I've been creating BOINC project brainstorming threads over at steemit, I'll be updating this thread once a month when the brainstorming thread is reposted on steemit.

Here's the link to the 3rd brainstorming thread: https://steemit.com/boinc/@cm-steem/brainstorming-boinc-projects-003

We've had suggestions for TensorFlow, Apache Spark (BARK), Sensor arrays, Proof of Scientific Research, Decentralization of the BOINC project web server, Full-Node distribution projects, TOR/I2P Proof of Bandwidth, Hashcat over BOINC, etc.

Don't like Steemit? Post your ideas in this thread!

No so much a BOINC project but I have suggested a MPI (Message Passing Interface) BOINC client. It would run on say a Beowulf cluster head-node and control the compute nodes using MPI. BOINC communicates with the science apps via shared memory so that would need to be translated to use MPI.

The end result would be a BOINC client that could run on a cluster without having to have the BOINC client installed on each node.

---

MarkJ