This week the Adapteva team put together more Parallella prototypes and configured eight of these into a Beowulf cluster, running an MPI program in order to test that an 8-way configuration using the final boards will work as expected.
Once operation of the cluster had been confirmed it was dismantled, those boards should now be on their way to backers and we’re now eagerly awaiting reports of the first community experiences with Parallella in the wild!
Last week I mentioned in brief how Mark Dewing (markd) had been giving thought to Python support and had done some initial work with the simulator. This week I’m delighted to announce that Mark has agreed to lead on the Python effort, and he’s posted some further thoughts to the forum. If Python is your language of choice, get involved!
Running the MPI test program.