Folding@home's End User License Agreement (EULA) and why it's there

The admins have wisely deleted that thread as it had no place in this forum for many reasons.
 
Already posted that in NVIDIA preliminary numbers thread. This post was in response to my PM to Vijay about the multi-gpu mod and It's also me who asked Lethal to delete it.

There is a reason Stanford forbid anyone to host the client files outside their own servers and the multi-gpu support is also a security risk,

 
Doesn't seem to be stopping a certain team that's had a tremendous ramp the past week or two. although after poking around there, it doesn't appear they're modifying any code. It seems to be simply exploiting an option that's already in the client. It appears they have several users running two and possibly in at least one case, three cards per PC.

Cheesy, dangerous for the science...yes... violating the EULA... not so sure. Certainly seems like it could put the project at risk IMO though. It could either hasten or delay the release of multi-gpu support.

0.png
 
Yes, I know some people will take advantage but leave the ethics issues to them. We are the [H]orde and we abide by the rules, even if that put us in a precarious situation.

Trust Stanford and we will be rewarded sooner than you think ;)

 
I have no doubts we'll get our shot at the multi-gpu thing as soon as practical, and had no intention of bending the rules either way.... I wouldn't have even started folding if the only impetus was how many points I could get. IMO the points are just a way to measure your progress and forward motion. Although they do inspire healthy competition and that's a good thing..:cool:

Holy shit...I just quoted Martha Stewart...:rolleyes: somebody get a gun.. :D

Thanks for keeping on top of this issue..;)

It'd be kinda ironic if Stanford got pissed and pulled the Nvidia support (not that they would) as a certain individual at a certain un-named team would be sitting on an ass load of kinda useless video cards....:p

0.png
 
Ahh..., who's this Martha Steward and who really cares who the frigg she is anyway:confused:? Is she the one violating the Eula :confused:? and who the heck is this guy/girl Eula anyway (a street walker?) (only kidding, hey nomad8u, don't sweat the small stuff, even the crookedest people say good things once in a while :D)

FOLD ON :p

 
Doesn't seem to be stopping a certain team that's had a tremendous ramp the past week or two. although after poking around there, it doesn't appear they're modifying any code. It seems to be simply exploiting an option that's already in the client. It appears they have several users running two and possibly in at least one case, three cards per PC...
Click to expand...
Which team are you referring to specifically (IIMA)? Both OCAU and the Russians in the top 5 have seen very nice ramps. In the case of OCAU it's the seasonal change that permitted higher OC to their machines, and their effort is an honorable one. Russia, OTOH, is a bit of a mystery to me... :confused:

EVGA has also seen a massive ramp of late... :eek:

 
Guys - if there is a specific allegation, then take it to your team leadership quietly, decide on your team's position, then take it to the relevant people (if it is against us, possibly us first in order to discuss in a friendly manner and try to come to a resolution, after which we would be happy for you to take it to Stanford if you still had misgivings).

Please do not post insinuations about either us, or any other team in your forums. It isn't good for inter-team relations, especially since our two teams are on pretty friendly terms with each other.... not as if we ban each other's accounts on the forums etc etc....in fact, a number or us are in here on a daily basis, and some of your members are in our forums every day too... reconnaissance missions!!

We have posted how we are doing it - all with nVidia and Stanford files and as such we have nothing to hide. A number have already seen it I should imagine. Some of you may have even followed our instructions. Good on you if you have - it's more science done.

If one particular person wishes to exploit the current points incentives by purchasing that many video cards, good on them. Yes it is kind of insane but that person is insane (and in a good way)....they bought a heap of Q6600's when they were still in B3 stepping and well before any price cuts. I am sure that many of you bought Quads after the price cut last year in order to exploit the two WinSMPs + Affinity Changer points bonanza.

The GPU's do more science than a Quad does anyway, if you can get three of them in a boxen and reduce the costs of dedicated machines and get bigger outputs, then you should do it if you can. It has to be more power efficient I am sure.

In the large scheme of things, it's not so much the overclocking in winter that is helping us. My Q6600 G0's run at 3.6Ghz year-round, and many of us have the same SOP now (and yes it gets bloody hot at myplace in summer!!). The ramp is simply down to one thing - more clients running, more people running them.

Many people on our forums who were not folders before, yet are active in the GPU overclocking areas have joined as they see it as something easy to run (and it is) while making a difference to our team, the science and humanity. They see it as something that their shiny new $800 GPU can do while they are at work.

Fold on.
 
Well said Misti :) We are not here to point fingers at anyone, only to get rid of questionnable methods and I'm sure team OCAU or any other team would consider the quality of science more important than raw output. Just don't let some rotten apples trash the whole basket :)

Fold on !

 
I don't think there was any questionable methods in the first place. Obviously recompiling clients is a bit hard since we do not have the source code.... and I have better things to do with my time than disassemble the client executables...

At the moment, we are in a state of testing previously run proteins in order to test that the vectors are the same or within acceptable limits of the previously crunched proteins. If people never pushed the boundaries, we would never have seen the affinity changer.

If running this sort of thing hurts the results and Stanford knows who is causing this by whatever means (trying to cheat, modifying clients, or simply running two video cards) then their account should be locked, and measures taken to stop it happening. I'd prefer them to try and resolve the situation and unlock the account too.
 
Mitsimonsta said:
Guys - if there is a specific allegation, then take it to your team leadership quietly, decide on your team's position, then take it to the relevant people (if it is against us, possibly us first in order to discuss in a friendly manner and try to come to a resolution, after which we would be happy for you to take it to Stanford if you still had misgivings).

Please do not post insinuations about either us, or any other team in your forums. It isn't good for inter-team relations, especially since our two teams are on pretty friendly terms with each other.... not as if we ban each other's accounts on the forums etc etc....in fact, a number or us are in here on a daily basis, and some of your members are in our forums every day too... reconnaissance missions!!

We have posted how we are doing it - all with nVidia and Stanford files and as such we have nothing to hide. A number have already seen it I should imagine. Some of you may have even followed our instructions. Good on you if you have - it's more science done.

If one particular person wishes to exploit the current points incentives by purchasing that many video cards, good on them. Yes it is kind of insane but that person is insane (and in a good way)....they bought a heap of Q6600's when they were still in B3 stepping and well before any price cuts. I am sure that many of you bought Quads after the price cut last year in order to exploit the two WinSMPs + Affinity Changer points bonanza.

The GPU's do more science than a Quad does anyway, if you can get three of them in a boxen and reduce the costs of dedicated machines and get bigger outputs, then you should do it if you can. It has to be more power efficient I am sure.

In the large scheme of things, it's not so much the overclocking in winter that is helping us. My Q6600 G0's run at 3.6Ghz year-round, and many of us have the same SOP now (and yes it gets bloody hot at myplace in summer!!). The ramp is simply down to one thing - more clients running, more people running them.

Many people on our forums who were not folders before, yet are active in the GPU overclocking areas have joined as they see it as something easy to run (and it is) while making a difference to our team, the science and humanity. They see it as something that their shiny new $800 GPU can do while they are at work.

Fold on.
Click to expand...

Mitsi... if this is concerning my post in this thread, I in no way meant to insinuate your team was cheating, and I don't want there to be any hard feelings between the teams.... that's unnecessary and would be unfortunate. If that's the way you took it, I apologize, and your issue is not with the team so much as with me.. let's keep it that way..;)

That being said... this is a discussion forum, and I was replying to a topic. If you'll notice, I did say I didn't think this issue (the multi-GPU client thing) was a violation of the EULA but merely an exploit that exists in the client due to it being based on the multi GPU capable ATI code. I stand by that as well... In either case, it should be no cause for you to get pissed at the team here. Nothing wrong's been done right? So there's no real issue...

Yeah I visit your forum on occasion, and like here have learned alot about how to improve and tweak my production to benefit the cause. Your forum has been a valuable resource as has the info here and at several other forums I visit.. I haven't posted at your forum yet, but will in the future (if I'm not banned after this...:p).

I agree that your "trials" with multi-gpu support have been openly discussed on your forums and I don't see why they wouldn't be... and I also agree that the sick cat with all the video cards is one wacky dude... and yes in a good way when you get down to who this benefits...

In any case, I stopped back in here to check on what's happening and saw your posts, and don't want you or OCAU being pissed @ the team. I don't think anything I posted was earth shattering news, and hell I didn't even mention your team but since you take issue with it, I guess I don't have to.

I stand by any comments I made and if you have an issue with them, it's with me and not the [H]orde...

Peace Brother and Fold On!

0.png
 
Its ok we know over at OCAU that the [H] are just upset. We will have #1 soon enough, and you guys can take comfort in knowing that your secure in the #2 spot. :)

BTW multiple GPU clients is fine by Stanford, just like running two instances of SMP ;)
 
The[H]ardenator;1032688673 said:
Its ok we know over at OCAU that the [H] are just upset. We will have #1 soon enough, and you guys can take comfort in knowing that your secure in the #2 spot. :)
Click to expand...

You guys have put a serious run on no doubt.... congrats! It looks like you've had alot of new folders join up recently... good stuff all the way around.
But ya can't count us out just yet. :cool:

BTW multiple GPU clients is fine by Stanford, just like running two instances of SMP ;)
Click to expand...

Touche.. nicely played...:p:D

0.png
 
The[H]ardenator;1032688673 said:
BTW multiple GPU clients is fine by Stanford, just like running two instances of SMP ;)
Click to expand...

Not what I read:

There's been some recent discussion of aspects of Folding@home's EULA and what constitutes a violation of the EULA. One key issue is whether donors (end users of the FAH software) can make modifications to the code in order to make enhancements, such as multi-gpu support for NVIDIA. While such modifications may be made with the very best of intentions, these modifications do not go through our QA (either at Stanford or our relevant partners such as NVIDIA, ATI, or Sony). This is a big problem with 3rd party client modifications is that even very subtle modifications can create problems for the science involved. Since there's no way for donors to QA end user modifications, we cannot support their use and must remind everyone that any modifications to any of the client binaries is a true EULA violation.
Click to expand...

 
Which part of that specifically says no multiple gpu clients?

I dunno if you guys realise, but there is no modifications needed. The -gpu switch works in the 6.12 beta 8 client straight from stanford with no mods at all.

Its as simple as installing 2 copies and changing machineid and using -gpu switch. No different to running multiple cpu clients.
 
chestRcopRpot2 said:
Which part of that specifically says no multiple gpu clients?

I dunno if you guys realise, but there is no modifications needed. The -gpu switch works in the 6.12 beta 8 client straight from stanford with no mods at all.

Its as simple as installing 2 copies and changing machineid and using -gpu switch. No different to running multiple cpu clients.
Click to expand...

The GPU- switch only works with the ATI client at the moment.
I've seen no word from Stanford that the Nvidia client been fixed to support multi-clients yet.
Hence the slight mod to a couple of the files which lets you run multi-clients on nVidia cards.
If you could run multi-clients without the file mods then I'd say go for it.
But with the client so new, the file mod may just possibly break something else.
Hence Stanford's reluctance to condone it.

Just my 0.02 worth.

Luck ......... :D
 
Tigerbiten said:
The GPU- switch only works with the ATI client at the moment.
Click to expand...

Tiger, if this is the case, why is it that you can get it working with only a client.cfg edit?

I might be totally stupid, but since when is a client switch & modifying the client.cfg supposed to be modifying the client itself? There is no editing, reverse engineering or decompilation of any of the binaries. If there was, I'd be right there with you, pointing the finger. In fact, I would be very angry about it.

The fact is that the quote that Kendrak posted does not apply, nobody is using a special patcher, nobody is hacking the binaries. It is simply doing what we have done since the P4 HT days - a second client on available hardware.

The client doesn't even come with a client.cfg, the client generates one itself on first run. I routinely modified my client.cfg in the old V5 console client days.

Mind you, I am not rich enough to run multiple cards in one machine, so it's all a bit meh to me.

As I posted earlier, if you seriously believe that there is an EULA breach by OCAU or any other team, then I suggest you discuss it privately (as in away from somewhere we can see it, PM maybe?) with your team leadership. As I also posted earlier, you could throw it past OCAU leadership first in order to try and come to an agreement before annoying Vijay (or other Pandegroup member), it might be a good idea.

Heck, both teams could go together and ask for a determination from Vijay.
 
Tigerbiten said:
The GPU- switch only works with the ATI client at the moment.
Click to expand...

This is what i was referring to before. It actually does work with nvidia cards too! No silly eula violating dlls or mods required. Although there is still caveats such as needing displays attached to each card in vista, sli needing to be disabled etc. Hopefully stanford are ironing out the bigger bugs as we speak. :)
 
The beta 8 client is the same for both cards.. it installs both the AMD/ATi CAL dlls and Nvidia's CUDA dll.

I might piss off some of you but, I'm with the OCAU guys on this one. Modding the client.cfg and adding switches isn't cheating or violating the EULA.. at least not anymore then running a SMP client in a VM or 2 clients on a quad. Something that most, if not all, the top producers on this team are doing. Vijay has said he doesn't want that to happen either. Yet they do it and encourage others to do it too, all for the sake of increasing the (point) production of each machine. If it was only about the science, they'd want the WUs returned as fast as possible (Vijay). Knowing that the SMP client scales pretty crappy, and that after 2 cores the returns drop like a rock, most seem fit to limit the clients to 2 cores and use the other 2 for another instance. Even though it slows down the return time. "But its in by the deadline!!" They proclaim.. as if that makes going against what Stanford wants ok.... :(

Sorry if my rant was a bit much for some of you.. well ok not really... But after the way some of my fellow teammates are treating the people running dual gpu.. acting like they take the high road and are better. When really, they are nothing more then hypocrites



 
I already discussed directly with Vijay on this. On a side, it's a "benign" violation by his words since it's not really in the client itself but with a driver dll to enable a hidden feature. On the other side, since there is no QA testing done with 2 GPU2 clients running together, we don't know if the results is valid or just garbage (I know that right now, we are just crunching test units with no real scientific worth but if the test results is garbage, we are delaying updates because some people decide to run outside the beta testing boundaries). It's also pointless to try to make this mod official because Stanford and NVIDIA already have something in QA testing.

If you want to benefit from the official one, I suggest you submit to join the beta testing group on FCF and with some luck, you will be able to use the official one to gain lots of points ;) If you care about doing quality science, I suggest you stay within the boundaris (for SMP, it's a different game because no modification is required).

I'm a programmer myself and I care a lot about doing quality beta testing and introducting unknown variables in the equation make things more complicated. Debugging may prove to be useless when we find out it was due to that unknown variable.

 
You must log in or register to reply here.
Back
Top