RE: [PATCH] sched: Folding nohz load accounting more accurate

[Doug Smythies]

>> On Tue, 2012-06-19 at 08:50 -0700, Doug Smythies wrote:
>  On 2012.06.20 02:45 -0700, Peter Zijlstra wrote:

>> My automated 63 hour test has been terminated, the code changed
>> and the test re-started. 

> There's another boo-boo in there:

O.K. I let the automated test finish the 3 processes stuff
(which is 1/2 way).
The newly compiled code is now running, but I only resumed
the test, I did not re-start it.

A comparative graph of the 1 & 2 process results is attached.
(the previous version of this same graph was on another thread).
Legend:
Control: older kernel compiled with CONFIG_NO_HZ=no
c308: Kernel with the earlier commit
5aaa: Basically, Kernel 3.5 RC2
Wang: Charles Wang's first proposed solution.
Peter35: Peter Zijlstra's current proposed solution.

Summary: Extremely good.

Attachment: peter_compare.png
Description: PNG image