Folding Forum

Kebast wrote:

NoMoreQuarantine wrote:In an effort to make a predictive ranking of GPU performance, I took the specs of all of the desktop GPUs since 2016 for both AMD and NVIDIA ...

I think you should at least go back one more generation. Several of the Nvidia 900 series cards would fall toward the middle of your list, and can be found cheap on ebay.

Personally I don't think that's necessary because - and that's my only point for improvement - there is no rank for efficiency. Like TFLOPs/Watts.
Imho it's not useful to buy a cheap 'old' GPU which drives your electricity bill high.

For many long-term folders efficiency of the GPU is most important.

MeeLee wrote:

Juggy wrote: There are quite a few motherboards that support more than 2x PCIe 8x. X399 and X299 supports 16x/8x/16x/8x in 4 slots for instance (differenmt combo's of this as well) simultaneously on PCIe 3.0.

On paper, yes.
But you have no idea how many motherboards I had to return, because they had issues running more than 2 GPUs. Some ran 3 (on paper 4), but once I added another GPU, they would either not boot, or the GPU wouldn't be recognized (more the former).

Interesting, I have run tri and quad SLI in more than one configuration in the past and run quite strenuous benchmarks for 3dmark scores from as far back as a 6GHz+ clocked e8600 on dice and hardly ever had board or CPU issues.

GPU in SLI yes, but standalone cards just about no issues. More recently I have run 8 x 6 card RX580 mining rigs with PCIe risers and also never experienced a problem

Maybe the luck of the draw.

foldinghomealone2 wrote:

Kebast wrote:

NoMoreQuarantine wrote:In an effort to make a predictive ranking of GPU performance, I took the specs of all of the desktop GPUs since 2016 for both AMD and NVIDIA ...

I think you should at least go back one more generation. Several of the Nvidia 900 series cards would fall toward the middle of your list, and can be found cheap on ebay.

Personally I don't think that's necessary because - and that's my only point for improvement - there is no rank for efficiency. Like TFLOPs/Watts.
Imho it's not useful to buy a cheap 'old' GPU which drives your electricity bill high.

For many long-term folders efficiency of the GPU is most important.

While I understand that thinking, it would take a long time to offset energy costs when looking at spending $700+ on a recent RTX card.
Also, I wouldn't want someone thinking that a GTX 1030 was a good buy when better performing cards are the same price on ebay.

In my case, I'm not in the market for a new video card for gaming, and didn't want to spend much on additional hardware for folding. I live where energy is fairly cheap, it made much more sense for me to spend a small amount on old cards now. I spent a bit of time researching performance of various cards looking at the FAHBench results, comparing price, and watts. I also wanted to stay within a small budget.

If the 3000 series was out now, I probably wouldn't be writing this

For me after a lot of analysis I've settled on getting the "70" version of the previous series so currently that's the 1070. They can be picked up off eBay for under $200, use a relatively modest 140W or so and produce decent output - 800Kppd under Core22.

If you wear one out, it's not a big loss and with decent cooling that is unlikely.

When the 3xxx range come out I will look at moving to 2070 cards once their prices fall.

1070 is a good card, I had it myself till I upgraded to 1080ti.
However I think (but I'm not 100% sure) that a 1660 super is a better card for a similar price and you have warranty.
But that really depends on your region and it's price tags

Kebast wrote:
While I understand that thinking, it would take a long time to offset energy costs when looking at spending $700+ on a recent RTX card.
Also, I wouldn't want someone thinking that a GTX 1030 was a good buy when better performing cards are the same price on ebay.

In my case, I'm not in the market for a new video card for gaming, and didn't want to spend much on additional hardware for folding. I live where energy is fairly cheap, it made much more sense for me to spend a small amount on old cards now. I spent a bit of time researching performance of various cards looking at the FAHBench results, comparing price, and watts. I also wanted to stay within a small budget.

If the 3000 series was out now, I probably wouldn't be writing this

RTX 2060 Super Gaming OC from Gigabyte, $400. Excellent value for money for Folding

Mod Edit: Fixed Quote Tags - PantherX

NoMoreQuarantine wrote:In an effort to make a predictive ranking of GPU performance, I took the specs of all of the desktop GPUs since 2016 for both AMD and NVIDIA from here https://www.techpowerup.com/gpu-specs/ and placed them in tables sorted by FP32 TFLOPS at boost frequency. The calculation for the TFLOPS is TFLOPS = FP32 * Frequency * 2. The last two columns show watts/TFLOPS & dollars/TFLOPS at boost frequency. The prices are the lowest prices for a new video card I could find on Amazon or Newegg. This ranking assumes that memory bandwidth has negligible impact on FAH performance.



Please let me know if you see any errors in my methodology. If there aren't any error, I think this is about as good as I can do from theoretical standpoint and I'll focus on gathering real-world data which should supersede these tables in the future.

First let me preface this by saying I'm really not trying to crap on your efforts. To the contrary, I appreciate that you are trying to find good data to help people in picking GPUs for the task at hand, and their personal preferences/desires.

But in doing so, I think you are not seeing that there seems to be no single performance metric (I could be wrong, maybe one of the FAH benches?) that indicates folding performance. All the charts showing FP32 operations are great for any use that relies mostly on FP32 operations. But the charts https://docs.google.com/spreadsheets/d/ ... tput=html# in this spreadsheet show that FP32 is not the answer for folding metrics. I would say overall that having read quite a few threads, looked at data, etc since joining here there is not one clear cut data point that really helps in calculating PPD returns. If there was (and once again, there might be) such a data point everyone would use it. But I'm not sure there is, other than looking at data and overall averages of actual PPD returns. From there, everyone could make decisions on their choices and preferences that suit their own needs and desires.

Until proven otherwise, I think the metric that will have the most impact is going to be based on real world returns. I've been around computer geekery long enough to know that you can't always look at charts and determine the "best" for much of any particular task.


As stated before, I appreciate your efforts. I think they could help a lot of people. But I think it's a task that would take a lot of time and effort, and probably take direct user input in a similar format to the link from Overclockers. Even beyond that, the maintenance on such a spreadsheet would be crazy. It might make sense to see if it could be made without pricing and such input, but the formulas in place. That way anyone using it could input current pricing and go from there.


Oh, and one error I noticed just because it's a card I'm considering. The 1650 Super is actually a TU116 card.

There is a bug in the drivers for the older AMD cards that prevents them from folding some projects, the bug does not affect Navi GPUs.

If someone (or more folders) would distribute their recent HFM-logs I could try to make a database.

BobWilliams757 wrote:I think you are not seeing that there seems to be no single performance metric (I could be wrong, maybe one of the FAH benches?) that indicates folding performance. All the charts showing FP32 operations are great for any use that relies mostly on FP32 operations. But the charts https://docs.google.com/spreadsheets/d/ ... tput=html# in this spreadsheet show that FP32 is not the answer for folding metrics.

You are correct, this is a complex multi-variable problem. Far more complicated than I initially thought it was
That said, I have been learning a ton by going through this exercise and I am refining my process as I learn more. There is an open question about how much memory bandwidth affects performance on FAH. Also, while FAH uses mostly FP32 for calculations, it also makes use of some amount of FP64 (maybe even FP16, but I haven't heard anything about that). Additionally, PPD performance varies with the size of the projects available.

BobWilliams757 wrote:But I'm not sure there is, other than looking at data and overall averages of actual PPD returns.

This is the goal long term.

BobWilliams757 wrote:But I think it's a task that would take a lot of time and effort, and probably take direct user input in a similar format to the link from Overclockers. Even beyond that, the maintenance on such a spreadsheet would be crazy. It might make sense to see if it could be made without pricing and such input, but the formulas in place. That way anyone using it could input current pricing and go from there.

You ain't kidding. Good suggestion.

BobWilliams757 wrote:Oh, and one error I noticed just because it's a card I'm considering. The 1650 Super is actually a TU116 card.

Fixed! Also, I noticed while reviewing that the Pascal GPUs have double the number of FP32 & FP64 cores than I gave them

foldinghomealone2 wrote:If someone (or more folders) would distribute their recent HFM-logs I could try to make a database.

did I miss how we provide them to you?

HaloJones wrote:

foldinghomealone2 wrote:If someone (or more folders) would distribute their recent HFM-logs I could try to make a database.

did I miss how we provide them to you?

No, but unfortunately I don't know what is a good way to share the HFM-exports.

Is there a free service that several/different users can upload their exports so I can download them?

foldinghomealone2 wrote:No, but unfortunately I don't know what is a good way to share the HFM-exports.

Is there a free service that several/different users can upload their exports so I can download them?

Google Forms https://docs.google.com/forms/u/0/?ftv=1 and allow attachments https://support.google.com/a/users/answer/9308632?hl=en

They would all get added to a folder in your drive.

Thanks, I was not aware of that.

Maybe give it a try and upload an HFM-export.
https://forms.gle/4QGhU2ifzPyUnCny9

foldinghomealone2 wrote:Thanks, I was not aware of that.

Maybe give it a try and upload an HFM-export.
https://forms.gle/dpBT7S498D55MpPJ7

Sure, I just set up HFM.NET let me run it for a while to get something to export and I'll add it.

Why does the form require an email address?

Edit: sent it