There is a macro that defines CLOCKS_PER_SEC, which can be used to turn the result of clock() into seconds. Note that this may not reflect the actual resolution, e.g. CLOCKS_PER_SEC may be 1000, but it may count up 14 clock-ticks each time, because it is running on a 700-odd Hz timer.
If you want reasonable portability, then the right thing to do is to write a wrapper function that hides the actual implementation, e.g. a "starttime()" and "stoptime()" that return a floating point value (double) in seconds. By returning a floating point value, there is no reason to worry about how the internal representation of time is done.
You can then choose the best function for each architecture to solve the problem in the best way on that OS.
--
Mats