numbers don't lie ...

Kris Kennaway kris at obsecurity.org
Thu Sep 14 10:35:58 PDT 2006


On Thu, Sep 14, 2006 at 01:28:03PM -0400, Gary Corcoran wrote:

> The confusing thing is that I thought 'real' time should be >= 'user' + 
> 'sys'.

No.  This is at best only ever approximately true on a uniprocessor
machine when there is no blocking I/O being performed.

A bit of thought about the nature of a multiprocessor system will lead
you to the conclusion that when two processes are executing in
parallel, the real time will be less than the sum of both user and
system times - that's what it means to have parallelism.

> But here 'user' is much greater than 'real' for both machines!  The sense I
> got from the other messages in this thread is that 'user' time is somewhat
> meaningless (i.e. unreliable as a measure) in a multi-CPU and/or 
> hyperthreading
> environment.  Can you clarify?

See my earlier messages; it's only meaningless if you don't understand
its meaning.

Kris
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 187 bytes
Desc: not available
Url : http://lists.freebsd.org/pipermail/freebsd-hackers/attachments/20060914/c101a22e/attachment.pgp


More information about the freebsd-hackers mailing list