Re: Moore's law (was Re: [patch 0/9] mutex subsystem, -V4)

From: Andi Kleen
Date: Thu Jan 05 2006 - 11:22:01 EST


On Monday 26 December 2005 01:33, Pavel Machek wrote:

[cc list from hell trimmed down]

> > Another example: Ingo's VFS stresstest which is hitting i_sem hard: it only
> > does ~8000 ops/sec on an 8-way, and it's an artificial microbenchmark which
> > is _designed_ to hit that lock hard. So if/when i_sem is converted to a
> > mutex, I figure that the benefits to ARM in that workload will be about a
> > 0.01% performance increase. ie: about two hours' worth of Moore's law in a
> > dopey microbenchmark.

Moore's law actually doesn't say anything about performance increases,
just about the number of transistors available.
>
> :-) Expressing performance increases in Moore's hours seems like
> neat trick. OTOH I do not think it is valid any more. Single-threaded
> performance stopped increasing 2 years ago AFAICS.

It's not true. e.g. a 2.6 Ghz FX-57 is significantly faster than the
top end CPU you could get 2 years ago. And I'm sure this years CPUs
will be still faster than last years.

> Plus people are
> pushing Linux onto smaller machines, that were unavailable 2 years
> ago.

Even smaller systems are still getting faster.

-Andi
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/