As before, all the dot products of M1, M2, M1* and M2* can also be
shifted out of the loops for large speed increases.

Regards,

Edward


On 20 June 2014 17:42,  <[email protected]> wrote:
> Author: tlinnet
> Date: Fri Jun 20 17:42:56 2014
> New Revision: 24212
>
> URL: http://svn.gna.org/viewcvs/relax?rev=24212&view=rev
> Log:
> Moved the calculation of the matrix exponential out of for loops for ns mmq 
> 3site mq.
>
> Task #7807 (https://gna.org/task/index.php?7807): Speed-up of dispersion 
> models for Clustered analysis.
>
> Modified:
>     branches/disp_spin_speed/lib/dispersion/ns_mmq_3site.py
>
> Modified: branches/disp_spin_speed/lib/dispersion/ns_mmq_3site.py
> URL: 
> http://svn.gna.org/viewcvs/relax/branches/disp_spin_speed/lib/dispersion/ns_mmq_3site.py?rev=24212&r1=24211&r2=24212&view=diff
> ==============================================================================
> --- branches/disp_spin_speed/lib/dispersion/ns_mmq_3site.py     (original)
> +++ branches/disp_spin_speed/lib/dispersion/ns_mmq_3site.py     Fri Jun 20 
> 17:42:56 2014
> @@ -62,7 +62,7 @@
>  # relax module imports.
>  from lib.float import isNaN
>  from lib.dispersion.ns_matrices import rmmq_3site, rmmq_3site_rankN
> -from lib.linear_algebra.matrix_exponential import matrix_exponential
> +from lib.linear_algebra.matrix_exponential import matrix_exponential, 
> matrix_exponential_rankN
>  from lib.linear_algebra.matrix_power import square_matrix_power
>
>
> @@ -150,6 +150,16 @@
>      # Z- matrix component.
>      m2_mat = rmmq_3site_rankN(R20A=R20A, R20B=R20B, R20C=R20C, dw_AB=dw_AB - 
> dwH_AB, dw_AC=dw_AC - dwH_AC, k_AB=k_AB, k_BA=k_BA, k_BC=k_BC, k_CB=k_CB, 
> k_AC=k_AC, k_CA=k_CA, tcp=tcp)
>
> +    # The M1 and M2 matrices.
> +    # Equivalent to D+.
> +    M1_mat = matrix_exponential_rankN(m1_mat)
> +    # Equivalent to Z-.
> +    M2_mat = matrix_exponential_rankN(m2_mat)
> +
> +    # The complex conjugates M1* and M2*
> +    M1_star_mat = conj(M1_mat)
> +    M2_star_mat = conj(M2_mat)
> +
>      # Loop over spins.
>      for si in range(NS):
>          # Loop over the spectrometer frequencies.
> @@ -162,20 +172,24 @@
>                  # Loop over the time points, back calculating the R2eff 
> values.
>                  for i in range(num_points_i):
>                      # The M1 and M2 matrices.
> -                    M1 = matrix_exponential(m1_mat[si, mi, oi, i])    # 
> Equivalent to D+.
> -                    M2 = matrix_exponential(m2_mat[si, mi, oi, i])    # 
> Equivalent to Z-.
> +                    # Equivalent to D+.
> +                    M1_i = M1_mat[si, mi, oi, i]
> +                    # Equivalent to Z-.
> +                    M2_i = M2_mat[si, mi, oi, i]
>
>                      # The complex conjugates M1* and M2*
> -                    M1_star = conj(M1)    # Equivalent to D+*.
> -                    M2_star = conj(M2)    # Equivalent to Z-*.
> +                    # Equivalent to D+*.
> +                    M1_star_i = M1_star_mat[si, mi, oi, i]
> +                    # Equivalent to Z-*.
> +                    M2_star_i = M2_star_mat[si, mi, oi, i]
>
>                      # Repetitive dot products (minimised for speed).
> -                    M1_M2 = dot(M1, M2)
> -                    M2_M1 = dot(M2, M1)
> +                    M1_M2 = dot(M1_i, M2_i)
> +                    M2_M1 = dot(M2_i, M1_i)
>                      M1_M2_M2_M1 = dot(M1_M2, M2_M1)
>                      M2_M1_M1_M2 = dot(M2_M1, M1_M2)
> -                    M1_M2_star = dot(M1_star, M2_star)
> -                    M2_M1_star = dot(M2_star, M1_star)
> +                    M1_M2_star = dot(M1_star_i, M2_star_i)
> +                    M2_M1_star = dot(M2_star_i, M1_star_i)
>                      M1_M2_M2_M1_star = dot(M1_M2_star, M2_M1_star)
>                      M2_M1_M1_M2_star = dot(M2_M1_star, M1_M2_star)
>
>
>
> _______________________________________________
> relax (http://www.nmr-relax.com)
>
> This is the relax-commits mailing list
> [email protected]
>
> To unsubscribe from this list, get a password
> reminder, or change your subscription options,
> visit the list information page at
> https://mail.gna.org/listinfo/relax-commits

_______________________________________________
relax (http://www.nmr-relax.com)

This is the relax-devel mailing list
[email protected]

To unsubscribe from this list, get a password
reminder, or change your subscription options,
visit the list information page at
https://mail.gna.org/listinfo/relax-devel

Reply via email to