Bad News for us SMP Folders

[aldamon]

You will also find bonus point clients to suck you in to more and faster hardware only to find if you choose to remain competitive you will buy even more hardware and on and on it goes. Worse VJ and crew treat any and all information about project changes like it was papers in an eyes only ultra top secret folding lying on the president’s desk.

Perhaps we could get more information if we filed under the freedom of information act.

Hehe. This killed me.

If F@H doesn't show me they can reverse their Cranio-rectal impaction, I'm probably going to go the way of Aldamon and many others to WCG.

I'm glad you guys still tolerate me after I jumped ship.

 

[Sunin]

Why your still folding for a good cause just under what appears to be a more organized system. If you quit all together well then that could change

 

[Xilikon]

BillR, I will play devil's advocate and comment about hiring professional programmers. To make the nvidia GPU2 client ready for release, Vijay contacted a programmer from nvidia (I made the suggestion in the effect that we have Mike Houston representing AMD so we need someone to represent NVIDIA) and Scott Le Grand came and he is the mastermind behind the CUDA engine. This show that they were able to hire good ones even if it's paid by manufacturers.

On the GPU2 side, things is going well and you will be surprised at how good it will run and I can bet it will be out of beta fast enough since it's so easy to install and to use as the console client for massive production. OTOH, on the SMP client, there is no really great programmer specialized into parallelized computing (under windows in addition) so being open source (the MPI engine and the GROMACS engine), all we can hope is to see someone step on the plate and do something...

Anyway, beside this note, I do agree with BillR that the Stanford team should be better with the PR side.

 

[jws2346]

BillR you are one smart dude and I'm doing my best to learn from fellow folding people such as yourself, people like relic (which I have great respect for, as I do you) and people that have been in this "very serious game" called F@H much longer than myself.

I want to wholeheartedly appologize if I have sounded sarcastic or "eat up" with any kind of cynicism. The only excuse I can come up with is I'm still in the learning stage of the etiquette in the "Web Community" (yeah, I know, excuses are like ♥♥♥♥♥♥♥♥s, everybodys got one )

I'll admit I have made some real unconsious idiotic mistakes in the past, but the best I can do is try not to repeat the same mistakes. I have always adhered to the old expression "the moving finger writes and after having written, no tears or words of piety can change what is written" (or something like that ) What that means to me, short of God (or whatever you believe in), there's no way to go back and change the past, all you can do is try to do better in the future. (pay attention all you politicians and "would be" politicians out there )

With all that said, I'm having a ball. One of the first and hardest lessons I learned from members such as yourself, no matter what, do not over leverage (go into unmanageable debt) yourself for the sake of buying better hardware to fold with. Also to appreciate and sometimes be amazed at the hardware some people have, but always and I mean always, try to be respectful.

It's always good to be on the same team with intelligent and helpful veterans of the campaign (I think it's called "teamwork" or something like that )





.

 

[marty9876]

So much complaining about such the same issues over such a long time from the same names quoted over and over.

Not saying anything stated is wrong, just either fuck the dead horse already or BBQ his ass!

 

[alan2308]

I'm glad you guys still tolerate me after I jumped ship.

Well, last I checked this was the Distributed Computing section, not the Folding@Home section. Officially supported or not, there's no reason why F@H, WCG, SIMAP, DIMES or anything else can't coexist here.

If the WCG team gets stronger, maybe we can crack the top 100 on the DC Vault.

 

[alan2308]

So much complaining about such the same issues over such a long time from the same names quoted over and over.

Not saying anything stated is wrong, just either fuck the dead horse already or BBQ his ass!

Tough call there, not sure either option there would be any more or less tolerable than the other.

 

[jws2346]

I'm glad you guys still tolerate me after I jumped ship.

Please correct me if I'm wrong, but I thought the only way you could "jump ship" is if you stopped using some kind of "Distributed Computing" program altogether.

The way I look at it whether you do WCG, SETI, F@H or, and I'm not familiar with them, SIMAP or DIMES it's all about "freedom of choice". I was maybe mistakenly under the impression "freedom of choice" was what the "Democratic environment" was all about

Please correct me anyone if I'm under a false impression. If I was cynical, which I'm not, I could say that expression "freedom of choice" sounds good, but hardly applicable in todays politics.

 

[APOLLO]

If the WCG team gets stronger, maybe we can crack the top 100 on the DC Vault.

IMO, that should have happened a long time ago. It might be a good thing some members are tackling other projects with their resources. [H] deserves far better than 111th. For perspective, my province is ranked 63rd.

Please correct me if I'm wrong, but I thought the only way you could "jump ship" is if you stopped using some kind of "Distributed Computing" program altogether.

I think he meant in the sense that F@H is the most supported DC project around here and AFAIK the only one with official recognition, although I could be wrong about that.

 

[Xilikon]

I think he meant in the sense that F@H is the most supported DC project around here and AFAIK the only one with official recognition, although I could be wrong about that.

You're correct, F@H is the only one officially endorsed by Kyle. Others is accepted but in a unofficial way. I have pushed for the recognition of WCG as another official team but Kyle rejected it to avoid any new e-drama...

 

[SmokeRngs]

You're correct, F@H is the only one officially endorsed by Kyle. Others is accepted but in a unofficial way. I have pushed for the recognition of WCG as another official team but Kyle rejected it to avoid any new e-drama...

Trust me, you're not the only one that pushed for it. I was tempted to switch the vast majority of my machines over to WCG if it was going to be recognized as an official team. As it is, I have a few single core machines, a couple borgs and my overclocked X2 running it anyway.

 

[jws2346]

Hey thanks a bunch Xilikon and APOLLO I was really scratching my head there for a minute.

Being a little new to DC'ing I didn't quite understand the "jumping ship" thing, as far as I knew aldamon was still part of the [H]orde forum, but just doing a different DC type program (other than F@H)

I use F@H mostly because it's the first time I've tackled anything DC connected. It seems the more knowledge you get the more knowledgeable choices you can make, of course, bias-ness (is that a word?) is always present, just like with most everything else.

I just figured aldamon was DC'ing with another type of DC program, the way the "jumping ship" deal sounded to me was he'd left the whole [H]orde forum (intelligence challenged )

Edit: While I might have close to 1 and 3/4 million pernts I don't come close to having the experience, time or effort it takes to get over 20,000 WU's like some of our more experienced members have accumulated. (hell, when you come right down to it, I don't even have a 1000 WU's)

 

[alan2308]

IMO, that should have happened a long time ago. It might be a good thing some members are tackling other projects with their resources. [H] deserves far better than 111th. For perspective, my province is ranked 63rd.

You should have seen where we were before I got all our teams tracked. I don't remember what the actual rank was, but I'm pretty sure it was sub-1000. Half of whats on our team page wasn't there, and the wrong WCG team was tracked.

Heres what APOLLO and I were talking about for anyone who doesn't know:

http://www.dc-vault.com/showteam.php?team=98

 

[Mohonri]

I think Vijay and the whole F@H team are in an interesting situation. For a long time, all computers just had a single core processor and a video card that was useless for folding. Now, in the space of only a few years, they not only have dual-core processors becoming the norm, but quad-core CPUs are far from rare. In addition, in the space of a year, both big GPU design companies now have immensely powerful GPUs capable of high-performance folding calculations. And in the same time frame, the PS3 is released.

In other words, there are four (SMP, nV, ATi, PS3) very new, very promising platforms that have emerged in a very short time. Vijay et al now find themselves like kids in a candy store--there's so much to do that it's hard to focus. SMP was first, so they worked really hard on that. But then the GPU clients became possible. And then the PS3. And then GPU2. And now nVidia. With each step, they have a huge potential gain in performance, so it makes sense to shift focus to the "new and hot" technology. As a result, the older ones get left behind.

Just be patient. It'll settle down soon enough.

 

[Xilikon]

I think Vijay and the whole F@H team are in an interesting situation. For a long time, all computers just had a single core processor and a video card that was useless for folding. Now, in the space of only a few years, they not only have dual-core processors becoming the norm, but quad-core CPUs are far from rare. In addition, in the space of a year, both big GPU design companies now have immensely powerful GPUs capable of high-performance folding calculations. And in the same time frame, the PS3 is released.

In other words, there are four (SMP, nV, ATi, PS3) very new, very promising platforms that have emerged in a very short time. Vijay et al now find themselves like kids in a candy store--there's so much to do that it's hard to focus. SMP was first, so they worked really hard on that. But then the GPU clients became possible. And then the PS3. And then GPU2. And now nVidia. With each step, they have a huge potential gain in performance, so it makes sense to shift focus to the "new and hot" technology. As a result, the older ones get left behind.

Just be patient. It'll settle down soon enough.

That's exactly my opinion and to add to this, Vijay said there will be a points rebalance down the road to put a correct value to the science generated by the various clients. It's also why it's now a good idea to diversify and not rely solely on SMP to have a stable points output.

Look at my points output, it's very stable despite the incoming 2665 and that's due to some variety. Now that I started folding on my 8800GTS, I'm currently planning some changes within the next month to have 2-3 GPU folding.

 

[EvilAlchemist]

With the GPU client putting out 4-6k PPD ... look like us SMP folders are not the big producers anymore.

I think, that if the GPU2 stick around, you will see all the SMP folders back their CPU OC back down to stock and throw in video cards ....

And if Stanford can get the GPU2 client to feed from more then one core, this will be a great setup

 

[schenks]

With the GPU client putting out 4-6k PPD ... look like us SMP folders are not the big producers anymore.

I think, that if the GPU2 stick around, you will see all the SMP folders back their CPU OC back down to stock and throw in video cards ....

And if Stanford can get the GPU2 client to feed from more then one core, this will be a great setup

I don't know, I'm actually thinking about INCREASING my OC. By just running 1x SMP, 1x standard, 1x GPU, the GPU only uses about 20% of one core. The resulting heat drop gained me about 9C TEMP DROP on my quad @ 3.1, giving me more room to increase my overclock and reap further benefits.

I had been running at 60/60/54/54, now i'm running at 48/51/45/45, which is just ridiculous considering I gained about 3500 PPD on this machine.

 

[Xilikon]

Evil, I told ya yesterday when we talked about the potential boost from GPU2. I think they are in a interesting position to be able to do some changes without angering a big SMP crowd because GPU2 is there as a different kind of candy

To add some weight, Adam Beberg himself said that the ultimate goal is to have SMP+GPU2 coexist on a dual core

 

[EvilAlchemist]

Evil, I told ya yesterday when we talked about the potential boost from GPU2. I think they are in a interesting position to be able to do some changes without angering a big SMP crowd because GPU2 is there as a different kind of candy

To add some weight, Adam Beberg himself said that the ultimate goal is to have SMP+GPU2 coexist on a dual core

I know .. All I said was in a good way ....
Just looking to use my boxen to their best.

 

[SmokeRngs]

While I'm not sure we've seen many examples with the nVidia GPU client, I know the older GPU clients had trouble being fed if you didn't have a rather fast CPU to feed it. I doubt this will cause many if any to low their overclocks, I know I won't. Then again, I use the machine for other things as it is and every extra mhz counts with some of the software.

 

[Xilikon]

SmokeRngs, it's the AMD cards which is having issues with the CPU feeding them. OTOH, NVIDIA have issue with releasing cycles when you need it.

Both clients are at two extremes and the goal is to be on the middle (minimal CPU usage and releasing GPU cycles when you need it for gaming or video).

 

[APOLLO]

SmokeRngs, it's the AMD cards which is having issues with the CPU feeding them. OTOH, NVIDIA have issue with releasing cycles when you need it.

Do you know if this is also true of the 4000-series? These cards are starting to look like performance bangs for the buck when it comes to gaming according to early released benches and rumors. It would be great to have an awesome performing card in both folding and games. The 4850 will probably destroy an 8800GT in games.

 

[Xilikon]

Do you know if this is also true of the 4000-series? These cards are starting to look like performance bangs for the buck when it comes to gaming according to early released benches and rumors. It would be great to have an awesome performing card in both folding and games. The 4850 will probably destroy an 8800GT in games.

It's the same with the 4000 series. However, Mike Houston said he have some tricks up his sleeve to fix this so watch for some news shortly (NVIDIA seems to have kicked him in the butt to hurry and tweak ).