clock.8 1.1 KB

12345678910111213141516171819202122232425262728293031323334353637383940414243
  1. .\" $Id$
  2. .TH CLOCK 8 "$Date$" "(c)1994 Larry McVoy" "LMBENCH"
  3. .SH NAME
  4. clock \- report timing attributes
  5. .SH SYNOPSIS
  6. .B clock
  7. .SH DESCRIPTION
  8. .B clock
  9. reports the timing interval,
  10. .BR ENOUGH ,
  11. timing overhead
  12. .BR TIMING_OVERHEAD ,
  13. and loop overhead
  14. .BR LOOP_OVERHEAD .
  15. The required timing interval is determined experimentally to
  16. ensure that the timing results are accurate to within 1% in
  17. the spite of the system clock resolution and measurement
  18. overheads.
  19. It is necessary to measure
  20. .B ENOUGH
  21. experimentally because system clock resolution varies widely
  22. across systems, from as large as 10 milliseconds to as small as
  23. 1 microsecond, a range of four orders of magnitude.
  24. The timing overheads and loop overheads are used by the
  25. timing harness
  26. .I timing(3)
  27. to remove the various overheads from the measured times.
  28. .SH OUTPUT
  29. The various results are reported in micro-seconds, e.g.,
  30. .IP ""
  31. \fCENOUGH=5000
  32. .br
  33. TIMING_OVERHEAD=7.000000
  34. .br
  35. LOOP_OVERHEAD=0.000004
  36. .br
  37. # version 2.0\fR
  38. .ft
  39. .SH ACKNOWLEDGEMENT
  40. Funding for the development of
  41. this tool was provided by Sun Microsystems Computer Corporation.
  42. .SH "SEE ALSO"
  43. lmbench(8), lmbench(3), timing(3).