Version 16 (modified by Sam Hocevar, 16 years ago) (diff)

0.47-2 packages

Help libcaca research projects using CPUShare

Image processing research sometimes requires expensive computations, for instance analysing the quality of an algorithm on thousands of different images. You can help the project by selling or donating your idle CPU cycles. We use CPUShare for this purpose: it provides all the infrastructure for reliable, secure and anonymous distributed computing.

Note: we are not using your data, only your CPU. CPUShare is secure and prevents malicious code from using anything else than the network socket.

Getting started

If you are using Debian or Ubuntu, just download one of the following packages:

Then download our sell order and put it in /etc/cpushare/. If your machine has several CPUs, you can place several copies of the same file in /etc/cpushare/, just make sure their names end in .cpu.

Finally, start the CPUShare daemons:

/etc/init.d/cpushare start

If you are not using Debian or Ubuntu, you can either try to install and run CPUShare from source (good luck!) or run this Live CD, either by burning and booting it, or by running it in an emulator (see more information on the CPUShare wiki):

How does it work?

Your machine's idle CPU cycles will be available to CPUShare users. There is no way to know whether your CPU will only be used for libcaca projects, but if you use our .cpu file, the CPUCoins earned by running other software will be credited to the libcaca project's account. We will in turn use these CPUCoins to buy CPU cycles on other machines; maybe yours, maybe not.

The CPUShare daemons run at a very low priority (nice -19 by default) but you can stop them at any time if the impact on your system is too important.

If you do not like giving CPUCoins directly to the libcaca project, you can create your own .cpu file by creating an account on cpushare.com and creating a new sell order.

What computations are currently running?

None at the moment. We are still porting our software so that it uses CPUShare efficiently.

Attachments (2)