linux patches related to time measurement accuracy?

Question

Is there a set of patches around for linux kernel that address accuracy of (system, rather than user) process time?

i've heard about RTLinux, but i don't really need the actual real-time scheduling constraints, i just need that the time measurements are accurate. If the process switch slipped by a couple milliseconds, so be it, but i need an accurate measurement of the actual time spent in the current process

Answer 1

If what you have in mind is the usage of time shell builtin, then 3 digit precision is the design limitation - as per man bash:

TIMEFORMAT
          The value of this parameter is used as a format string  specify‐
          ing  how  the timing information for pipelines prefixed with the
          time reserved word should be displayed.  The % character  intro‐
          duces  an  escape  sequence  that is expanded to a time value or
          other information.  The escape sequences and their meanings  are
          as follows; the braces denote optional portions.
          %%        A literal %.
          %[p][l]R  The elapsed time in seconds.
          %[p][l]U  The number of CPU seconds spent in user mode.
          %[p][l]S  The number of CPU seconds spent in system mode.
          %P        The CPU percentage, computed as (%U + %S) / %R.

          The  optional  p is a digit specifying the precision, the number
          of fractional digits after a decimal point.  A value of 0 causes
          no decimal point or fraction to be output.  At most three places
          after the decimal point may be specified; values  of  p  greater
          than  3 are changed to 3.  If p is not specified, the value 3 is
          used.

For seemingly more precise measurements, you can try using date with customized output formatting to show nanoseconds:

Tstart=$(date "+%s.%N")
some_command(s)
Tend=$(date "+%s.%N")
echo "Elapsed: $( Tend-Tstart |bc -l) nanoseconds"

However, notice that in both cases (time and date), you have some overhead, as the system takes some time to run the starting/stopping commands. The overhead is smaller for time, thanks to it being a builtin. So the three-digit limit is there for a reason: the rest is just random garbage.

See also this similar question, though the accepted answer contains a little bug currently.