[task #7807] Speed-up of dispersion models for Clustered analysis -- June 06, 2014 - 22:31

Follow-up Comment #2, task #7807 (project relax):

This seems not possible to implement.

If for example you have 4 spins, at 3 frequencies, the shape of
fact = r20a - r20b - k_BA + k_AB
=>
Psi = fact**2 - dw2 + 4.0*pA*pB*kex**2
=>
eta_scale * sqrt(Psi + sqrt_psi2_zeta2)

would be 12.

If the number of cpmg_frqs is 20, then this becomes impossible.
etapos = eta_scale * sqrt(Psi + sqrt_psi2_zeta2) / cpmg_frqs

The shapes does not fit.

Also, the number of cpmg_frqs can be different per frequency.

And you would anyway need to split up per spin.

Unless, one forms some matrix of correct size.
20x12 X 12x20 = 20x20 matrix

But that would alter the model to much!

Also looking at:
r20_kex = (r20a + r20b + kex) / 2.0
this will have shape 12

And then
R2eff = r20_kex - cpmg_frqs * arccosh( fact )
this will be shape 12 minus shape 20 * shape 12.






    _______________________________________________________

Reply to this item at:

  <http://gna.org/task/?7807>

_______________________________________________
  Message sent via/by Gna!
  http://gna.org/