TheWolf

-ZS-Carpenter
Don't listen to that windbag. You don't need inside info or pour over the f@h forums to see what has been going on the last few years. Anyone who is not wearing PG blinders has seen what way they lean. Unless something is done with the new core the QBR for GPU is lip service. Offer a bonus then kill tpf Do they really think anyone believes they did proper testing on this? at all? If kepler is that much different it should have gotten its own core and not be sharing one with fermi. Just like the flake feast of the 762x WUs, 30% reduction in points, 8-12c increase in heat and unstable on some cards. good job testing those WUs out.

Using what is basically a hack to get kepler card to fold is not day 1 support. Pointing the finger at nvidia is just finding a scape goat. Either do something or phase out GPUs like the PS3. No point in supporting something half tail if its not what you want to run your simulations on.

and yea, i did have a bad day /rant

+1 good rant, took the words right out of my mouth.
As pointed out by someone else, if the beta team saw the reduction in points in there testing
changes should have been made to correct this problem or released notes letting Fermi owners
know not to use this new beta core. So it all falls back to PG not being upfront enough with
its donors base. They should know by now any leak of something new is going to be tested by
the donors that are trying to keep up with anything new. So why not be more informative upfront.
That would stop a lot of the out pour of bad feelings towards PG.

Edit: I will say if this is to be the new standard for GPU folding on Fermi cards, my cards will not be folding
until such time that correction are made to put back the correct optimization for Fermi base cards.

I do not think that statement is fair Wolf it was posted for all to read in the beta forum back in May and EVGA does have members that are part of the beta team, that should have been aware of it and warned others of the problem. The reason Stanford opened the beta team thread in the first placeI was so all could see what was being tested and be more transparent. I also believe the only reason it has been released at this time is because of the unrest among kepler folders and Stanford was trying to give them something to appease them,  there may be no cure for the screen lag on kepler cards other than what Stanford has done with the FAH_GPU_IDLE environment variable, (I do not know) Stanford is pretty much in a loose loose situation here and I would not like to be in there shoes at this point in time.
http://foldingforum.org/viewtopic.php?f=66&t=21571
 

the latest beta drivers from Nvidia are Cuda5 from what was posted at FF
 
As for beta team complaining about points on WU's or the core they are not allowed to do so. all they can do is give feedback on tpf so they can fix issues if any, because any increase in tpf could signal problems or an issue. although I can't see much difference with such reasoning (tpf is points... the lower the tpf the more points)
 
Once a core or WU goes public then points be address in the general forums.
 
as for 8057
http://foldingforum.org/viewtopic.php?f=66&t=22808
 
 
it's on the beta psummary
http://fah-web.stanford.edu/cgi-bin/fahproject.overusingIPswillbebanned?p=8057
 
 
 

on the points issue
http://foldingforum.org/viewtopic.php?f=66&t=21571&start=15#p222421
 
 

Grandpa_01
I do not think that statement is fair Wolf

Its about as fair as what we been getting with a list of things that need fixing
and is way to long to point out each and everyone. Only now that the natives are
restless do they speak. Good luck with fair on that one.
 

Grandpa_01

As far as the deep pockets statement goes that is a bunch of bull, do you not think that a folder that can afford a 4P farm can not afford a GPU farm that will make just as many ppd when the new QRB for GPU is released if it is most cost effective way to fold. Most of the leaders have followed the trends since day 1 of folding they go where there contribution can do the most good and that is measured by production which is measured by PPD. The PPD discrepancy will remain as long as it there is a points system. 

Kepler support will happen when it happens, In the mean time I am happy to see the = pay for = work come to be, but I am sure there will be some griping about it from all camps because inevitability somebody will put there tin foil hat on.

Fold on what ever happens, happens

My comment is not being received as it is intended. And if you are going to make a comment off of one sentence, than reading all of my posts would be a good idea. Because those following along would have a clear understanding as to what the comment is about. 
 
It is about my "Perception". We all know perception is not always reality. Now go ahead and tell me I did not say that...
 
That said... I also mention the numerous teams I have helped start. Even with my personal thoughts in my mind not yet being placed here... Just about a little over a month ago I directed a business colleague to several forums to get an idea of what F@H offers. What he took from the "Looking around" was that in order for him and or any team he would start... They needed to spend $4000+ per system (4P) in order to feel like their efforts can have any kind of impact. This was very unattractive to him. He made a point to tell me a few weeks later in a business meeting we both attended.
 
This was unsolicited and caught me off guard. This is how a business man read into the four forums I directed him to. And to be clear... EVGA was not one of them.
 
So push my "Perception" off to the side all you want. And now points have found their way back into this thread so I will respond to that (Very little because it is where threads get locked ) The facts as they are out there today are clear.

1. Bonus points are provided to those who have built "Well" above average systems. Their is no incentive to use what you have because it all adds up as PG has made the "Points" the incentive. Not the project.
2. F@H can still be run on nearly any CPU or GPU. The slower they are, the less PPD they get. 
3. You cannot take four Single CPU computers and get the same CPU PPD from those four as you can from a single 4P. Yet four x58's cast a LOT less than a 4P can. (Deep pocket zinger)
4. Kepler is not optimized. 
5. Communication while improving, is not even close to other "Scientific" projects.
6. Perception can win and lose volunteers. 
7. The majority of folding power comes from "Single" GPU and or CPU systems. Not free HP clouds and 4P's.

 
This does not mean they are just simply producing less WU's per core. It means for a few of those on the outside looking in that it "Looks" like in order to have any positive impact they must build the big expensive machines or have multiple GPU's that in turn cost far more on electricity than CPU's do. Then you add to that, if they chose Keplers, they are running a HACK and wasting energy compared to other projects that are fully utilizing the Keplers and their power...
 
So if I may... This perception, whether right or wrong, is a perception that a reasonable amount of folks have. 
 
Now let's remember this little math lesson Punchy offered us as a food for thought on the "Deep pockets" perception... http://forums.evga.com/fb.ashx?m=1765957
 
For me deep pockets it is very simple. Most of us got going with a "Bring what you have, it is what we are looking for!" kind of energy. Today... You need multiple systems or CPU's or GPU's to even score the exact same amount of PPD we once scored with the original "Bring what you have" way of helping.
 
You do not see me attacking those who have nothing but rainbows and flowering things to say about PG. I am glad folks feel the way they do. But does that mean their opinion/perception is any less or more valid than mine?
 
No...

I bet they adjust the points on it soon.

Afterburner

Raises eyebrow...

Yes spook!!.... me, you, boney and the scotch man will be investigating the phenomon
 
so low has the helm, and O'Hurry is communicating and Check Off is on the "Kan!!!!" 
 

Simba123
But then I don't know what GPU they used as the 'reference' , but unless it's a GT 210 or something, these numbers will be corrected.

Check the FAQ.

Macaholic

Simba123
But then I don't know what GPU they used as the 'reference' , but unless it's a GT 210 or something, these numbers will be corrected.

Check the FAQ.

Just remember that the FAQ has not be edited for the new changes yet, that's the old way of benchmarking. All units are to be benchmarked on one system, the one for SMP I would guess, to determine the base points. I haven't seen anything posted yet (doesn't mean it's not out there somewhere and I just missed it) about the QRB calculation, so I would think it is the same formula that has been used on SMP/BA. If it is the same work being completed that much quicker, there really is no reason to adjust the points scheme. Adjustments in the past were made because there was no direct way to compare the work being done by one type of client versus the other. Even between SMP and BA there isn't a definite direct comparison due to scaling variances and (AFAIK) not doing the same type of work (I don't think the BA units are just larger SMP units).

Oh and...
 
Every GPU in the EVGA Marketplace just jumped in value if this holds true once released from the BETA team...
 
This makes me want to grab (4) additional GPU's. Really looking forward to what CUDA 5 will bring as well. And I wonder if the points will drop once that support releases to the masses as well...

Xavier Zepherious
as for the points... this could change - this is a beta run to test out the QRB for GPU's
give them a baseline so to speak - to see if it needs tweaking

 

I would only argue that a change to the QRB for GPUs should result in the same change being made for QRB on other platforms. If we look at it for face value, equal compensation for equal work performed means that everything should be judged on the same basis. So, maybe a problem with QRB could be realized, but any corrections should be to the one calculation, not developing different calculations for different hardware.

This is hilarious... support... yeah, right.
 

If you do not join the BetaTeam, you can read the discussions but you cannot participate in them. Moreover, NO SUPPORT is given to Non-Beta Team Testers if they encounter issues with Beta Projects or Beta FahCores. If you need help, please make sure that you are NOT running/folding a Beta FahCore or WU. We will ASK YOU TO REMOVE THE BETA CONFIGURATION before helping you troubleshoot your issue.