Folding@Home Support Thread

Well, I don’t know if this is the right place to put this, but I think that somebody at NVidia should have a talk with somebody at stanford and see how they can get this amazing humanitarian project running on the NVidia platform. If ATI can get it to work, why couldn’t NVidia?

I’m a proud owner of a 7600GS and a 6200TC… and love NVidia as a whole. However, when I think that I could be helping researchers cure cancer but some technicality is stopping us from achieving this goal, I have to wonder… why?

Many people are in the market looking to buy a new graphics card, and the new ATI card might be promising… but the 8800 is proven as the #1 card on the market right now. However, the ATI cards let you fold while you’re not gaming, and that for me is worth a lot, having lost relatives to various illnesses that can be helped by folding.

So there you have it. Why not work together, put out a press release promoting Nvidia’s part in curing disease?

Warmest regards,

Chris

I’d love to see nVidia join the Fold. I don’t game much anymore but, as ATi has shown w/ their GPU client, there are other ways to harness a video card’s horsepower.

I actually did run a BFG 8800GTS for about a month before giving up on the idea that I’d be able to make full use of the card. I sold it and “downgraded” to an HIS x1950xt which, while slower, Folds like a banshee. And I KNOW that the 8800 series should have the potential to blow at least that generation of ATi hardware out of the Folding water. So do it already! :D

Indeed. The Playstation 3 uses a graphics chip produced by Nvidia, the PS3 is now joining the Fold so why not Nvidia? With ATI joining the Fold it would look better on Nvidia in my opinion. It helps show to other people that the company cares. :)

I currently use the GeForce 6800 GT and I know this thing has potential in Folding and would love to be able to see that happen. ^_^

GPU Folding FTW!

Actually, the graphic chip in the PS3 isn’t being use to fold – it’s doing normal graphics. It’s the IBM cell chip that’s doing the real work.

A better comparison is with ATI who managed to fix their drivers work successfully as a physics processor. In that regard, Nvidia has done a good job of advertising how the hardware has lots of processing power than can be used for non-graphics tasks, but the drivers for DirectX simply fail to do the job. I could care less about the advertising – what I want to see is a good set of drivers.

I fully agree, both of my households PC’s and laptop are running F@H, as is my PS3.
All 4 of them are running nV based graphics as well. I would love to see the results of being able to include them it the project. With the jaw-dropping output of PS3’s Cell processor adding the processing potential of the RSX would really be useful.

Stanford’s GPU project has already teamed up with ATI on the matter and is seeing good results. Adding nVidia would only further a good cause. It would be nice to see two competitors come together to support this project. Not to mention the good PR to be gained.

nVidia, please consider this suggestion.

Supporters, please contact the Folding@Home team to suggest it to them aswell.

I will take this one step further. I am planning on buying 2 to 4 high end graphics cards in the next 2 months for Folding@Home. I don’t play games but am ready to put up some serious cash to build systems to run F@H and BOINC. (Under the guise that the “kids” need new computers. ;) )

nVidia: PLEASE make a public announcement that you are actively supporting F@H and BOINC. (And actually do it - not just an announcement.)

I am really trying to avoid buying a steaming pile of x1950’s :X

I thought GPGPU programming could be done through the openGL driver as well as the DX driver… true?

Would this mean that there might actually be some current use for an AGEIA card? I’ve been Folding for 2-3 years now fairly steadily although not at all for the past year as I have been deployed to Iraq; but I recently read about ATI bragging about its GPUs being able to Fold, so I googled for nVidia Folding and found ths forum. I currently own 3 Ti 4400 cards, 1 Ti4200, and 1 6800GT in 5 seperate comps - 4 of which have been used for Folding 24/7 when I was home. If I can get use out of all of my vid cards on the Folding machines I could more than double my efforts! Either way I’m proud to say that I have remained an nVidia fanboy all through the years and have never owned an ATI card. Go nVidia!

When the PhysX was first released, I tried finding some kind of information about how one might use it as a coprocessor for general purpose numerical work. Unfortunately, the Ageia website had no information about the low level capabilities of the chip, or whether they might be interested in people using it for scientific computing. Ageia is totally focused on Windows game programmers, and email queries about using the PhysX for non-game computing went unanswered. Their current SDK is Windows-only, and has a hefty licensing fee unless they think your program is going to drive huge sales of PhysX cards. Unless you are serious game programming house, the PhysX is pretty much useless.

Thankfully, this search eventually led me to GPGPU approaches, and CUDA, which has been much more approachable for people in scientific computing.

Perhaps you should try again? Or maybe get a large number of gamers who are also Folders to mass email them about it? It’s been a while now since the AGEIA came out and I don’t see them being bought in large numbers nor have I seen any games that take advantage of it being a dedicated physics card. Perhaps they’d be more open now to anything that might garner them appeal with the public at large and the advertising would be free… I’d do it myself, heck I’ll email them about it; but have no time to get a bunch of people motivated enough to email the company… MaxPC has a Folding team and their staff are fans of the idea of the AGEIA card… maybe if you got them interested in trying to bait AGEIA out of its shell… I’m not only thinking of the benefits to Folding, but with all of the increased interest - and maybe stellar Folding results - we might just see a deluge of new games finally taking advantage of the card’s capabilities. It’s worth a try, don’t you think?

CUDA meets my needs for now, and I don’t have to pull any teeth to get programming documentation. If Ageia wants to start selling cards to the scientific computing community, they know where to find us. :)

What is actually the problem in porting folding@home to CUDA? If it is suitable for ATI their CTM it should certainly be portable to CUDA, as that appears to be more flexible.

Time. The people who write Folding@Home would have to dedicate resources to doing the port. NVIDIA does not write Folding@Home – perhaps you should ask the Folding@Home team?

Mark

I was just curious if the issue was time or some showstopper bug, I didn’t mean it like ‘hurry up NVidia’. I know the responsibility of writing Folding@Home lies with Folding@Home.

Nvidia support for F@H would be great. Please add it.

Might be too late for all the people above but I’m gonna upgrade in a few months and if nvidia can’t fold, I’ll take an ATI. If I take an ATI, my next mobo will be Crossfire-ready, so…

You should bug mhouston about it. He is at Stanford on F@H project.

Just to add that Folding@home are working on the Covid19 cornavirus protiens. I would dedicate a few Jetson Nanos to this if could just help them to get it working on them - cmon this is your chance to shine in the worlds hour of need!

The number of hosts contributing to F@H has exploded by more than 10x over the past two weeks. The sustained throughput is currently about 470 (FP32) PFlops, making this the most powerful distributed super computer in the world, thanks largely to GPUs.

Sadly, the project’s servers cannot keep up although they already added several new ones. I have been folding on GPUs for team whoopass (#131015; the original CUDA folding team) since 2008, and my two GPUs are currently idling 60%-80% of each day because they cannot get work units and/or cannot report results due to server overload. Maybe NVIDIA could donate an additional server to F@H to speed things up?

As an alternative, I am now also contributing to Rosetta@home, which however is CPU only. They are tackling problems (the lowest energy protein configurations) that are complementary to those tackled by F@H. This project has doubled their computational throughput since the start of the coronavirus crisis, to 1.3 PFlops. FWIW, I was listed as predictor of the day at Rosetta@home for March 25, 2020:

https://boinc.bakerlab.org/rosetta/rah/rah_old_predictor_of_the_day.php

On the upside: F@H claims to have broken the exaflop barrier: https://twitter.com/foldingathome/status/1242918035788365830
On the downside: I haven’t been able to get any work units from F@H today, and 20+ attempts at uploading results for previously completed units have failed