"Paul J. Reder" wrote:
> 

> >
> > btw, can we get rid of the the second index ( [0] ) ?  Earlier in 2.0,
> > we were allocating 10 scoreboard images; I assume this is some residue.
> >

> 
> It is my understanding that this is to allow the generic scoreboard to represent
> process oriented mpms ([processes][no threads]) as well as thread oriented mpms
> ([processes][threads]). So for mpms with no threads you just use a 0 in the
> second index.
> 

Thanks...makes sense.  

hmmm...we could make it go faster still by pre-computing a
one-dimensional worker number to use for the scoreboard index at worker
initialization time.  

Greg

Reply via email to