New Nvidia core?

[toTOW]

Who knows ? ... you might have some good surprises after all ...

[codysluder]

VJ Pande & the rest of the group needs to take a serious look at how WCG is run from top to bottom and consider making some changes. Until then, this project will continue to lose contributors because you are basically telling them that their contribution means nothing. I know because I have slowly regressed in my F@H status because of all of the frustrations & now my equipment is going towards a cause that is very well documented to be doing exactly what they say it is.

That's a pretty biased statement.

The Pande Group has said that a day's work on a GPU is worth 1500 points. That's not nothing. A days' work on the CPU is worth 110 points.

The only problem here is that you've been getting an inflated number of points and you figure that you're entitled to continue to profit from the bugs in the earlier versions of the core. As these bugs are corrected and more realistic (i.e.-larger) proteins arrive, almost anything can happen, but lower points have been predicted from day one.

Perhaps you'd be happier if the development code had remained hidden until a final version is ready to release.

[Indignity]

That's a pretty biased statement.

Cody, how are my comments biased? It's been voiced time & again that people are upset about the point distribution . Case in point FLECOM (I do not know him individually nor have I had conversation with him). For the most part, he was given the cold shoulder until it started looking like bad "PR".

The Pande Group has said that a day's work on a GPU is worth 1500 points. That's not nothing. A days' work on the CPU is worth 110 points.

The only problem here is that you've been getting an inflated number of points and you figure that you're entitled to continue to profit from the bugs in the earlier versions of the core. As these bugs are corrected and more realistic (i.e.-larger) proteins arrive, almost anything can happen, but lower points have been predicted from day one.

Point taken & that's been discussed/argued many a time. I've even been one to stand up & defend the process. Here's the problem; When we were all cruising along on 5000PPD with our 8800s, the response was soo great that it took out the project's servers constantly. The majority of us were fine with these occasional outages because we were still able to get WUs to do while the completed ones sat. With these particular WUs, there weren't the constant EUEs & turning our silicone to mush from the heat generated by these units when I could keep them working.

Along the same lines, what happened to these WUs? Did they go away because the research wasn't needed any longer?

Perhaps you'd be happier if the development code had remained hidden until a final version is ready to release.

That's what BETA is. We have been forced to update cores on more than one occasion.

**Puts on Tinfoil Cap**

Since it's been voiced that Stanford has absolutely no ownership in this whatsoever, I've been very skeptical that this "Project" is even doing what we all believe it was supposed to be doing. "Canned" answers via PM to people inquiring about proof just further puts that doubt in my head.

Take a look at my contribution & ask me if I was dedicated to F@H. I personally feel that I have been deceived and my personal investment in time & equipment was all for nothing!

Mr. Pande, I hope you are getting everything you want out of this project and will eventually become very wealthy!

[7im]

Not sure what you mean by no ownership. Vijay posts here all the time. Very few DC project leads actually post in their project's own forum.

The previous smaller WUs didn't go away, they are just becoming larger and more complex as Stanford's ability to more accurately model proteins improves. This is just like the CPU client and WUs did before this. Actually, read the Project News page. Pande Group just published a new paper based on work done by the GPU client. I haven't checked recently, but 60+ published papers is likely more Papers than all the other DC projects combined. That's hardly nothing!!!

And it's rather difficult to become very wealthy when they share the results and are published openly. Please check your facts before slamming the project with unsupported accusations.


News Post about new Paper published with help by GPU client contribution: http://folding.typepad.com/news/2008/12 ... ished.html

[WickedPixie]

Actually, read the Project News page. Pande Group just published a new paper based on work done by the GPU client. I haven't checked recently, but 60+ published papers is likely more Papers than all the other DC projects combined. That's hardly nothing!!!

News Post about new Paper published with help by GPU client contribution: http://folding.typepad.com/news/2008/12 ... ished.html

Am I reading it right that the published paper was based on classic clients (Tinker & Gromacs?), and follow-up simulations using GPU2 & SMP clients will follow?

Will there be an announcement when the follow-up simulations start, especially on the SMP side?

The reason I ask is so that I can return all the machines I pulled out from SMP, and a good portion of my GPUs.
To my understanding, the bulk of these SMP projects are geared toward Influenza Virus research.

If there is a direct correlation between viral infections and AD/PD, kindly explain it in layman's terms.
If there is, I would gladly return them all to F@H.

[Indignity]

7im, with the utmost respect, this is what I am referring to:

After a phone call to Stanford’s Provost and then being referred to one “Richard Zare” I am frankly a bit stunned. The reason for this call was to see if we could get some assistance with our F@H problems.

Within the chain of command at Stanford Professor Pande would be answerable to Professor Zare.

After a very polite phone call I learned a few things everyone should know and consider.

Pande is only responsible to Zare as far as “Ethics” are concerned.

Pande is totally autonomous within the structure of Stanford. Basically, he answers to know one.

All decisions regarding the Folding at home project are the sole responsibility of Pande.

Stanford University takes no responsibility for Pande and or his project, save he do something actually illegal for which they may or might become libel.

How Pande runs his program is of NO INTEREST to Professor Zare or Stanford University.

If the program should fail the entire responsibility will fall on Pande, not Professor Zare and not Stanford University.

Should the entire program fail do to mismanagement, it again falls on Pande, Stanford doesn’t care, and it is not their responsibility.

Professor Zare informed me, “I do know over one million people are involved”. My next question was, “what if they all quit?” The response from Zare was clear “then they quit”. I asked, “Wouldn’t this be a possible publicity nightmare for Stanford” the response “this is Professors Pande’s project and if it fails it fails I think you will find all universities are run this way”

I then commented that Pande had in essence told one of our members that was contributing over 800 computers to “go pound sand”. Professor Zare’s response to that was “exactly, it is Professor Pande’s project and he has that right”.

Keep this in mind; my question about “what if they all quit” was a spontaneous response from me born from my own incredulity because frankly I was a bit shocked. Professor Zare in no way took the question as a threat, he started laughing with me. In fact Professor Zare was a very polite gentleman and chuckled all through this entire conversation and thought at first I was calling about VJ’s forum.

Simply put, Stanford University as per Professor Richard Zare simply doesn’t care if the project works or not.

In the end while Professor Pande uses Stanford University’s name this project is VJ Pande’s project, not Stanford University’s project. This project from day one has belonged to VJ Pande.

This information was shared by a member of a community that is one of the larger contributors to the entire F@H project. I am often skeptical of these types of shared information, but this gentleman was sincere & only had good intentions as a reason to pursue the information he shared.

7im, I have read the information posted about F@H. I have both asked for & given help with the project at various times. I was a big proponent of the project for a very long time. People like me don't have the inside information that you & the Pande team do, therefore it's easier to be skeptical of what the project represents.

The bottom line is if I didn't care about what the ultimate result of F@H is supposed to represent I would not be as passionate about expressing my feelings as I am. I am also certain that I speak for many many frustrated people that have not taken the time to share their feelings.

[bruce]

Any university expects it's professors to do research. It's called "academic freedom" for a reason. Decisions about science should be made by scientists. I'm sure he will find exactly the same sort of response from the Provost of any other university if you ask about a specific professor or a specific research project. The idea of calling the Provost's office or Prof. Zare demonstrates a fundamental misunderstanding of the nature of academic freedom. In a corporate environment, such officials might have a supervisory role for the content of the science, but such is not the case in most US universities. The concept of academic freedom is that administrative officials have supervisory roles for finances, personnel issues, and legal or ethical issues. Academic research is completely in the hands of professors (of course the overall quality is evaluated and certified at the department, school, and university level through mechanisms such as the tenure process but they do not control the research). Thus, not only do the chair of Chemistry, Dean, and Provost approve of Folding@Home, but they would feel it inappropriate to interfere in Dr. Pande's work barring a serious ethical or legal problem. The paradigm of "may I speak to your supervisor" simply does not apply at any university.

I'm not sure what your objective is. The title of this topic is "New NVidia core?" and your post has nothing to do with a new nVidia core so it should be deleted as "off topic" I can split this new discussion off into a new topic but it's not clear what i should call it or which posts will make a coherent discussion of some other topic. (If you have a suggestion, PM me.)

The next posts had better be on-topic or they'll be removed.

[EvilAlchemist]

Seriously, I don't understand what happens to the "old" WUs & why the hell the cores are forever in constant change!!!!

The old WU's were completed. You Folded / Unfolded them. That project is done. Time for a new one set of work units.
It would not do any good to fold the same work units over and over and over again ... no science would be done.

The cores change because there are improvments / changes that need to be made.
Sometimes the change is to update the code for better scientific results.
Sometimes the change is to optimize the core and give better PPD.
Sometimes the change is to fix something that is wrong / broken in the code.

[VijayPande]

**Puts on Tinfoil Cap**

Since it's been voiced that Stanford has absolutely no ownership in this whatsoever, I've been very skeptical that this "Project" is even doing what we all believe it was supposed to be doing. "Canned" answers via PM to people inquiring about proof just further puts that doubt in my head.

Take a look at my contribution & ask me if I was dedicated to F@H. I personally feel that I have been deceived and my personal investment in time & equipment was all for nothing!

Mr. Pande, I hope you are getting everything you want out of this project and will eventually become very wealthy!

The above post is simply wrong. It's important to keep in mind what are the facts and what is rumor. In the end, all of this comes purely from a discussion about my not raising the points for the classic client. Here are the facts:

- All of FAH is owned by Stanford. Stanford owns the software, the copyright, the Folding@home trademark, everything. I (and my team) have no ownership stake in any part of it and see no profits from the calculations of FAH donors. This is discussed in detail in the FAQ.

- Stanford's Office of the Associate Dean of Science directly supports FAH financially, paying for our server room, networking charges, etc. He is very aware of FAH and I report to him in my capacity of Chair of Biophysics and in terms of FAH issues that need to be resolved.

- To date, FAH has resulted in great achievements, both in terms of quantity (see our numerous papers http://folding.stanford.edu/English/Paper) and quality (numerous awards from the scientific community http://folding.stanford.edu/English/Awards). This has all been open and public.

I hope that clears up the facts. I think that rumor originated from a phone call of a folder to Stanford's offices then misinterpreted the answers. Stanford has these sorts of details publicly listed on various web sites. For example, the Research Policy Handbook (http://rph.stanford.edu/) discusses many of these details (and this document may be interesting in general to those who want to know how research institutions work).

[mklvotep]

I shouldn't have worried, these new wu's are really nice. I'm probably in the mid 90k range from 70k. Very nice.

[spazzychalk]

i guess i weighed in on this a while back on the cant we forget about the damn points and focus on the cause platform, but now for a more technical review. this is a beta. besides the note that the higher points were bugs, its still in testing. youre testing new software. and even throughout final releases codes are optimized to speed up then computations which gets your points faster anyway through revision histories

[John Naylor]

And as far as points go, this warning is in bold on the download page so while the core is still beta I would think it still applies:

Beta clients and servers performance may vary significantly from standard FAH clients during the development process, including but not limited to work unit shortages, server downtime for upgrades, short notice for client upgrades, and Points Per Day that differs a little or a lot from the developmental benchmark level.