Re: ZVC: Increase threshold for larger processor configurationss

[Christoph Lameter]

On Thu, 29 Jun 2006, Andrew Morton wrote:

> A lot of the counters only ever count in one direction!  So we could even
> skew them by an entire STAT_THRESHOLD.

I just tested both approaches separately. If I leave the STAT_THRESHOLD at 
32 and just add the STAT_THRESHOLD / 2 trick we can scale cleanly up to 
160p in the synthetic test.

I have a bad feeling though for 512p 1024p and 4096p with the static 
approach with STAT_THRESHOLD /2 but it will take some time to get tests 
done on those.

I think the best right now is to just take the dynamic threshold patch. I 
expect that to do just fine at the higher numbers. If we have more trouble 
at 1024p then maybe try STAT_THRESHOLD/2 before going to 
aggregate counter consolidation.

-
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/