• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Tools for Programming Distributed Computing Projects?

NoEcho

NoEcho

2[H]4U
Joined
Aug 14, 2001
Messages
3,250
Anyone know what tools are used to created the Distributed Computing projects we crunch?

I'd like to get into the guts of the settings for some projects of my own.
 
There are many frameworks and/or libraries to use in building a DC projects. Really depends on what you are looking for.

Clusters - MPI, Corba, ACE, etc....
Single MP Machines - pthreads, J2EE containers( java ), boost, etc....
 
There's also the BOINC client for end users to run if your doing a public program.
 
Hmmmm.

I mean apps to create the equivalent of master work units, apps to tweak how info is brought in, how it analyzes the data... like that.

Like with Seti At Home (they still have that, right?), it'd be cool to cross analyze wavelengths, locations where the data was gathered, waveform stuff. So, instead of just downloading apps with predefined tasks, being able to tweak the kind of work we're doing with the data.
 
Are you saying you want to explore existing DC projects in more detail or create own project?
 
I'm talking about tweaking existing projects by using the same tools as the people who developed the work units. Obviously the result wouldn't be admissible to the original project but it could result in separate insights from the same data.
 
How about a threaded MPI version of the fah cores. Use MPI to give off large chunks to machines in the cluster and let threads work on the large chunk give to a machine. Best of both worlds :D
 
There has been a cluster version of the FAH client tested, but the final verdict was that it made no sense to pursue it further at that time. The only way it worked efficiently, was if the cluster nodes used high-speed interconnects, and those interconnects cost more than the entire rest of the cluster.

BOINC has a development kit for those that want to develop a new project with it, but I fear the data that you want to analyze for an existing project, would not be included in a simple work unit, because they want to save bandwidth, and it wouldn't be necessary. That data would only be needed when the work units were created.

Best to contact the project that interests you directly, imo.
 
You must log in or register to reply here.
Back
Top