Now GN3 specifically states that for both initial verification and for periodic testing that insulation resistance testing should be undertaken as a block measurement of a group of circuits (such as distribution boards), with simple installations measured as an entire unit.
The preference is to measure insulation resistance between live conductors and live conductors and live conductors and earth (with cpcs connected), but there is the option to test between earth and the live conductors connected together if the first method is not practical.
I have only just realised that initial verification permits group testing, though I would normally use this for EICRs.
From the point of view of the regulations this is permitted and correct.
For fault finding and diagnosis over time the testing of individual circuits provides a much better indication of the condition of circuits and allows detailed investigation and assessment to be undertaken. It will also provide a much better indication of the state of the installation on a circuit level.
If you have measured IR on separate circuits and levels are low then it is possible that the overall insulation resistance measurement may be below that specified in the regulations (overall resistance is lower than the lowest value of parallel resistances), but I believe that if all individual circuits meet the regulatory limits then the installation is satisfactory (this is perhaps suggested in GN3, but not stated).
I would normally test individual circuits on initial verification and group testing on periodic inspections.
However either way round can be done in either circumstance and provide an allowable result that can be recorded.
Perhaps a quick calculation may be useful if individual IR results are low just to check total resistance as a reference but it is very easy to drop below the limits on lots of circuits (e.g 10 circuits at 15MΩ would drop very close to 2MΩ overall).