Global fits with NANOGrav data sets

[Former user]

The new noise modeling implemented by NANOGrav uses parameters such as ECORR that are included as part of the par files. When these noise model parameters are recognized inside of doFitAll(), the typical fitting routine of tempo2 is rerouted to getTempoNestMaxLike() circumventing any of the global fitting apparatus in tempo2 allowing for things like clock signal measurements or the formation of A+Ax time series. It would be ideal if NANOGrav could continue to use the new noise models that have been implemented and regain the ability to form clock signals, A+Ax time series, etc. . .

With attached files, try command

tempo2 -f B1855+09.par B1855+09.tim -f B1937+21.par B1937+21.tim -f B1953+29.par B1953+29.tim -global clock.par

When I run this with a version of tempo2 from 2014.11.1, everything executes without any issues, a function called ifunc.dat is created that supposedly contains the best-fit clock signal, but it is identically zero. That is the problem.

[Michael Keith]

This will be resolved by the NEWFIT routines. Similar to issue #2

[Michael Keith]

Not sure who submitted this, but progress is being made. The new fit routines are now starting to work, and I will begin to merge the getTempoNestMaxLike stuff into t2fit to get a unified interface. Related to work for issue #13

[Michael Keith]

I believe that this is resolved as of 2016.09.0