Re: [Devil-linux-develop] RTOS patch for DL

[Heiko Z. <he...@zu...>]

On Wed, July 27, 2005 08:46, Serge Leschinsky wrote:
> Dear Sirs,
>
>
> Saturday, July 23, 2005, 1:35:54 PM, you wrote:
>
>
> SL> it. Unfortunately vhz patch and performance patch are incompatible
> SL> now.
> I've forgotten to say I applied this performance patch without PAX and
> grsecurity. I've tested my system by SysBench (system performance
> benchmark, http://sysbench.sourceforge.net/) with performance patch and
> without it.
>
>> # ./sysbench --num-threads=500  --max-requests=10000 --test=threads
>> run sysbench v0.4.0:  multi-threaded system evaluation benchmark
>>
>> Running the test with following options:
>> Number of threads: 500
>>
>>
>> Doing thread subsystem performance test
>> Thread yields per test: 1000 Locks used: 8
>> Threads started!
>> Done.
>>
> Without Perf patch    |  With Perf patch
> |
>
>> Test execution summary:                        |
>> total time:                  236.9318s     |  359.9786s total number of
>> events:      10000         |  10000
>> total time taken by                        | event execution:
>> 113411.5786   |  179826.3306
>> per-request statistics:                    | min:
>> 0.0010s       |  13.8871s
>> avg:                    11.3412s      |  17.9826s
>> max:                    93.7898s      |  21.2820s
>> approx.  95 percentile: 61.2178s      |  18.1917s |
>> Threads fairness:                              |
>> distributin:                 45.45/79.44   |  100.00/100.00
>> execution:                   42.06/78.00   |  99.62/99.93
>>
>
> The most interesting result is "approx.  95 percentile:", i.e. the
> most evident time of thread execute.

Are you sure you have the columns right? Right now it looks to me like
without the perf patch it's faster.
(It could also be that I can't think straight, because I had too much for
lunch)

-- 

Regards
  Heiko Zuerker
  http://www.devil-linux.org