In initialise.cu, the initial condition of a field, that is, Phi and PhiDot are being set with PreFFTInitPhi and PreFFTInitPhiDot.
We already have the option thermal_spectrum but it is insufficient, since
In the BE distribution, the mass of the field and the temperature are naively set to 1,
Normalization is naively set to get the dimension in units right,
PhiDot gets the same scale as Phi.
To improve,
Mass and temperature should be obtained as arguments, to be suitably used to calculate the distribution,
Normalization needs to be better considered.
Use the correct dispersion relation to get omega for given kx (assuming free field for now?).
Then, instead of setting Phi and PhiDot separately, set the two modes of the oscillation with BE distribution. (Maybe this is just same as giving PhiDot additional factor of omega?)
In
initialise.cu
, the initial condition of a field, that is,Phi
andPhiDot
are being set withPreFFTInitPhi
andPreFFTInitPhiDot
. We already have the optionthermal_spectrum
but it is insufficient, sincePhiDot
gets the same scale asPhi
.To improve,
omega
for givenkx
(assuming free field for now?). Then, instead of settingPhi
andPhiDot
separately, set the two modes of the oscillation with BE distribution. (Maybe this is just same as givingPhiDot
additional factor ofomega
?)