>Mach 3.0 employs continuations in the kernel itself, and allows cross adress  
>space RPC's to execute 14% fsater, and exception handling to over 60% faster.
>(/doc/published/threadmgnt.ps)

So, as far as I can tell, Mach 3.0 was sufficiently _slow_ that just
about nobody here (at CMU) wanted to run it on their desktops.

Apparently, that's gotten better with the work done at Utah, but i'd
trust what I've heard people say about whole system performance rather
than the numbers that the Mach group published...



cgd