Re: Help Resource Counters Scale better (v4)

[Andi Kleen]

Prarit Bhargava <prarit@xxxxxxxxxx> writes:
>
> On a 64p/32G system running 2.6.31-git2-rc5, with RESOURCE_COUNTERS

This is CONFIG_RESOURCE_COUNTERS off at compile time right?

> off, "time make -j64" results in
>
> real    4m54.972s
> user    90m13.456s
> sys     50m19.711s
>
> On the same system, running 2.6.31-git2-rc5, with RESOURCE_COUNTERS on,
> plus Balbir's "Help Resource Counters Scale Better (v3)" patch, and
> this patch, results in
>
> real    4m18.607s
> user    84m58.943s
> sys     50m52.682s

Hmm, so resource counters on with the patch is faster than
CONFIG_RESOURCE_COUNTERS compiled out in real time? That seems
counterintuitive. At best I would expect the patch to break even, but
not be actually faster.

Is the compilation stable over multiple runs?

Still it looks like the patch is clearly needed.

-Andi

-- 
ak@xxxxxxxxxxxxxxx -- Speaking for myself only.
--
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/