Feature/subsecond time

[jlnav]

The Timer module now measures in milliseconds behind-the-scenes. libE_stats displays additional millisecond values for Start and End times:

Start: 2019-12-11 10:27:50.227 End: 2019-12-11 10:27:50.733

However, for compatibility with current timing mechanisms timer.elapsed and timer.total times are returned as seconds. If necessary, I can implement additional settings or parameters for milliseconds if we want to be more precise throughout the project.

[coveralls]

Coverage decreased (-0.04%) to 94.692% when pulling 867580c on feature/subsecond_time into 1bdf010 on develop.

[jmlarson1]

@jlnav Is it possible to replace the two uses of time with datetime? For example, using datetime.strftime and time_tuple which should be like time.localtime(), see:
https://docs.python.org/3/library/datetime.html#datetime.date.timetuple

Thoughts?

[jlnav]

@jlnav Is it possible to replace the two uses of time with datetime? For example, using datetime.strftime and time_tuple which should be like time.localtime(), see:
https://docs.python.org/3/library/datetime.html#datetime.date.timetuple

Thoughts?

Good point. Fixed.

[jmlarson1]

I don't think this is working as expected. When I run, I get the following output:

Worker     2: Calc     0: sim Time: 0.00 Start: 2019-12-11 00:00:00.405 End: 2019-12-11 00:00:00.406 Status: Not set
Worker     3: Calc     0: sim Time: 0.00 Start: 2019-12-11 00:00:00.406 End: 2019-12-11 00:00:00.406 Status: Not set
Worker     3: Calc     1: sim Time: 0.00 Start: 2019-12-11 00:00:00.409 End: 2019-12-11 00:00:00.409 Status: Not set
Worker     2: Calc     1: sim Time: 0.00 Start: 2019-12-11 00:00:00.408 End: 2019-12-11 00:00:00.408 Status: Not set
...
...

The time is not correct, and the Status is Not Set for some reason.

[shuds13]

It will be Not Set unless you are using the calc_status feature. I may be inclined to only write the milliseconds to file if run-times are smaller than some cut-off, as these lines can get very long - esp. if you include job run-times (as well as calc times).

[shuds13]

I'm also getting milliseconds only on my laptop (running forces test).

[jlnav]

Sorry about that. I think I just fixed the milliseconds-only issue.

[shuds13]

Yes, the time issue seems to be fixed now.

[shuds13]

I'm happy to pull this into develop. We may want to think about variable precision to output, maybe depending on runtimes. But that can be an optional follow up.