Re: [PATCH 0/9] Scalability requirements for sysv ipc - v3

[Nadia Derbey <Nadia.Derbey@bull.net>]

KOSAKI Motohiro wrote:
> Hi
> 
> nice improvement.
> but...
> 
> 
>>   2.6.25-rc3-mm1   2.6.25-rc3-mm1+   2.6.25-mm1   Patched 2.6.25-mm1
>> 1         1168441           1064021       876000               947488
>> 2         1094264            921059      1549592              1730685
>> 3         2082520           1738165      1694370              2324880
>> 4         2079929           1695521       404553              2400408
>> 5         2898758            406566       391283              3246580
>> 6         2921417            261275       263249              3752148
>> 7         3308761            126056       191742              4243142
>> 8         3329456            100129       141722              4275780
>>
>> 1st column: stock 2.6.25-rc3-mm1
>> 2nd column: 2.6.25-rc3-mm1 + ipc patches (store ipcs into idrs)
>> 3nd column: stock 2.6.25-mm1
>> 4th column: 2.6.25-mm1 + this pacth series.
> 
> 
> this result is slightly odd.
> 
> similar to 2.6.25-rc3-mm1 and patched-2.6.25-mm1
> similar to 2.6.25-mm1 and 2.6.25-rc3-mm1
> 
> but
> 
> not similar to 2.6.25-rc3-mm1 and patched-2.6.25-rc3-mm1
> not similar to 2.6.25-mm1 and patched-2.6.25-mm1
> 
> Is patched-2.6.25-rc3-mm1 and patched-2.6.25-mm1 applied the same patch?

No

> or I misunderstand how to see your chart?
> 
> 

Well, looks like the description was not clear, sorry for that!
1st column: 2.6.25-rc3-mm1 with original ipc implementation
                            (i.e. ipcs stored in an array)
2nd column: 2.6.25-rc3-mm1, with ipcs stored in an idr tree
             actually, that's when the performance regression
             had been introduced.
3rd column: 2.6.25-mm1, still with the same implementation
             i.e. still with degraded performances.
4th column: 2.6.25-mm1 + rcu-based implementation of the idr,
             which is the current patch series.

Hope this makes things more clear.

Regards,
Nadia