[Omp] Reproducing results from randomizer

[ThanhVu H. Nguyen - Gmail]

I use the MT rng since it is suggested by my professors for being very
popular and runs quite well.  I was not aware it's not thread-safe.  Do
you have any suggestion a thread-safe rng so I can convince my prof to
change the RNG for this project ?  



--- C J Kenneth Tan -- OptimaNumerics <cjtan at OptimaNumerics.com> wrote:

> ThanhVu,
> 
> > --- "Meadows, Lawrence F" <lawrence.f.meadows at intel.com> wrote:
> > 
> > > Look at http://sprng.cs.fsu.edu/ 
> > 
> > I was suggested to assign a random num gen to each thread but I was
> not
> > sure exactly how it is done. Some code example perhaps ?
> >  I try to avoid using a different number generator since all my
> > projects have been using the Mersenne Twister rng.  Thanks
> > 
> 
> View the RNG as a disc.  If you are not using distinct sequences,
> then
> you would just have the parallel threads all sampling from the same
> disc.  So if you use a serial RNG that generates just 1 sequence, as
> far as the randomized simulation is concerned, you may just as well
> have 1 thread running.
> 
> Unfortunately, MT19937 is a serial RNG.  What you need is a parallel
> RNG.
> 
> 
> Kenneth Tan
>
--------------------------------------------------------------------------
> News: OptimaNumerics Enhances Performance on SGI Altix Supercomputers
>
--------------------------------------------------------------------------
> C J Kenneth Tan, PhD
> OptimaNumerics Ltd                    Telephone: +44 798 941 7838
> E-mail: cjtan at OptimaNumerics.com      Telephone: +44 871 504 3328
> Web: http://www.OptimaNumerics.com    Facsimile: +44 289 066 3015 
>
--------------------------------------------------------------------------
> 


tvn,

ThanhVu H. Nguyen