> From: "Paul Stimpson" > That's what I was thinking too but the machine is a dual core 2.4 and the
> 2 cores are alternating between 43% and 57% (one on each then swapping).
> I assume that random generation is a compute-bound activity so if that
> was the bottleneck I would have expected near on 100% both sides. ... > Don't know why this thing is so slow.
It's a bit academic but I wonder if the urandom algorithm is being split
into _dependent_ parts (43+57=100%) ? A bug list somewhere might like
to know....
