On Tue, Feb 25, 2003 at 12:28:05AM -0800, Rick Lindsley wrote:
> This version makes the disk stats on struct gendisk per-cpu.
> I am working on making the per partition stats per-cpu too (struct
> hd_struct).
>
> In general I'm in favor of this. It seems intuitive to me that counters
> of this type should be per-cpu. But the question is, do we actually
> see any gains? At the very least, are we sure we've not introduced any
> degradation? Has any of your testing so far been measuring performance or
> just checking for correctness?
Tests on this patch have been for correctness. However, we had done some
measurements sometime back (actually loong time back) to verify if we actually
saw any gains with similar statistics counters. The results are at
http://lse.sourceforge.net/counters/statctr.html
Since, we already know that per-cpu counters are good from our experiments
earlier, I did not bother doing it for this patch.
-
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/
This archive was generated by hypermail 2b29 : Fri Feb 28 2003 - 22:00:27 EST