(Dag-Erling Sm=F8rgrav) writes:
> Gary Corcoran writes:
> > The confusing thing is that I thought 'real' time should be >=3D 'user'=

+ 'sys'.
> > But here 'user' is much greater than 'real' for both machines! The sen=

se I
> > got from the other messages in this thread is that 'user' time is somew=

> > meaningless (i.e. unreliable as a measure) in a multi-CPU and/or hypert=

> > environment. Can you clarify?

> real >=3D (user + sys) * ncpu

umm, other way around of course.

real * ncpu >=3D (user + sys)

Dag-Erling Sm=F8rgrav -
_______________________________________________ mailing list
To unsubscribe, send any mail to ""