Project Future

Message boards : News : Project Future
Message board moderation

To post messages, you must log in.

AuthorMessage
Tomáš Brada
Project administrator
Volunteer developer
Avatar

Send message
Joined: 3 Feb 19
Posts: 276
Credit: 106,040
RAC: 162
Message 3244 - Posted: 26 Mar 2019, 13:45:17 UTC

I was asked about the project future. You can read about plans in this thread.

Any, If at all, continuation of this project will not happen before summer.

Makarova has more computation available with her experiments. We could run those.

DesA has some interesting fluid simulation ideas. I wrote him what properties should the boinc apps exhibit to run comfortably. We will see what he prepares until then. He has some other ides, not only the simulation, more computational or even neural-networkey.

Next is a Gridcoin microgrid. It is an ambition to provide various levels of commercial boinc hosting and support. Development of boinc apps is part of the ambitions. Customers would give us their non-boinc programs and problem-space and we will make it run on boinc. Non-profit "customers" can be backed by the gridcoin whitelist. This server will be the testbed for this initiative.

There are some improvement to the boinc codebase that I want to make, and this server will be the testbed. For example torrent or ipfs asset distribution that would ease up the load on boinc file-server. And making the BOINC API easier to use (proper library, out-of-tree builds, docs).

Next is a CPU fuzzer. It's goal is to discover what undocumented instructions are available in the various processors owned by crunchers.

I would like to explore possibility of building and CI of software on the boinc clients. Including distributing the boinc apps themselves in source form, to be compiled on the crunchers' computer. In linux, programs were meant to be compiled on the target computer since ever, so why fight with static linking and binary incompatibility?

Another inspiration is PlanetLab non-boinc project. I would like to use boinc to quicly deploy experimental P2P software, see how well it forms a p2p network in close to real world controlled scenario and undeploy it equally easily.

Hentai@Home is a interesting project, because it is unlike many, storage and bandwidth intensive instead of computation. I realize that home networks are not optimized for heavy uploads.

As you see thre are many ideas. Probably I will choose two in the summer. And of course, nothing will happen without consent of the crunchers.
ID: 3244 · Rating: 0 · rate: Rate + / Rate - Report as offensive

Message boards : News : Project Future

©2019 Tomáš Brada