Align data by timestamp

[godelstheory]

Current methodology converts timestamps to seconds since day start. Though convenient and quick, it runs into the issue when midnight occurs. When moving to larger scale automation of the experiments, where an experiment could easily be performed with a local clock near the midnight hour, errors wills occur.

• In data_munge.R the methods should either merge the data sets by timestamp, or ensure that the conversion of seconds accounts for the day.

[godelstheory]

Actually, Intel Power Gadget doesn't appear to add day information! This might be added if the data is collected past midnight. Needs to be tested.

• Determine if day information is added to IPG output, if the clock goes past a midnight.

[godelstheory]

Maybe RDTSC solves the problem? From what I can glean, it doesn't.