[Python-Dev] (time) PEP 418 glossary V2 (original) (raw)

Victor Stinner victor.stinner at gmail.com
Tue Apr 24 12:38:21 CEST 2012

Monotonic ---------

This is a particularly tricky term, as there are several subtly incompatible definitions in use.

Is it a definition for the glossary?

 C++ followed the mathematical definition, so that a monotonic clock only promises not to go backwards.

The "C++ Timeout Specification" doesn't have any monotonic anymore. It has a steady_clock, but it's something different. http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2010/n3128.html#time.clock.monotonic

 In practice, that is not sufficient to be useful, and no Operating System provides such a weak guarantee.  Most discussions of a "Monotonic Clock" will also assume several additional guarantees, some of which are explicitly required by the POSIX specification.

What do you mean for POSIX? The definition of CLOCK_MONOTONIC by the POSIX specification is:

"The identifier for the system-wide monotonic clock, which is defined as a clock whose value cannot be set via clock_settime() and which cannot have backward clock jumps. The maximum possible clock jump shall be implementation-defined." http://pubs.opengroup.org/onlinepubs/000095399/basedefs/time.h.html

time.monotonic() of the PEP 418 gives the same guarantee (cannot go backward, cannot be set), except for "system-wide" (Python cannot give this guarantee because of Windows older than Vista).

 The tradeoffs often include lack of a defined Epoch or mapping to Civil Time,

I don't know any monotonic with a defined epoch or mappable to the civil time.

and being more expensive (in Latency, power usage, or duration spent within calls to the clock itself) to use.

CLOCK_MONOTONIC and CLOCK_REALTIME have the same performances on Linux and FreeBSD. Why would a monotonic clock be more expensive?

 For example, the clock may represent (a constant multiplied by) ticks of a specific quartz timer on a specific CPU core, and calls would therefore require synchronization between cores.

I don't think that synchronizing a counter between CPU cores is something expensive. See the following tables for details: http://www.python.org/dev/peps/pep-0418/#performance

CLOCK_MONOTONIC and CLOCK_REALTIME use the same hardware clocksource and so have the same latency depending on the hardware.

Victor

More information about the Python-Dev mailing list