Info Welcome to the Tom's Hardware F@H Team Thread (Team 40051)!

Page 5 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
That's the thing, on stuff like x264 it is half as fast at the same frequency (I checked), not a fifth the performance, than the i5. On the same token, if that was true the Ryzen score would be 10 times that of the i5, not the more reasonable 3.
 
Jun 15, 2020
2
1
15
0
With the Pfizer vaccine out at 90% effectiveness, can I stop folding for Covid? My office has been very hot this whole year with all the machines running 24/7!
 
With the Pfizer vaccine out at 90% effectiveness, can I stop folding for Covid? My office has been very hot this whole year with all the machines running 24/7!
it's not out of trials. The Chinese vaccine got stopped at final trials in Brazil. The Pfizer one uses ARN transport to work, it's a first - barely even tested on animals.
 

RTX 2080

Estimable
Anyone on the team still active and watching here? I fired up a few GPUs and have another on the way, so my production is back up for now.
Watching yes, active no.

I folded from February 2020 to January 2021 but had to stop because it got hard to explain to my wife why it was ok to spend an extra $250 a month on electricity. I enjoyed the competitive aspect and enjoyed trying to stay as high on the top 10 contributers list as I could, but after that I only folded on cold nights to keep my bedroom warm. If I moved to a place that had rooftop solar I'd fire it back up though.
 
Completely understand that - I cut back quite a bit in the summer (a lot easier to handle in the winter as I heat with a heat pump). What systems do you have to add up to $250/month? Or do you have expensive electricity?
 

RTX 2080

Estimable
Completely understand that - I cut back quite a bit in the summer (a lot easier to handle in the winter as I heat with a heat pump). What systems do you have to add up to $250/month? Or do you have expensive electricity?
Well I was folding with both a RTX 2080 Founder's Edition and a EVGA 2060 KO. The 2080 I bought when I built my PC and the KO I bought a few months after I started folding because there were rumors that some KOs (being based off a RTX 2080 chip with some stuff disabled) could fold just as good as a 2080 at less than half the price. Unfortunately, my KO inherited a large percentage of a RTX 2080's appetite for power without much of a performance gain over a regular 2060, but at least it could hold a strong overclock compared to my 2080 FE; surprisingly, the fans were a lot quieter too.

Anyway, the two GPUs running together in the one tower drew about 450 watts and my overclocked i7-9700k keeping them fed probably drew another 100 watts. Underclocking them to achieve better points per watt could have been done, but I was super motivated to stay in the top 10 contributers for Tom's Hardware, and with the increase in interest in folding due to Covid-19, I had to fold flat out to keep up with the heavy hitters. Doing this 24/7 quickly added up, especially when the A/C also had to be run 24/7 during the already hot summer months.
 
Last edited:
Still folding here as always.
My total power bill is $250 or less.
And that is with 3 computers and 4 video cards running, plus AC to cool the house.
You must have some pretty expensive electricity. Mine is 11.6c per kilowatt hour.
 
Ah, yeah. CA makes sense. I spend probably $250+ on computer power and cooling, but that is a bunch of computers on f@h and boinc. At the moment I am folding with 3 Radeon VII and a Titan Xp. I still have a decent amount on boinc though, including a 2p Epyc 7601, a 2p 2698v4, and a 2p 2697v3, and some other smaller single socket CPUs.
 

RTX 2080

Estimable
Ah, yeah. CA makes sense. I spend probably $250+ on computer power and cooling, but that is a bunch of computers on f@h and boinc. At the moment I am folding with 3 Radeon VII and a Titan Xp. I still have a decent amount on boinc though, including a 2p Epyc 7601, a 2p 2698v4, and a 2p 2697v3, and some other smaller single socket CPUs.
Interesting. What made you go for the three Radeon VIIs? Are they particularly good for folding?
 
I bought one originally when they were new, and the other two when they were on fire sale for $500 new. I bought them mostly for boinc, as some projects make heavy use of double precision (Milyway@home) and the VII is the only recent consumer card by AMD or NVIDIA not to have artificially limited or terrible DP performance. They actually are decent folders though. I have mine underclocked and undervolted (should be pulling 100W under stock, maybe 120-150W per card?) and they still get around 2.4M PPD (on larger projects).
 
Apr 4, 2020
5
0
510
0
Due to a number of longstanding chronic issues with the previous threads, we present to you all a brand new thread, stickified with some extra strong adhesives, for our illustrious folding team! You can find the original thread here, which may still be useful, albeit with outdated information.

Welcome to the Tom's Hardware F@H Team Thread (Team 40051)!

With coronavirus in the air (quite literally), it's time for us all to come together and do our part. Don't just sit at home and twiddle your thumbs during the quarantine, fold for science! With your help, and enough processing power, we can collectively fold proteins and find a cure for the coronavirus. No this isn't science fiction, this is Folding@home!

The Tom's Hardware Community staff has created a new folding@home team for the Tom's Hardware Community. We've included detailed instructions on how to join the Tom's Hardware Folding@home team below. But first, what is folding?

What is Folding@home?
Folding@home is a project started by Stanford University that allows anyone to donate spare processing cycles (from your CPU and GPU) to compute complex calculations in a massive distributed research network. You can learn more about folding@home here.

Among the more well-known diseases that benefit from this research are Alzheimer’s, Huntington’s, Parkinson’s disease, AIDS, many cancers, and most recently, the novel coronavirus, covid-19. If you’ve ever known anybody who has been afflicted by any of those, then you know how devastating it can be for them and their families. This is your chance to help combat these illnesses and all it takes on your part is leaving the program running while you’re asleep or at work! Every contribution will help in advancing treatments and, hopefully, will eventually lead to cures.

How can I join the Tom's Hardware Folding@home Team?
  1. First head to the official folding@home website and download the correct client based on your OS. Here is a direct link to the download page: https://foldingathome.org/alternative-downloads/
  2. Install the client. If you are windows, clicking advanced will allow you to decide when the software will start. We recommend for you to star the software at login.
  3. Start folding@home!
  4. A browser window should automatically open with the client running. Make sure to click the "Set up an identity" bubble before clicking start folding
  5. On the next screen set your name and team number. Make sure to join team 40051. Your name can be anything you wish, but we recommend for you to use your forum username if available. At the time of this writing the passkey generator was not working. The servers are really busy and there are not enough work unit's for all current participants. Some user's have had to wait a day or more for passkeys. To learn more about passkeys, click here.
With this set up in place your computer will provide work units to the team. Keep in mind, closing the browser does not stop folding. Folding is a background process. To stop folding navigate to client.foldingathome.org and click "Stop Folding".


FAQ
TO check the team's current stats, and your own, head to http://folding.stanford.edu/Stats.


All these letters are just short type for a full words:

WU = Work Units
PPD = Points per Day
PPW = Point per Watt or Points per Week.
TPF = Time per Frame
WTF! = Way to fold!


  • Bonus point is when you turned in WU earlier than what the deadline says and you earn extra points because it.
  • Qualified WU's mean certain WU's that can be turned in early for bonus points.
  • Except for some older Single core CPU WU's and older GPU WU's, All WU are qualified for bonuses.
While the F@H group doesn't "officially support nor oppose overclocking", you can overclock and gain more points.

The catch is however that F@H is more sensitive to instability than most other stress programs will detect. Even if your computer is "Prime 95 stable" for a week straight, it isn't 100% fool-proof on the cpu when it comes to folding.

F@H recommend using these stress programs that closely mimic how F@H run for both your GPU and CPU.

Conclusion
Wow you made it this far?! Kudos to you! Be sure to sign up now for the team. In the coming days there will be an announcement for a friendly competition between this and two other PC hardware focus sites. Stay tuned!

Happy folding :giggle:
 
Apr 4, 2020
5
0
510
0
I have done some work about a year ago which I never followed up on. Here are some caculations I made. If all the processing for F@H were done on efficient graphic cards (say RTX2080) I calculated F@H uses about $5 million dallrs per month in electricity. Obviously, the actual amount is much higher.

Using an ammeter to calculate my own power usage on several computers, I can say that using the CPU as a slot for F@H on computers with efficient graphics units (say newer than 3 years) will actually decrease points per day while increasing power. Unfortunately, the default F@H chooses is to use the CPU as a slot.

Personally I think the F@H organization should discourage people from using old computers (or even all CPUs) in their operation as they simply are not "green", unless they replace electrical resistance heating of a building.

One might also be interested in this link:

https://www.nature.com/articles/d41586-020-03348-4
 

RTX 2080

Estimable
I have done some work about a year ago which I never followed up on. Here are some caculations I made. If all the processing for F@H were done on efficient graphic cards (say RTX2080) I calculated F@H uses about $5 million dallrs per month in electricity. Obviously, the actual amount is much higher.

Using an ammeter to calculate my own power usage on several computers, I can say that using the CPU as a slot for F@H on computers with efficient graphics units (say newer than 3 years) will actually decrease points per day while increasing power. Unfortunately, the default F@H chooses is to use the CPU as a slot.

Personally I think the F@H organization should discourage people from using old computers (or even all CPUs) in their operation as they simply are not "green", unless they replace electrical resistance heating of a building.

One might also be interested in this link:

https://www.nature.com/articles/d41586-020-03348-4
The purpose of distributed computing is to use hardware that would otherwise sit idle to solve problems. While it would be nice to do so only on efficient hardware, the reality is that most people willing to volunteer their hardware don't have the most up-to-date tech. By limiting the hardware that can be used for folding@home to only newer more efficient hardware, there would be significantly less processing power available and research would proceed much more slowly. Folding@home has no incentive to limit their software to only efficient computers as that would limit their ability to crunch numbers. Since it is the person volunteering their hardware who has to pay for electricity, it is up to them to be incentivized to save electricity and buy more efficient hardware.

Folding@home defaults to CPU folding for (as best as I can tell) two reasons:
  1. Probably 90% of computers out there don't have a discrete graphics card, they only have integrated graphics built into their CPU.
  2. There are some folding operations that can only be done on a CPU as a GPU doesn't have the ability to perform all the same kinds of mathematical calculations that a CPU does. Thus, there is a reason to do CPU folding period, no matter how efficient it is compared to GPU folding.
Folding@home is for scientific reasearch, not environmentalism. If they were concerned about energy usage, they wouldn't outsource projects at all, they would build their own supercomputers with the latest and most efficient hardware, but this is not within their focus nor their ability.
 
Apr 4, 2020
5
0
510
0
The purpose of distributed computing is to use hardware that would otherwise sit idle to solve problems. While it would be nice to do so only on efficient hardware, the reality is that most people willing to volunteer their hardware don't have the most up-to-date tech. By limiting the hardware that can be used for folding@home to only newer more efficient hardware, there would be significantly less processing power available and research would proceed much more slowly. Folding@home has no incentive to limit their software to only efficient computers as that would limit their ability to crunch numbers. Since it is the person volunteering their hardware who has to pay for electricity, it is up to them to be incentivized to save electricity and buy more efficient hardware.

Folding@home defaults to CPU folding for (as best as I can tell) two reasons:
  1. Probably 90% of computers out there don't have a discrete graphics card, they only have integrated graphics built into their CPU.
  2. There are some folding operations that can only be done on a CPU as a GPU doesn't have the ability to perform all the same kinds of mathematical calculations that a CPU does. Thus, there is a reason to do CPU folding period, no matter how efficient it is compared to GPU folding.
Folding@home is for scientific reasearch, not environmentalism. If they were concerned about energy usage, they wouldn't outsource projects at all, they would build their own supercomputers with the latest and most efficient hardware, but this is not within their focus nor their ability.
My entire purpose is to try and make the FAH organization more efficient-not to prove anybody right or wrong. I think there are numerous things the FAH organization could do to be more efficient. They should have a separate section on their website for efficiency. 1) encourage people to send them their old graphic cards they no longer want to be donated to people who are active FAH users that might improve their efficiency drastically. 2) Make it clear that people who have good, separate graphics cards may be better off not running their CPU; personally, I would not hazard a guess at to what that percentage that is, but I would guess it to be much higher than 10%. 3) since we do not know how much CPU-specific folding needs exist for FAH, it is impossible to make any suggestions on the allocation of CPU verses GPU resources.
 
Long time folder here.
Once graphics cards were able to do folding, work started shifting over to them.
But a lot of the projects were started with CPU work units that can not be ported to GPU work units. The nature of their coding.
Some can be either.
And most newer ones are GPU only.
I support efficient folding and run the most PPD/watt graphics cards of each generation.Along with efficient power supplies.
GTX 960 FTW
GTX 1060 6GB Gamer
GTX 1070 SC
RTX 3060TI
I no longer CPU fold.
But the massive amount of giving of resources we saw with the covid research was awe inspiring.
I saw estimates of beating the top 10 super computers in the world. For many months. And a peak of beating the top 100.
With that much raw power given to researchers, companies, academics, Universities, research centers.....
We now have vaccines for covid with unseen before speed. Months, not years.
The research continues.
 
I agree that it is an area where they could communicate better. How much of the CPU work can/is duplicated on the GPU? How much isn't feasible/efficient to run on the GPU? I still CPU fold a bit for testing purposes, but generally just GPU fold. One challenge right now is that the GPU market is crazy (and has been for some time), so any semi-modern ("efficient") GPU is worth a fortune right now. So many aren't upgrading as they don't have a choice. My main GPUs are Radeon VIIs, Titan Xp/1080Ti, 1070/1060, and 3070.
 

ASK THE COMMUNITY