001 17:37:54 -0500 (EST)


On Tue, 4 Dec 2001, Marcelo Tosatti wrote:

> Yes, throughtput-only tests will have their numbers degradated with the
> change applied on 2.4.16-pre2.
>
> The whole thing is just about tradeoffs: Interactivity vs throughtput.
>
> I'm not going to destroy interactivity for end users to get beatiful
> dbench numbers.
>
> And about your clients: Don't you think they want some kind of
> decent latency on their side?

It depends on the machine. For a server the thing you need to feed clients
is throughput. I don't see how feeding the data slower is going to be GOOD
for latency. Particularly servers which push a lot of data, like mail and
news or certain web sites, need to push it now.

Latency is more of an issue for end user machines.

-- 
bill davidsen <davidsen@tmr.com>
  CTO, TMR Associates, Inc
Doing interesting things with little computers since 1979.

- 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/