On 5 September 2012 22:52, Arjan van de Ven <arjan(a)linux.intel.com> wrote:
On 9/5/2012 10:19 AM, Rajagopal Venkat wrote:
> On 5 September 2012 22:39, Arjan van de Ven <arjan(a)linux.intel.com> wrote:
>> On 9/5/2012 9:56 AM, Rajagopal Venkat wrote:
>>>> measure1:
>>>> ev3.start
>>>> ev1.end <<<<<
>>>
>>> evX.end <<<<<
>>> These events are causing numbers to go wrong.
>>
>> but out of a 20 second window.. this is a tiny tiny window...
>> if you see 100.1% I'd buy this reasoning.
>> but you're seeing much more than that.
>
> How about generating a report for 1sec duration?
even for 1 second... still it's miniscule compared to this whole 1 second
the amount of setup/teardown time just is not that huge.
Here are some perf timestamps,
(3979299431)
(3979303554)
(4079217947)
(4091306943)
(4091322535)
(4091336882)
When 1sec report is generated and if above timestamp gets
added to timer accumulated_runtime, no wonder why such
huge usage is reported.
--
Regards,
Rajagopal