>
>So is there a plan to use a Pentium RDTSC based version of microtime?  

There was once a plan to support nanotime(), using the NTP
"nanokernel", but other developers were obstructive.  The nanokernel
is quite general, and can handle SMP systems where, like some Alphas,
the CPU cycle clocks are not synchronised.

(There are hard problems doing interpolation with variable-rate
clocks, especially if the rate-change mechanisms (like APM bioses)
dont tell the time-interpolation code when they change the clock
rate).