TR Forums

I'm just curious. I've read a few posts here that sound a bit down on folding, this forum seems much quieter than it used to, and some folks have even been shutting down their farms.

I would like to know if anybody in authority has said anything about the progress of the overall project on a % complete basis.

Is it even worth it for somebody to get into it now, especially for those who might soon be building a new quad-core or hex-core i7 system with a couple of new and fast GPUs? Or is it not even worth the bother?

I don't think you can meaningfully measure percentage completion on a project like this. It's not like (say) the projects distributed.net promotes, where there's a well-defined target. With Folding, the results learned from existing WUs guide future investigations, i.e. they determine what types of WUs they issue down the road. It's an open-ended sort of thing.

The shutdown of my Folding clients was a pragmatic decision, driven by the hot weather this past spring/summer and the cost of electricity (hot weather is especially problematic because Folding increases the load on the A/C as well, so you get hit on the electricity costs twice). With the huge improvements over the past few years in CPU and GPU power management, the "if the system is going to be on anyway you might as well put the cycles to good use" argument doesn't carry as much weight either, since the system consumes significantly less power when it is idling.

With the arrival of colder weather I will be turning a few systems back on, since the additional electricity usage will be helping to heat my house instead of fighting with my A/C.

I'll probably turn the client on my work desktop back on as well. I moved to another area about a months ago, and this area is on the same cold air feed as the data center next door. As a result we actually have to run the heat (which is electric) in here to maintain reasonable temperatures even in the summer, so that electricity would be getting used anyhow. (Yes, I agree that's goofy... but I did not design the climate control systems on this floor, I just have to live with them!) This is in contrast to my old office, where it was always too hot...

I still believe in the overall goals of the Folding@home project, but my participation lately has been tempered by practical considerations.

I've shut down Folding on all my machines. Electric bills almost doubling during spring/summer months was a factor like JBI's . Additionally, there doesn't seem to be a lot of interest here at TR on the forums for it, which contributed to my decision to shut down my machines.

I still plan to participate in the annual Frankenbot challenge...

(Here's the thread from last year)

For me it is because super computers these days are so fast.

Just two or three years ago, and especially like 5 years ago, the FAH network was way way more powerful than the fastest super computer, but these days many super computers are way more powerful in FLOPS and also way way more efficient because they actually have an interconnect. So for me I feel like FAH is not as useful to the scientific world as it used to be.


Also FAH in general has been in decline since late 2010.

mph_Ragnarok wrote:

For me it is because super computers these days are so fast.

Just two or three years ago, and especially like 5 years ago, the FAH network was way way more powerful than the fastest super computer, but these days many super computers are way more powerful in FLOPS and also way way more efficient because they actually have an interconnect. So for me I feel like FAH is not as useful to the scientific world as it used to be.

That doesn't really make sense as a justification. Those fastest supercomputers are hugely expensive; only government research labs can afford them, and it is unlikely that smaller research projects like F@h could afford to lease time on them (if that is even an option). The F@H network is probably still more powerful than any dedicated supercomputers outside of NSA/DOE laboratories.

The project has been structured such that a fast interconnect isn't needed. Since each WU is self-contained, a fast interconnect would be of little or no benefit.

Thanks for the responses guys. When the colder weather comes, maybe I'll decide to go online. I need to learn more about the project to be sure that it aligns with my own .. um.. "morality", for lack of a better term.

I had read one or two threads where somebody who was running older systems for several years suddenly got left in the dust when a person with newer hardware went online and started racking up WUs faster by an order of magnitude. Sort of like the F@H version of the "rage quit". Not really an imporant issue, I just find human dynamics to be a fascinating subject.