numpy performance and random numbers
Gib Bogle
g.bogle at auckland.no.spam.ac.nz
Mon Dec 21 18:23:56 EST 2009
sturlamolden wrote:
> On 19 Des, 16:20, Carl Johan Rehn <car... at gmail.com> wrote:
>
>> How about mulit-core or (perhaps more exciting) GPU and CUDA? I must
>> admit that I am extremely interested in trying the CUDA-alternative.
>>
>> Obviously, cuBLAS is not an option here, so what is the safest route
>> for a novice parallel-programmer?
>
> The problem with PRNG is that they are iterative in nature, and
> maintain global states. They are therefore very hard to vectorize. A
> GPU will not help. The GPU has hundreds of computational cores that
> can run kernels, but you only get to utilize one.
>
> Parallel PRNGs are an unsolved problem in computer science.
My parallel version of ziggurat works fine, with Fortran/OpenMP.
More information about the Python-list
mailing list