Model Buildings.ThermalZones.EnergyPlus_9_6_0.Examples.SmallOffice.IdealHeatingCoolingWinter fails verification

[AndreaBartolini]

The variable ECoo.y of the model Buildings.ThermalZones.EnergyPlus_9_6_0.Examples.SmallOffice.IdealHeatingCoolingWinter fails the verification vs the Dymola reference file.
That variables integrates the sum of the five outputs coo[x].conPID.y in coo block, each one multiplied by a gain, and the signals coo[x].conPID.y pass the verification.

Detailed analysis is available in OpenModelica/OpenModelica#12809

The question here is to decide if the accumulation of the errors of the signals coo[x].conPID.y (multiplied by the relevant gains) shall or not be verified vs the Dymola reference file, maybe this verification is too hard to get by using different compilers and solvers.

Keeping @casella in the loop.