I'd like to share that I've been holding off on buying a kepler card specifically since it's not been optimized yet for F@H.  I really want to upgrade since I'm a gamer and I want more powah!  :(  
 
I wonder how many others are in the same shoes as me?
 

Afterburner

P4 systems = deep pockets. The select few, not the many.

P4 not 4P, as in Pentium 4
 
I'm talking way way before EVGA every got into folding. Back when the Socket A was a brand new thing.
 
AMD started selling chips that you could get a 33-50% OC on for a fraction of what Intel wanted for a P4 rig.
 
So we shifted. Not a couple fast systems, but boxen upon boxen of slightly slower AMD rigs that cost 1/3 of what the Intel ones did. People where literately putting them in pizza boxes. From this the term boxen came about.
 
We had another guy that used thumb tack and "papered" the walls of his study with systems. Cops got called on him because the windows was opened in the middle of winter and they thought he was a meth lab.
 
EVGA started up about the time that we where changing from an almost all GPU focused team to an SMP/-bigadv team because of the release of the i7 920.  I know I sold of my 3x 9800GX2 righ and a few months later was one of the first to be playing with a dual 1366 rig (I think i was the second one on the team with one).
 
The point is types of hardware fall into and out of favor. PG focuses on different things at different times.
Just have to put the sail up where the wind is blowing. The EVGA members that are building 4P rigs are doing just that. Just don't get too comfy....  Always something on the horizon.
 
Fold on!

The only question is what is the scientific value of the WU's the 2 different rigs produce. Is the scientific value of machine x 8 time greater than machine y. I do believe Stanford sets that mark by what they believe and use it as a guide for the available path's to choose from.

Grandpa_01

The only question is what is the scientific value of the WU's the 2 different rigs produce. Is the scientific value of machine x 8 time greater than machine y. I do believe Stanford sets that mark by what they believe and use it as a guide for the available path's to choose from.

Good question. One would think out of all of us here, that one of us would know the answer to that very question.
 
Making a point here. I did not, have not, complained about the points in this thread because I feel it is totally a waste of time as I tried to be a part of that conversation last year. All that does is muddy the waters. Like Politicians do when they want folks to follow the noise or pretty lights instead seeing the actual issue or problem at hand.
 
To me the value is in "Production". Can we agree on that? I Crunch two computers 24/7 and frankly do not care about points. (Link) Each one hits me in one or multiple ways. So the points have zero value to me. Fun to talk about them? Yes of course. But you do not see me stop WCG that I have been doing for many years now because I can use Prime Grid for HUGE amounts of more points than WCG will ever give me.
 
Or even POEM over eOn2. It is about my systems being used at their full potential on projects that...

1. I feel my productions has true value to the project
2. The projects are keeping up with the CUDA Cores
3. Projects help, respond and or at least keep us in the loop as to what is going on and is in the "Real" future.

This is all I am voicing my thoughts on. Not one of those three items is happening that I am aware of at this time in F@H. Now if you asked me to use my electricity and parts, that you should not keep me up to date? Just how long would folks "stay" in the program exactly?
 
Like we both agree on... Progress happens. That is a two way street. 
 
Now instead of me just having a hissy fit and screaming I quit. I am voicing my thoughts to gauge what is my next step. I am looking for improvement. Otherwise... My point remains the same... They are losing the largest audience and fast. That can only be covered up by bonus points for so long. Someone has to look at the WU's processing on time and accurately at some point.
 
Or is that the point in all this. They no longer need those that got them where they are today? They are to big/successful for any time to be spent on us now...

edit: retract my post.
No one could have answered the questions.
Because no one knows, not even PG knows them selves.

My post was trying to stop the points argument it does not belong in this thread, this thread is about the fundamental problems of F@H with the lack of support and communication with the folding public. Both from the FF and Stanford itself which I believe we all agree could use some fixing. I crunch from time time as well but not near as often as I use to. I also know that there support is no better than Stanford's is when it comes to keeping the public informed (at least it was not in the past). They have just as many server issues as Stanford if not more and there are allot of Boinc projects that I have absolutely no interest in but you do have a choice of what you want to crunch.
 
As far as Production goes I completely agree with that and I will let Stanford guide me with there values as to what I will use to fold with. The deep pockets comments people make just irritate me to no end. I have no deep pockets I still work a 40+ hour week in construction people act like the points is why everybody fold's and quite frankly that is a thoughtless comment, When I sit next to my wife who is a 4 year cancer survivor and listen to her struggle to breathe with her 1 lung and look at the oxygen line hooked to the machine that motivates me to fold. When I think of my mother that died at the age of 52 from stomach cancer that motivates me to fold. When I think of my grandfather who passed of Alzheimer's at age of 67 that motivates me to fold. When I think of my 4 children and 9 grandchildren and possibly finding a cure or treatment to battle these diseases before it affects them, that motivates me to fold. I really get tired of the thoughtless comments made by people. Perhaps they should think about what would motivate an average person like myself to spend so much time and money on something like this.  

1. I feel my productions has true value to the project
2. The projects are keeping up with the CUDA Cores
3. Projects help, respond and or at least keep us in the loop as to what is going on and is in the "Real" future.

This is all I am voicing my thoughts on. Not one of those three items is happening that I am aware of at this time in F@H. Now if you asked me to use my electricity and parts, that you should not keep me up to date? Just how long would folks "stay" in the program exactly?

1. All WU's have a value just some are needed more urgently than others.
 
2. I am pretty sure they are trying VJ has said before and not too long ago that he believed the future of folding was in the Graphics card. But from as long as it is taking them I am guessing that Kepler and cuda are having a hard time working with Gromacs and the larger and more complex WU's that Stanford need's
 
3. This one needs some major work done on it and I do know that VJ is concerned about it from conversations I have had with him. I definitely would not want to be in his shoes here because when he does post it usually makes 1 group or another mad. If they said a new working core was going to be out in a month then did not deliver I am sure there would be some pretty big rants going on. But they do need to communicate better with the folding public. And the FF needs to be a little more friendly to the donors.

Like we both agree on... Progress happens. That is a two way street. 

Now instead of me just having a hissy fit and screaming I quit. I am voicing my thoughts to gauge what is my next step. I am looking for improvement. Otherwise... My point remains the same... They are losing the largest audience and fast. That can only be covered up by bonus points for so long. Someone has to look at the WU's processing on time and accurately at some point.

Or is that the point in all this. They no longer need those that got them where they are today? They are to big/successful for any time to be spent on us now...

Discussion and venting is a good thing, I would question whether it is the largest audience or not though not every computer has a folding capable graphics card but every graphics card needs a computer, but that does not matter any loss of any folder in any group of folders is a loss period. I do not know the answer or the remedy for the problem, I do know that any given group of folders is going to want an advantage over the next group that is the human factor which comes into play because the point system creates a competition thing and due to that fact I very seriously doubt that Stanford will ever be able to win in this area. It is pretty much a loose loose situation for them here. What I hope though is that most folders will realize that the true winners of our folding is the future of humanity and the e-peen thing will take a back seat in most of our minds.

Just wanted to pop back in and say, I like that this thread is comprised of a variety of Folders across different teams.   This is the way we should operate all the time.  Work together, think out-loud together and above all be one for the project.  Kudos to how people have been able to share ideas and keep the thread functional.  I don't personally believe that we will affect anything (well maybe we will affect each other's thinking a bit), but sometimes it is therapeutic to just share what we are thinking too.
 
Liked that brief history of [H] Kendrak, learned some things I didn't know about you guys.  Grandpa, my wife is COPD, on full-time oxygen with her mother and grandmother lost to Alzheimers.  Taking care of her is the most important thing I do and I appreciated your explanation of why you Fold.  I too think of my Daughter, Grandkids and others when I consider why I do this work (Folding and Crunching).  Any time I get too far into the frustrations of computing DC, I only need to think about them to get re-grounded.  Wish I could say I was always perfect at that, but I'm not. 

My comment goes along with Grandpa's. Ultimately, it's about the science and also about what you are willing to sacrifice to build a multiple processor rig. Wanting support for dedicated rigs for bigadv is legit. Complaining about the speed at which other equipment gets optimized and supported is kind of petty and obviously about points, as again, folding is supposed to be secondary for a typical rig. Again, my Kepler is doing fine, and I'm not exactly an expert. There will always be hiccups. We just deal with it and move on.

I've pretty much given up on folding
F@H basically looks like a project that's going nowhere

I certainly understand the frustration and the points of view expressed here, but I think I may be able to contribute something regarding the nature of academic research. 
 

• First and foremost, I am not defending PG or F@H, simply trying to shed some light on the subject based on my experience.  Academia is very different than the "real world."

 

• Second, PG is not neglecting Kepler by choice and at this point they are probably as distraught about the fact that they are having difficulty getting it up and running as we are.

 

• Third,  as much as we like to sit here and ask questions, Dr. Pande does not have all the answers.  PG stands for the Pande Group - essentially a loose collaboration of academics led by Dr. Pande.  What this means is that somewhere at a university (more than likely not Stanford) there is someone working on the Kepler core, and its not Dr. Pande.  And its not the same person working on the BigAdv SMP, GPU2, Unicore, ect WUs.  Chances are, they meet face to face at most a couple of times a year at conferences.  As much as we want to turn to Dr. Pande for answers, he more than likely has already asked those same questions and not gotten any meaningful information to pass along.  Unless its significant progress, information tends to flow slowly in academia.

 

• Its impossible to know how close to being finished something is.  You can estimate all you want, but its never done until its done.  Any estimate you provide will not be accurate.  Something always goes wrong, something you hadn't anticipated always comes up.  There are no cutting corners or "eh - good enough" on a project of this nature as it would lead to erroneous and meaningless research.  Would it be preferable to have a date followed by a constant stream of setbacks and missed deadlines?  I'm inclined to think that may just get the general masses more riled up.

 

• Lets also keep in mind that this research is essentially done voluntarily by those performing it.  Though there are loose correlations between research output and pay raises, due to the nature of the tenure-based system in academia incentives to complete this work by most of the individuals are primarily the personal accomplishment and intellectual prestige.  Perhaps a reason we are seeing more emphasis on bigadv is because at the moment a lot of Dr. Kasson's research is BigAdv WUs and he is a recent graduate that is in the pre-tenure phase of his career and tenure decisions are highly influenced by research output in your first 7 years after graduating - providing a strong incentive for him to work extremely hard on his areas. 

 

• Having a bit of experience academia (about to finish grad school), academics at a T1 research institutions performing this type of research typically have huge egos.  They are also often less than socially adept (not at all a slight to them - its just a fact that when you devote that much of time to your work/research to perform at that level there are sacrifices required and socialization is typically is minimal to non-existent).  That makes them inclined to be less than forthcoming when they are having difficulty solving a problem - particularly to those that don't have a good grasp of the intricate and complex of the problem.  When you are performing research on this level, there are only a handful of people that can provide meaningful feedback/suggestions to your work.  Guess what?  No one here is one of them.  I don't mean this to be insulting, as smart as we all are in our particular areas of expertise (and there are a lot of smart people here), we're not experts in the issues they are facing.

 
I know none of this helps us get any closer to releasing a properly working Kepler core, but in reality there is nothing any of us can do.  Is it frustrating?  Yes, I am as frustrated or more than most around here.  But at the same time, I am not at all upset over it, I am not going to stop folding, and I am not going to waste my time worrying about it.  Short of getting a PhD in Biophysics (if anyone wants to sponsor me through school, I'd be inclined to go), there is nothing I can do but fold what I've got and hope for the best.  Just remember why we're all here - for 10 EVGA Bucks a month! 

Good post, thank you Troy. Like I said I am not upset. I feel deflated. All of what you posted is basically what many of us already said. Where did I say the DR. needs to fix anything? I have been saying F@H or PG...
 
My three challenges have not changes thus far. They are valid. 
 
Not one person should stop doing anything because I voiced my thoughts and am looking to see what others think. Is is very helpful to read of those looking past these issues and continuing the good fight. That is what I want to do. Just having a hard time overcoming them at this time... I want to help. Heck how many people on the planet, as individuals, have a combined 100,000,000 + in contribution numbers? 1000? 2000? Out of millions!
 
Keep the thoughts coming folks... This is definitely helpful for a lot of us!

troy8d

Wasn't directed at you, or anyone in particular.  Just my thoughts on the peculiar and sometimes frustrating nature of academia in regard to general sentiments expressed by many on the forum over the last year or so.  At this point, I'm in the same boat as you being disappointed.  I prefer to focus on what I can do right now rather than what I can't.

Gotcha... Still, your post has a lot of value 

watching this discuss and wanting to chime in many times
 
the only issues I see:
 
1. FF mods and folding forum: two issues
a) censorship
 
As many have noted  - mods censoring valid criticism and ideas
if you have issues with points and even provide ways to change it to make it more fair you still get deleted as a gripe
 
even when I posted a idea how to change the client (for core counts) - it felt like my head was bit off for suggesting it
 
if you want input and want criticism and ideas then you have to foster it by letting the complaints or ideas come in and stay. even if you think it's unwarranted, unfounded you let the statement stand. you may challenge but keep the discussion friendly. That's how things get fixed or become aware of the outstanding issues 
 
Id like to see how many programs would get fixed if no one griped or wrote in about the problems because no one listened or they get censored or get their head bit off.
 
just imagine windows or Apple (with all its errors) that no error/information was sent back(crashes/bugs) or gripes about the client the look or feel ...how far it would have come without any user input
 
I don't care how you solve this - but the best way is involving the donors on what the new guidelines for posting and new guidelines for he Mod's to follow would be in order. Censorship is unbecoming , specially from a university that prides itself on that.
 
b) beta forum unwillingness to talk
 
the beta forums aren't working well- as the team has noted they talk less for fear of leaking info...so there is less posting
 
 
2. points fairness
 
this is an issue that has to be solved by stanford. 
Yes I believe some Units are worth more than others
 
yes I believe that a computer should make about the same points for the same amount of work 
 
ie bigadv down to smp
140k to 35k  not a fair step considering you haven't changed the hardware
 
that being said I wouldn't have minded 50-70k
or make WU's that fit in the 70-100k range for SMP
for the point discrepancy between the two
 
but to come out with bigger Work units that earn less points?
 
Yes I believe 2p or 4p can earn more  points (because of the investment on the systems), just not the current sky high numbers. 
I don't mind that they earn more for a particular Project either
 
change some SMP wu's values or create ones that provide balance between the types of systems, so user with i7 don't get bumped down. or move Bigadv more inline with SMP (although you still get better points)
 
Yes I realize those people at the top will curse when you mention points changes - but how could that be any different than everyone else that's had to take a hit.
 
which is why I suggested padding SMP wu's
 
honestly I just wish they went to a more linear QRB than what they have now
 
 
3, Management of FAH
 
FAH has to start operating as a biz
it grown so big they have to operate this way
(you can even sell out your services to others to help support the university)
 
Vijay can be CEO (like bill gates and do development while having final say)
 
just get a COO
which could handle projects(getting work from other institutes or clients) and ongoing points development & fairness, the forums, PR, and all the other stuff with operating a biz like finances
(all the stuff Vijay may have no interest in)
 
he can also make sure the system stays up, updated and working (stay up on the operations of the facility)
leaving Vijay with more scientific endeavors
 
and a CIO
which would develop the collection and distribution system and keeping it all running while bug fixing and software development
 
this is strictly the system development,maintenance and working with Vijay on the clients with others software developers. which might mean a development staff
 
 
Yes Troy8d is right about the academia part.
 
I've seen businesses run out of an personal office at some Universities. doing millions in contracts with govt and millions with businesses.
your collecting money.. doing research for yourself, gov, and clients for that money. in return the govt gets something the client gets something and so do you ..your research and $$$
 
if you running it small time...what happens is you get someone juggling things rather than treating it as such a BIG BIZ. You bring $$$ to the university...treat it as a biz and it will grow.
 
 
-----------
 
with the nature of scholar work he does  and the amount of press and everything else - it's time they started to think of it as a biz for the school. and act that way.
 
if Dr. Pande wants to do research then hire a CEO to operate it all and he can go back to what he did best
 
what is needed is more officers and people to manage it all
 
--------------------------------------
as far as kelper core- well it works - I don't say well
 
the only issue I see here is that stanford has to work with Intel and AMD and Nvidia more closely to add functionality at an earlier stage if possible. I don't believe SMP takes advantage of the full power from the CPU's yet like AVX or AVX2 or FMA2 or FMA3
 
And whats this going to be like when we hit Maxwell(GPU) and IVY-E (8,10,12 core CPU) Haswell -E(14-16? core CPU)
 
faster clocks, more cores, more abilities
GPU autonomous (maybe with a Arm CPU)
 
lots of things to add functionality to
 
And are the points going to grow to 2M PPD or 10M PPD for a system?
keep growing the points rather than capping(like 300k for best 4p system) and fitting everything in-between fairly for all
-------------------------
 
Now Me - I'll fold anyways.  even with the kick in the lower regions. Just don't keep doing it and expect people to stay.
 
Fix the issues outstanding and everyone will stay and come back
or continue ignoring/censoring the donors and people will get dissatisfied with the operations and Policy of FAH and leave
 

From a developer standpoint (I'm not a programmer, but I know many people who are), I feel that PG should consider opensourcing some of this. Maybe crowdsourcing their code would allow more people to chip in. Is nvidia giving their developers all the proper information to develop a core? What is WCG doing that's making it so much better compared to folding? How is their product developed and could fah adopt such a type of system? Why are people so happy to hand over cpu time, and not offer development assistance?
 
As Troy pointed out, this IS rocket science. There can be no grey area, or errors in the precision of information. This is why the work unit is trashed if your gpu gets too hot, despite it being stable at 70c.
 
You can see my post is filled with many questions. Questions likely not to be answered. Despite our vehemence towards PG for not keeping up with things, you have to hand it to them because they're out there doing something meaningful. This is compared to SETI which appears more hellbent on burning as much power as we can produce.

Wow, miss a few days and look what happens...