Suggested Change to the PPD System

Moderators: Site Moderators, FAHC Science Team

Grandpa_01
Posts: 1122
Joined: Wed Mar 04, 2009 7:36 am
Hardware configuration: 3 - Supermicro H8QGi-F AMD MC 6174=144 cores 2.5Ghz, 96GB G.Skill DDR3 1333Mhz Ubuntu 10.10
2 - Asus P6X58D-E i7 980X 4.4Ghz 6GB DDR3 2000 A-Data 64GB SSD Ubuntu 10.10
1 - Asus Rampage Gene III 17 970 4.3Ghz DDR3 2000 2-500GB Segate 7200.11 0-Raid Ubuntu 10.10
1 - Asus G73JH Laptop i7 740QM 1.86Ghz ATI 5870M

Re: Suggested Change to the PPD System

Post by Grandpa_01 »

7im wrote:Maybe Grampa_01 would like to comment on this points inflation as compared to his recent rant about low PPD on P8101s. :twisted:

7im wrote:

Grandpa_01 wrote:1. What is the problem ? Some People feel cheated
2. How big is the problem ? Very small a very small minority and very vocal group of folders comment 1 way or another.
3. Is there in fact really a problem at all ? .NO the current system works well.
4. What are the all the ways of coming at possible causes of symptoms? What factors are actually causing the symptoms? Are we defining the right problem?(does the "problem" we have decided to address actually cause the symptoms we are addressing) The answer to the 4th part of this question is yes The rest of this I feel are irrelevant since they are all answered by #1
5. Who is best informed to define the problem, and best able to measure the resolution? VJ / Stanford He knows what he needs and how to achieve it. He also knows he cannot win by answering any questions because he is going to make somebody mad. It is sad but us few donors put him and Stanford in a loose loose situation because of our selfish comments and actions and really do not give a rip about it or the progress of the science.

The numbers say the amount of science getting done goes up every year some say this is true some say it is false. I believe it is true and therefore see no problem. I do not mind looking at a bunch of zeros in a # but others might. I also do not believe in a handycap system never have never will. I believe if X is doing X them give him X plain and simple. Most of these discussions are nothing but political views and really have no place in science. :wink:




Not your thread. You don't get to make up whatever answers fit your only your personal viewpoint.



Not your thread either but yet you can answer with whatever fits your own personal view?

Now is that fair? :wink:
Hmmph I remember someone asking me to comment earlier in this thread :lol:

And yes that is my personal view just as your answers to the question were your personal view. And I really do not care whose thread it is if I have something to say about it. And you know who the people are I was referring to. One of them is myself. You can go back through the threads like this both here and other forums and it is always the same people starting these threads and voicing there opinions over and over again. And it is only a few people doing it. It appears that the majority of users do not see a problem, but there are a few people who are always stirring the pot. And no k1wi I am not referring to you your thread is about your idea of a solution to a perceived problem which arises from time to time which is the (H) factor.

I have seen the threads where some of these people have asked other members of there forums to come here to the FF and post comments about there dissatisfaction with the points and QRB system and at the most 1 or 2 of there members do so or none. It just is not that big of a deal in the folding community as a whole. It is no where near what some would like us to think it is.

As far as k1wi suggestion goes as I said before I do not know whether it is a good thing or a bad thing. I know there are some things I am concerned about such as how is a person that is just starting going to catch a person that has 500,000,000 points folded under the current system in a reasonable amount of time. ? Are you going to 0 everybody out, the system does not really give a person that ability as I see it. I believe the current system has been in effect to long to change it and still remain fair to new folders or even folders that have been at it for a while. The current system actually works quite well in enticing people to upgrade and invest in more powerful equipment. I do not understand why anybody would want to stop that it would not be in the interest of the science.

Some say that people would upgrade no matter what and that may be true to a certain extent but if you dangle a little carrot out there you will not get the same results as dangling a big carrot would. :wink:
Image
2 - SM H8QGi-F AMD 6xxx=112 cores @ 3.2 & 3.9Ghz
5 - SM X9QRI-f+ Intel 4650 = 320 cores @ 3.15Ghz
2 - I7 980X 4.4Ghz 2-GTX680
1 - 2700k 4.4Ghz GTX680
Total = 464 cores folding
k1wi
Posts: 909
Joined: Tue Sep 22, 2009 10:48 pm

Re: Suggested Change to the PPD System

Post by k1wi »

If we set Yn at 1, then it would have the effect of maintaining the current points distribution. I'm not advocating devaluing the points system to a previous time frame (although I suspect reverting to the point in time where QRB was introduced would have been ideal, just not realistic)... I'm certainly not going to zero everyone out either, I can't think of any purpose in touching users total point balances.
Patriot
Posts: 76
Joined: Sat Aug 16, 2008 2:04 pm

Re: Suggested Change to the PPD System

Post by Patriot »

I was trying to figure out where that quote was coming from...then I noticed a hidden post...heh
but he is right....6903 is 4000x bigger not 1000x...
anyhow...

So I have heard various talk about how the gpu units were (are) smaller and less complex.
I knew that GPU client was Single precision and had mistakenly thought CPU was Dual precision.
Logic connecting less precise wu with client type...
Anyhow Bruce shed some light onto the clients...
bruce wrote:AFAIK, Uniprocessor and SMP are single precision WUs, too. The fundamental difference between CPU-based processing and GPU based processing is a limitation to implicit solvent calculations. The GROMACS used for CPU-based calculations can process either an implicit or explicit solvent.
bruce wrote:An explicit solvent simulates a protein molecule immersed in a large field of water (usually) atoms. Each atom can move and when it does, it changes the forces on nearby atoms. This also means the water molecule can move, including rotating and since water molecules are polarized, the forces on the protein change, depending on what the water does. There have to be lots of H-atoms and O-atoms to represent a uniform solvent surrounding the protein molecule.

An implicit solvent model has no water molecules but rather is simulated by equations for a force-field that comes close to being equivalent to all of the water. Those equations are much quicker to evaluate than computing each H-atom and each O-atom but they're plenty good for an approximation and obviously MUCH, MUCH faster.

FahCore_11 (and whatever runs on the PS3) only had the implicit solvent equations and the hardware at that time couldn't really handle enough atoms for an explicit solvent. Projects 764x and 802x seem to have an unusually large number of atoms -- and a lot of folks are complaining that these projects don't behave "normally" which MIGHT suggest that they're explicit solvent models. They're operating on new versions of FahCores and only on relatively new GPUs so there's good reason to speculate that explicit solvent models may soon be commonplace on GPUs . . . or not.

EDIT:
Found an interesting statement in a recent abstract.
http://pubs.acs.org/doi/abs/10.1021/ja200310d

Note the words "non-bulk-like" which, to me means that the result can only be detected in an explicit solvent model.
Image
Grandpa_01
Posts: 1122
Joined: Wed Mar 04, 2009 7:36 am
Hardware configuration: 3 - Supermicro H8QGi-F AMD MC 6174=144 cores 2.5Ghz, 96GB G.Skill DDR3 1333Mhz Ubuntu 10.10
2 - Asus P6X58D-E i7 980X 4.4Ghz 6GB DDR3 2000 A-Data 64GB SSD Ubuntu 10.10
1 - Asus Rampage Gene III 17 970 4.3Ghz DDR3 2000 2-500GB Segate 7200.11 0-Raid Ubuntu 10.10
1 - Asus G73JH Laptop i7 740QM 1.86Ghz ATI 5870M

Re: Suggested Change to the PPD System

Post by Grandpa_01 »

If it would work in reality and not just in theory then I see no problem with it, as long as the proverbial carrot remains the same as today's. Which I believe is what you are trying to accomplish. But as I stated above I do not mind looking a bunch of 0's. I do not know if it will work or not though, my mind is not that complex when it comes to formulas. And thanks k1we you have done a good job of answering my concerns. (At least to the best of my mind's abilities)

I am sure I will have more concerns in the future since I do not feel I completely understand the formulas or how it is going to work.
Image
2 - SM H8QGi-F AMD 6xxx=112 cores @ 3.2 & 3.9Ghz
5 - SM X9QRI-f+ Intel 4650 = 320 cores @ 3.15Ghz
2 - I7 980X 4.4Ghz 2-GTX680
1 - 2700k 4.4Ghz GTX680
Total = 464 cores folding
jimerickson
Posts: 533
Joined: Tue May 27, 2008 11:56 pm
Hardware configuration: Parts:
Asus H370 Mining Master motherboard (X2)
Patriot Viper DDR4 memory 16gb stick (X4)
Nvidia GeForce GTX 1080 gpu (X16)
Intel Core i7 8700 cpu (X2)
Silverstone 1000 watt psu (X4)
Veddha 8 gpu miner case (X2)
Thermaltake hsf (X2)
Ubit riser card (X16)
Location: ames, iowa

Re: Suggested Change to the PPD System

Post by jimerickson »

@Patriot: thanks for that info! makes me rethink what has been said about gpu qrb.
Post Reply