This will be later in the future, as of right now I do not have Internet, and can only game on 4G from my phone.
On the topic of this, how much bandwidth does folding use?
I have 2 phones, 1 with 4G for gaming, and 1 with unlimited 3G @ 200kb/ps.
 
 
I have experience in crypto mining, so I'd like so see what I can do here.
 
 
Questions:
 
What is the "speed" at which you fold called. For mining it was kh/s.
 
 
Is AMD or Nvidia faster?
 
 
How do you get evga bucks doing it?
 
 
How "Fast" can these cards fold at:
 
HD 7850
GTX 760
GTX 970
GTX 980
 
 
Hold much evga buck a month could I make with the cards:
HD 7850
GTX 760
GTX 970
GTX 980

I currently own the hd 7850 and GTX 760 and I plan on upgrading. If the 980 is significantly faster at folding, and if the investment is worth it if I'd get a decent amount of return, then I would get a GTX 980 over the 970.
 
 

PPD is the primary measurement of folding power, as it represents the amount of value you're adding to FAH as a whole, and is normalized across all projects running on a given piece of hardware (or at least it's supposed to be- architectural differences between the benchmark hardware and all of the hardware out there means this may not be 100% true for your hardware).  TPF is only really relevant when discussing a specific project, as different projects can have wildly differing runtimes for a variety of reasons.
 
As to bandwidth, folding doesn't use much.  The work is broken up into discrete chunks called Work Units.  The client downloads one, has the proper core (the software that actually runs the calculations) process the WU, then uploads the completed work and starts the process over again.  Most WUs are at most 15 MB down and up, and ones that size tend to take several hours to complete, if not days on slower hardware.  Stanford has been beta testing a new distribution method where you essentially get a ton of rather short WUs, i.e. effectively streaming the work, but a) it doesn't appear to be anywhere near release to the public and b) I think each micro-WU is on the order of kB, and would still take a few minutes even on high-end hardware.
 
Currently the 970 and 980 are the kings of GPU folding (when there isn't some wonkiness going on, that is), with both doing somewhere in the range of 250k-330k PPD.  Right now the GPU cores that run on Maxwell are not optimized for that architecture, and as a result both the 970 and 980 are within about 5% performance of each other.  Given the large shader disparity between the two, I doubt they'll be so close once the cores have some Maxwell optimizations in them.  No one can say when, or even if, those improvements will be released, and obviously there's no telling what sort of a performance boost the 980 would see and whether it was worth the price premium.
 
I think a 7850 will do 50k PPD and a 760 will do around 60-70k.  I wasn't getting any solid Google hits, but that seemed to be the right range.  In the last couple of years performance between AMD and Nvidia has been roughly equal at most price points, a stark change from the few years before that when AMD was terrible.

Yeah, I'm not in it entirely for the money, but 10 EVGA bucks is 10 EVGA bucks. Even when I did crypto mining, I mined dogecoins just for fun. I still haven't sold any of them.
 
I would be nice to see in the future, reading in some school text book years from now, that we were the ones to help this.

Yeah, I think most of us save them for a good sized discount on upgraded hardware, which in turn can be used for more folding.  Basically you can get a new GPU once a year for $120 off.

Some very good info given above. We also have a folding section under the community tab at top of forum, with useful info.
It is also where your evga folding points for the month are shown, and also where you claim your bucks. We will be glad to have you aboard, when ever your ready. Welcome.
http://www.evga.com/folding/ 

Tropickz
Folding can also be done with the CPU, so if you have hyper threading your golden.

HT doesn't help that much.  FAH pegs the FPUs in a chip pretty hard as it is, so there isn't a lot of leftover processing power for extra threads to use.  The performance gain is usually around 15% over no HT, which is nice if you have it, but not all that great if you were looking at it strictly from a cost perspective.
 
HT does help for feeding GPUs, though.  Dropping from 8 to 6 threads on an i7 won't drop your folding PPD all that much while likely being enough to feed two GPUs without bottlenecking them.  If you had an i5, you'd probably have to reserve 2 full cores, which would really crimp your performance.  Of course, CPU PPD is so much lower than GPU PPD right now (and for the foreseeable future) that it doesn't make too much of a difference either way.

I addressed 970 vs. 980 in the third paragraph of my first post.
 
CPUs can have a wide range of PPD figures since some projects scale better than others.  From what I just checked, an 8350 can be as low as 12k PPD and as high as 28k.  I'd expect 20k to be about the norm.
 
GPUs are definitely king unless you have a multi-socket server, and even there that advantage will be going away come February 1.

Okay, so I let it run overnight, and I have 20,000 PPD
 
I have the slider to full, and my GPU is at 99%, but my CPU is not working.
The status says "ready" but it isn't downloading anything.
 

How do I set that, or remove it?
 
In the setup, it was at -1, I just changed it to 8, and still no change. Do I have to restart the client?

 

 
 

 

So should I finish this one then, or let it go? I had it running overnight.

Just for comparison's sake, I completed [3] 9201 wu's  on a 780ti.
Avg. TPF= 2:15  PPD Shows : 203,000 and credit= 32,000 each.
I think they took around 3 1/2 hours per WU to complete.
Hopefully as the WU runs the TPF decreases as the PPD increases.

What is a Work Unit?


Is that what the green bar represents? % completed on only 1 work unit?

Well, twice now I forget to finish it. I had it paused and resumed several times.
 
 
Is there a way to abandon the current WU?

dave851

iNoToRiOuS
Well, twice now I forget to finish it. I had it paused and resumed several times.
 
 
Is there a way to abandon the current WU?

Well its bad practice to abandon wu's. And that finish button is for if you want to finish your current wu and not start another wu.

 
How would I start another? I still haven't gotten the GPU up to full speed, and the CPU won't even start.