Re: [PATCH 0/4] Fix ebizzy performance regression due to X86 TLBrange flush v2

From: H. Peter Anvin
Date: Fri Dec 13 2013 - 17:39:20 EST


On 12/13/2013 01:16 PM, Linus Torvalds wrote:
> On Fri, Dec 13, 2013 at 12:01 PM, Mel Gorman <mgorman@xxxxxxx> wrote:
>>
>> ebizzy
>> 3.13.0-rc3 3.4.69 3.13.0-rc3 3.13.0-rc3
>> thread vanilla vanilla altershift-v2r1 nowalk-v2r7
>> Mean 1 7377.91 ( 0.00%) 6812.38 ( -7.67%) 7784.45 ( 5.51%) 7804.08 ( 5.78%)
>> Mean 2 8262.07 ( 0.00%) 8276.75 ( 0.18%) 9437.49 ( 14.23%) 9450.88 ( 14.39%)
>> Mean 3 7895.00 ( 0.00%) 8002.84 ( 1.37%) 8875.38 ( 12.42%) 8914.60 ( 12.91%)
>> Mean 4 7658.74 ( 0.00%) 7824.83 ( 2.17%) 8509.10 ( 11.10%) 8399.43 ( 9.67%)
>> Mean 5 7275.37 ( 0.00%) 7678.74 ( 5.54%) 8208.94 ( 12.83%) 8197.86 ( 12.68%)
>> Mean 6 6875.50 ( 0.00%) 7597.18 ( 10.50%) 7755.66 ( 12.80%) 7807.51 ( 13.56%)
>> Mean 7 6722.48 ( 0.00%) 7584.75 ( 12.83%) 7456.93 ( 10.93%) 7480.74 ( 11.28%)
>> Mean 8 6559.55 ( 0.00%) 7591.51 ( 15.73%) 6879.01 ( 4.87%) 6881.86 ( 4.91%)
>
> Hmm. Do you have any idea why 3.4.69 still seems to do better at
> higher thread counts?
>
> No complaints about this patch-series, just wondering..
>

It would be really great to get some performance numbers on something
other than ebizzy, though...

-hpa


--
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/