[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: parcellfun and rand
From: |
Jaroslav Hajek |
Subject: |
Re: parcellfun and rand |
Date: |
Sat, 4 Sep 2010 14:23:50 +0200 |
On Sat, Sep 4, 2010 at 6:55 AM, Muhali <address@hidden> wrote:
>
> Unlike cellfun, parcellfun appears to re-initialize the rand function in each
> call. This is counter-intuitive.
>
> octave:1> disp(parcellfun(1, @(x) rand, cell(1,1), "UniformOutput", false,
> "verboselevel", 0))
> {
> [1,1] = 0.32068
> }
> octave:2> disp(parcellfun(1, @(x) rand, cell(1,1), "UniformOutput", false,
> "verboselevel", 0))
> {
> [1,1] = 0.32068
> }
> octave:3> disp(cellfun(@(x) rand, cell(1,1), "UniformOutput", false))
> {
> [1,1] = 0.32068
> }
> octave:4> disp(cellfun(@(x) rand, cell(1,1), "UniformOutput", false))
> {
> [1,1] = 0.26259
> }
>
This has been already fixed; please upgrade your package. Parcellfun
now broadcasts the original seed to each forked process and
reinitializes based on that and the thread number. Unfortunately,
there is no feasible way to make parcellfun behave identically to
cellfun in this regard.
hth
--
RNDr. Jaroslav Hajek, PhD
computing expert & GNU Octave developer
Aeronautical Research and Test Institute (VZLU)
Prague, Czech Republic
url: www.highegg.matfyz.cz
- parcellfun and rand, Muhali, 2010/09/04
- Re: parcellfun and rand,
Jaroslav Hajek <=
- Re: parcellfun and rand, Francesco Potortì, 2010/09/04
- Re: parcellfun and rand, Francesco Potortì, 2010/09/04
- Re: parcellfun and rand, Muhali, 2010/09/05
- Re: parcellfun and rand, Muhali, 2010/09/07
- Re: parcellfun and rand, Francesco Potortì, 2010/09/07
- Re: parcellfun and rand, Muhali, 2010/09/07
- Re: parcellfun and rand, Francesco Potortì, 2010/09/07
- Re: parcellfun and rand, Jaroslav Hajek, 2010/09/08
- Re: parcellfun and rand, Jaroslav Hajek, 2010/09/08