Re: r23952 - /branches/disp_spin_speed/lib/dispersion/ns_cpmg_2site_3d.py -- June 15, 2014 - 10:34

Hi Troels,

Is this actually faster?  I just wrote a script to test this at
https://gna.org/task/index.php?7807#comment199.  The log file with the
results is also attached.  This is also archived at
http://article.gmane.org/gmane.science.nmr.relax.devel/6240.  If you
look at that, I think I can show a way to make the numeric models up
to twice as fast.  And that is simply to use the out argument for the
dot product.  At least it will minimise the time required for the dot
products, and maybe remove this as the bottleneck.

Regards,

Edward



On 15 June 2014 08:53,  <tlinnet@xxxxxxxxxxxxx> wrote:
> Author: tlinnet
> Date: Sun Jun 15 08:53:38 2014
> New Revision: 23952
>
> URL: http://svn.gna.org/viewcvs/relax?rev=23952&view=rev
> Log:
> Changed linked matrix/vector inner products into chained dot expressions.
>
> 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_cpmg_2site_3d.py
>
> Modified: branches/disp_spin_speed/lib/dispersion/ns_cpmg_2site_3d.py
> URL: 
> http://svn.gna.org/viewcvs/relax/branches/disp_spin_speed/lib/dispersion/ns_cpmg_2site_3d.py?rev=23952&r1=23951&r2=23952&view=diff
> ==============================================================================
> --- branches/disp_spin_speed/lib/dispersion/ns_cpmg_2site_3d.py (original)
> +++ branches/disp_spin_speed/lib/dispersion/ns_cpmg_2site_3d.py Sun Jun 15 
> 08:53:38 2014
> @@ -56,6 +56,7 @@
>  # Python module imports.
>  from numpy import dot, fabs, isfinite, log, min, ones, ndarray
>  from numpy.ma import fix_invalid, masked_less_equal, masked_where
> +import numpy as np
>
>  # relax module imports.
>  from lib.dispersion.ns_matrices import rcpmg_3d
> @@ -138,9 +139,7 @@
>
>          # Loop over the CPMG elements, propagating the magnetisation.
>          for j in range(2*power[i]):
> -            Mint = dot(Rexpo, Mint)
> -            Mint = dot(r180x, Mint)
> -            Mint = dot(Rexpo, Mint)
> +            Mint = Rexpo.dot(r180x).dot(Rexpo).dot(Mint)
>
>          # The next lines calculate the R2eff using a two-point 
> approximation, i.e. assuming that the decay is mono-exponential.
>          Mx = Mint[1] / pA
>
>
> _______________________________________________
> relax (http://www.nmr-relax.com)
>
> This is the relax-commits mailing list
> relax-commits@xxxxxxx
>
> 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