I wonder if this board is cost-effective for non-gpu datascience, especially in terms of power consumption.
I have read a paper on several single-board computers, saying that they were almost there, despite being ARM processors.
I have read a paper on several single-board computers, saying that they were almost there, despite being ARM processors.
https://petewarden.com/2014/06/09/deep-learning-on-the-raspb...
Aside from that, I don't think the performance will be acceptable. I've been running Spark and Jupyter on a 20-core, 5 node cluster and it's acceptable, but only useful for debugging (because stuff happens slower).
If I wanted to build a cluster out of them, I'd probably have to break out the USB connections in some way, probably meaning a whole custom PCB or something... Given that there are more powerful Pi clones out there I don't get the appeal for this particular usecase.
There are several actual papers on google scholar though.