Anyone here a participant in F@H?

Lil306

If so, what team do you fold for?

Custom PC - 35947

If you haven't got a team please join up for the healthy competition

esuhl

F@H? Is that the thing where you increase your CPU load to analyse radio signals in an attempt to find aliens?

Don't mean to be a party pooper, but is the increase in power/pollution/CO2 emissions really worth it, given the infintesimally small liklihood of finding little green men?

paddyrg

You're thinking of SETI at Home. Folding at Home is about folding proteins and is about disease prevention/cure.

That said, most processors waste so many clock cycles whilst you read an email or whatever that those syscles may as well be used by SETI or F@H

suicidebob

paddyrg wrote: »

You're thinking of SETI at Home. Folding at Home is about folding proteins and is about disease prevention/cure.

That said, most processors waste so many clock cycles whilst you read an email or whatever that those syscles may as well be used by SETI or F@H

The clock cycles aren't 'wasted'. They just don't exist.

It's like saying your trainers are being wasted because they're sitting in the cupboard.

asbokid

There's no sense to what SETI and F@H claim to be doing. I heard that SETI has closed down because it was all so pointless. The cost of distributing the data to be processed by home PCs is much more than the cost of installing streaming processor units at a single site and crunching the data locally.

A 3072 core GPGPU runs at close to 1MHz and costs $700. Built as a PCI card, you can have half a dozen GPGPUs on a single PC motherboard. All of the GPU cores execute the same processor instructions at the same time. This is different to a multi core x86 where each core can run its own separate process.

And so the gains from parallelisation with GPGPU come from chopping up the dataset so that each of the cores gets to crunch its own bit of data using the same operations. This is known as the SIMD architecture - single instruction multiple data. The multi-core x86, by contrast, is an MIMD architecture.

Parallelising was, of course, the whole idea behind SETI and F@H, but it was achieved through geographic distribution. The same code was run on lots of different machines, but each of those machines was allocated a different chunk of the dataset.

But with the availability of low-cost many core GPGPU devices, the distributed parallelisation paradigm of SETI and F@H are pointless now.

For those who like conspiracy theories, try this one for size...

My suspicion is that SETI and F@H were not what they seemed. The source code was never revealed, and the nature of the dataset was highly obfuscated, too. So how does anyone know for sure that the data that was sent to your PC by the SETI servers really did come from radiotelescope readings?

With no source code, and no explanation of the nature of the data, all we had to go on was trust.

For those who participated in the SETI project, perhaps you were unwittingly aiding in a secret CIA/MI6 programme to crunch PGP passphrases in a distributed brute-force attack on the encrypted communications of an unpopular overseas government.

Just a thought!

[Deleted User]

SETI is Open Source and many parts of F@H are too.

Lil306

asbokid wrote: »

Just a thought!

Do you think everything is a conspiracy ?

F@H does have a point, it's curing diseases.

It may not happen in this lifetime, but it's always worth making use of unused cycles on the computer for better causes.

I used to do SETI but it gets boring watching the screens and you go a bit giddy seeing a jump in the spikes, whereas the F@H project benefits people in the future

F@H has a personal meaning to me anyway, like a lot of others someone has been affected indirectly or directly through something which F@H could one day cure. I fold for my dad he died from cancer.

Some people just fold for the competition.

asbokid

anewhope wrote: »

SETI is Open Source and many parts of F@H are too.

All of SETI, the distributed client code and the search algorithms were originally closed source.

The idea that distributed parallel processing could be used in a brute-force attack on encryption keys is not a new "conspiracy theory"..

This SANS paper was mooting the possibility nearly a decade ago.

Whether you could commandeer an existing framework and use it for that purpose is another matter. There aren't any technical reasons why you could not.

It doesn't help that SETI are pretty weird people who figured they had some right to use our CPU cycles to process their mysterious "work units" but otherwise kept their dataset exclusive and secret from other parties, including other researchers.

Hardly in the spirit of open science.

http://www.sans.org/reading_room/whitepapers/awareness/distributed-computing-unstoppable-brute-force_1330