• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Invidia Drivers effect Folding preformance?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
You mean, does it raise/lower your PPD? It can, yes.

The problem with 'finding the best one' is that, I don't think there's a way to know what's 'best' because testing the same WUs over the same time is impossible. So, you may see 'improvement' but what if it's the WUs and not the driver?
 
You mean, does it raise/lower your PPD? It can, yes.

The problem with 'finding the best one' is that, I don't think there's a way to know what's 'best' because testing the same WUs over the same time is impossible. So, you may see 'improvement' but what if it's the WUs and not the driver?
Thank you that good to know Maby if I watch the WU'S when the new driver installs, I can keep a netbook of the points and the name of the WU. If that makes sense.:unsure:
 
Thank you that good to know Maby if I watch the WU'S when the new driver installs, I can keep a netbook of the points and the name of the WU. If that makes sense.:unsure:
You could do that, but I wonder if there's a more fool proof way to get the WU's for a given period than trying to manually record them.
 
I wouldn't spend time on this. Bigger fish to fry to gain more PPD, I'd imagine.

But as inferred, I'd look to see if there's a log that stores WUs/time it took/PPD and compare them. I just don't know if that exists.
When you mean bigger fish to fry and don't say Linux. :mad: What are they and can I download any of them?:)
 
Every now & then a new driver will bork F@H. It hasn't happened in a while though. Word spread quickly & everyone just rolled back to an older version. I seem to remember it taking a couple updates before nvidia fixed the problem, there were about 3 driver releases we had to skip. Otherwise the newest driver for our cards is probably best because we use the PCs for F@H and other tasks.
 
Every now & then a new driver will bork F@H. It hasn't happened in a while though. Word spread quickly & everyone just rolled back to an older version. I seem to remember it taking a couple updates before nvidia fixed the problem, there were about 3 driver releases we had to skip. Otherwise the newest driver for our cards is probably best because we use the PCs for F@H and other tasks.
Thank you for the information.(y)
 
Every now & then a new driver will bork F@H. It hasn't happened in a while though. Word spread quickly & everyone just rolled back to an older version. I seem to remember it taking a couple updates before nvidia fixed the problem, there were about 3 driver releases we had to skip. Otherwise the newest driver for our cards is probably best because we use the PCs for F@H and other tasks.
Exactly. NVIDIA has really done some work to make their drivers work for F@H and is perhaps a main reason why they produce better than AMD GPUs. AMD had been partners with F@H in the past but they have let it go. Hence, GPU talk in the F@H community is almost all NVIDIA. AMD cards do produce but I think that the ppd/$ win goes to NVIDIA.

If you're not folding, AMD cards can be very competitive with NVIDIA but if you are folding, NVIDIA is the best bang for your buck. So while drivers may negatively impact production, NVIDIA is our best bet.
 
I wouldn't spend time on this. Bigger fish to fry to gain more PPD, I'd imagine.

But as inferred, I'd look to see if there's a log that stores WUs/time it took/PPD and compare them. I just don't know if that exists.
You can skim this data from f@h log, but would have to make a script or something to pull the data without a bunch of leg work. I might investigate some options for this in the near future.

You could do that, but I wonder if there's a more fool proof way to get the WU's for a given period than trying to manually record them.
You should be able to back up a new wu and let it fold most of the way. Then make your change, swap the wu file with the mostly completed one and let it fold it. Then you could compare the same wu.
 
You can skim this data from f@h log, but would have to make a script or something to pull the data without a bunch of leg work. I might investigate some options for this in the near future.


You should be able to back up a new wu and let it fold most of the way. Then make your change, swap the wu file with the mostly completed one and let it fold it. Then you could compare the same wu.
Yeah....I'm just not sure the juice is worth the squeeze... especially for the person who asked initially. Like, get the dollars in order before the pennies, ya know? Get the right card in the first place (more PPD and less power use where applicable), type of thing.... but...so it goes! :)
 
Rule of thumb in regards to fah, do not upgrade drivers unless you have to. If it ain't broke don't fix it. :beer:
This is the sauce right here, when I ran half a dozen boxes barely breaking like 300k a simple driver update would cripple everything. Ended up just rolling back to a known stable driver and never updated unless there was a massive improvement for the CUDA stuff.

The past 8 or so years though, this hasn't been an issue during the times I'm able to get back in to Folding, especially with RTX cards. NVIDIA seems to have really stepped up their game, and I think we can thank coin mining for that.
 
Back