Speed up ZI measurements

[jenshnielsen]

mainly in fast_measurements.py

Two issues at least

The plotting is somewhat slow. This is partially because of the very long x axis from the zi (at least 4096) points. We could consider an option to down sample that axis as it's not always needed with than many data points. Also optionally reducing the number of plots will help with this.

The datas acquisition is slower than it should be.
For instance I find that running the Keysigth ramp at 2.8 KHz 1000 averages should take 0.36 s (1/2.8 KHz) but in reality it takes around a second.

The measured timings are (setting up the ZI involving multiple writes to the instrument (or data server) takes 0.15 S, Actually taking the data 1 s, Readout 0.05 s. So this is in large dominated by the actual measurement. It not clear if this is due to a too short rearm time of the ZI (currently 60 micros but increasing it to 100 microseconds did not seem to have an effect), some other limitation of the ZI, or something else entirely.

Without plotting the entire measurement takes 16.6 S for 10 rows. Not measured how the remaining overhead is distributed between stepping qdac, datasaving and other factors.