Re: [2.4.17/18pre] VM and swap - it's really unusable

Peter Wächtler (pwaechtler@loewe-komp.de)
Mon, 21 Jan 2002 18:12:58 +0100


yodaiken@fsmlabs.com schrieb:
>
> On Mon, Jan 21, 2002 at 05:33:50PM +0100, Peter Wächtler wrote:
> > yodaiken@fsmlabs.com schrieb:
> > >
> > > On Mon, Jan 21, 2002 at 05:05:01PM +0100, Daniel Phillips wrote:
> > > > > I think of "benefit", perhaps naiively, in terms of something that can
> > > > > be measured or demonstrated rather than just announced.
> > > >
> > > > But you see why asap scheduling improves latency/throughput *in theory*,
> > >
> > > Nope. And I don't even see a relationship between preemption and asap I/O
> > > schedulding. What make you think that I/O threads won't be preempted by
> > > other threads?
> > >
> >
> > I/O intensive threads block early voluntarily - while CPU hogs don't.
>
> Since the preemption patch only allows additional preemption in kernel
> mode, I'm curious to know what the compute bound tasks are doing in
> kernel mode. Did Linux add in-kernel matrix multiplication while
> I was not looking?
>

Dead right you are.
Then there are only slow system calls left. Umh, execve(), fork()
(with big address space) - what about page_launder etc.?

> > Statistically there is a higher chance, that a CPU hog gets preempted
> > instead of an IO bound (that gives up the CPU in some useconds anyway)
>
> "Statistically"? As far as I know, most I/O in Linux does not block.

You mean, the syscall returns without a reschedule?
Aehm, now it's time for some statistics where the kernel spents its time on ;-)

But what is a possible explanation for the people, who think their
systems behave better with preemption - strong believe?
-
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/