diff options
| -rw-r--r-- | xmds2/realistic_Rb/realistic_Rb.xmds | 4 |
1 files changed, 2 insertions, 2 deletions
diff --git a/xmds2/realistic_Rb/realistic_Rb.xmds b/xmds2/realistic_Rb/realistic_Rb.xmds index 5f7e437..75e7464 100644 --- a/xmds2/realistic_Rb/realistic_Rb.xmds +++ b/xmds2/realistic_Rb/realistic_Rb.xmds @@ -36,7 +36,7 @@ We are solving - dE/dz+(1/c)*dE/dt=i*eta*rho_ij, where j level is higher then i. + dE/dz+(1/c)*dE/dt=i*eta*rho_ij, where j level is higher than i. Note that E is actually a Rabi frequency of electromagnetic field not the EM field in xmds terms it looks like dE_dz = i*eta*rhoij - 1/c*L[E], here we moved t dependence to Fourier space @@ -406,7 +406,7 @@ <integrate algorithm="ARK45" tolerance="1e-5" interval="Lcell"> <!--<integrate algorithm="SI" steps="200" interval="Lcell"> --> <!--RK4 is good for large detunings when frequency like term are big, it does not try to be too smart about adaptive step which ARK seems to make too small--> - <!--When ARK45 works it about 3 times faster then RK4 with 1000 steps--> + <!--When ARK45 works it about 3 times faster than RK4 with 1000 steps--> <!--<integrate algorithm="RK4" steps="100" interval="1.5e-2">--> <!--SIC algorithm seems to be much slower and needs fine 'z' step tuning and much finer time grid--> <!--For example I had to quadruple the time grid from 1000 to 4000 when increased z distance from 0.02 to 0.04--> |