[Gretl-users] Re: Publication in "econometrics" using Gretl
Am 09.11.2020 um 17:38 schrieb Marcin Błażejowski:
On 09.11.2020 17:23, Riccardo (Jack) Lucchetti wrote:
> On Mon, 9 Nov 2020, Marcin Błażejowski wrote:
>
>> On 09.11.2020 16:44, Sven Schreiber wrote:
>>> BTW, what happened to the plans to use MPI parallelization for BMA and
>>> BACE?
>> BACE is parallelized, but needs documentation...
> Is it?
>
> I couldn't find the string "mpi" anywhere in the code. Plus, I tried
> running the example script with 200000 replications and I saw only 1
> CPU going.
Ok - is parallelized, but since it needs documentation it is not uploaded.
Maybe you can use the lockdown for that ;-)
Seriously, it would be great to publish that code and have that
accelerated. Also for others to have an example how to achieve the
speedup. IMHO the initial MPI documentation doesn't have to be complete.
OTOH if you have a 7-month old I could imagine the lockdown doesn't
necessarily provide lots of time for that?!
cheers
sven