Re: measuring time spent in kernel

[Andrew Morton]

Torrey Hoffman wrote:
> 
> "top" and similar tools don't seem to capture the time spent in kernel
> which isn't on behalf of a user process.
> 
> I vaguely remember mention on this list of a tool that soaks up as many
> cycles as it can get to obtain an accurate measurement of the true
> system time.
> 
> Can some one give me a pointer?  I've had no luck with Google...

http://www.google.com/keyword/keyword/%3Fcyclesoak  :)

-