Jump to content

Home

folding@teamLFN


ChAiNz.2da

Recommended Posts

  • Replies 746
  • Created
  • Last Reply

Updating those who are using the -advmethods switch on their Fermi GPU, it looks like Pande Group has released some new projects looking to take full advantage of the processing power of the Fermi GPU's.

Per Pande Group member yslin:

Please note that P6811 has the largest number of atoms ever (2150 atoms) for a GPU3/Fermi project. Donors with low/medium-end GPUs may notice a significant difference in PPD since low/medium-end GPUs may not perform well relative to high-end GPUs as system size increases.

 

My GTX 460 is crunching a Project 6811 GPU right now and it's worth a whopping 7202 points but my PPD has dropped from around 13K to about 10.5K ppd. Speaking for myself, I'm not concerned about the points drop off but I did notice my GTX 460 is being pushed harder than it was before as my temps are up about 5-6 degrees Celsius. I find this annoying as NVIDIA's System Monitor utility always says my GPU is 99% utilized whenever it's running the folding client. Apparently there are different levels of 99% GPU utilization. :mad:

 

 

@ ChAiNz,

If you're thinking about a GTX 460 then just be aware that rumor is the NVIDIA GTX 560 will be released early next year and will feature 384 stream processors. I knew this was going to come eventually but I ran out of patience. Source: Fudzilla

Link to comment
Share on other sites

A hearty congratulations to all members of Team LFN!!! We have forced Pande Group to express the sum of our collective folding efforts in 8 digits now. That's right party people. Our team has reached the milestone of... ***drum roll please***

[align=center][highlight]10,000,000 points!!!!!!!![/highlight][/align]

 

*** wipes tears of joy from eyes ***

You Team LFNers rock the house like no other. Who wants the first big kiss? :kisses2:

Link to comment
Share on other sites

  • 2 weeks later...

^^^^ Glad that that was expressed!!! Kudos to all our folders@home!

 

I am thinking about getting a second 460, but I do not have a 2x16 sli interface on my mobo, just a 2x8... there are some great deals for Cyber Monday!

 

Welcome to TeamLFN loony636 and JediExile!! Haven't seen you in a while, loony636 :waive1:

 

Negative Sun, you are closing in on your 50 WU's for that nifty badge! Well done, it is great to have members of LF, new and old, be active in our endeavor to fold them proteins!

 

Oh, and in case you haven't noticed Char Ell is folding at a breakneck pace! Someone close to him needs to get him some new case fans and a system blower :lol:

Link to comment
Share on other sites

It's great to have you on our team JediExile!

 

@ Qui-Gon

FWIW, when I installed my GTX 460 I moved my GTS 250 to the 2nd PCI-E slot on my main rig which runs a Core i7 860, meaning that I am limited to 8x for each PCI-E slot now. So far I haven't noticed any real difference in gaming at 1920x1200 resolution, playing games like Mass Effect 2 and Dragon Age: Origins. However I'm only folding on the GTX 460 though because I'm concerned my power supply may be a bit underpowered. I haven't encountered any power problems so far while playing any games but all I'm using the GTS 250 for is PhysX. It's not running any display. I tend to think the GTS 250 will get worked a lot harder and draw more power if I run a folding client on it and if I do that in conjunction with running the SMP client and GPU client on the GTX 460, well, I think that may be too much to ask of my 650W Corsair PSU.

 

If you're concerned about drop in points per day performance however, I'm not sure on that one. I seem to recall reading some posts on foldingforum.org that there was no appreciable PPD performance difference between running a GPU client on a PCI-E 8x slot and a PCI-E 16x slot but it's not something I can personally verify.

Link to comment
Share on other sites

Interesting on the non-difference on the x8 v x16 bandwidth. It was just a thought.

 

It is interesting that we have the same cards, and that I have the i7 870, nearly identical main setups. I had never thought of running the 250 as a second card on my main rig, as I knew I could not SLI the different cards. It never occurred to me that I could use them independently! I also did not know you were folding on the SMP client simultaneously with your Fermi GPU client... I have thought about that, and am interested how difficult the setup is, and how taxing/heat inducing it is on your system. My box stays cool, but I am also a worry wort!!

 

Thanks for the food for thought, Char Ell!

Link to comment
Share on other sites

Well, I can definitely say that I've found version 6.30 of the SMP client very simple to run. It's a console client so it runs in a command window. I understand you can set it up to run as a Windows service if you want but I've not bothered with that. I'm actually running it and the GPU client on my PC as I'm typing this post. I do let my SMP client run on all 8 CPU cores (of course that includes the 4 logical cores).

 

Ambient temp: 69 degrees F

 

Operating temps:

CPU 1 - 124 degrees F @ 100% utilization

CPU 2 - 124 degrees F @ 100% utilization

CPU 3 - 117 degrees F @ 100% utilization

CPU 4 - 117 degrees F @ 100% utilization

I use a Cooler Master Hyper 212 for CPU cooling with just the fan it comes with to push air across the cooling fins.

 

GPU - 128 degrees F @ 100% utilization (cooling fans reported @ 38% of max speed) running project 6811 work unit. So far I've found 6811 WU's work my GPU the hardest and run the hottest.

 

IIRC my PC was pulling around 350 W from the wall socket when I simultaneously ran the SMP client on my Core i7 860 and GPU client on the GTX 460. I measured that shortly after I started running both clients at the same time and before my GPU had started running the project 6811 WU's so I wouldn't be surprised if my rig is pulling an additional 10-20 W now.

 

A couple of other things I think worth mentioning.

 

- From what I've read on foldingforum.org, for optimal performance the SMP client should be configured with the default idle priority and the GPU client should be configured to run at slightly higher priority on systems running both clients.

 

- Some folders have reported problems running two instances of the GPU v3 client on systems with both a Fermi and non-Fermi graphics card. At the present time it appears the best way to deal with the problem is to run the GPU v3 client on the Fermi GPU (GTX 460) and the GPU v2 client on the non-Fermi GPU (GTS 250). I haven't tried this myself. You would still need to include the -gpu 0 and -gpu 1 switches in the configuration for each client.

 

- For those running the SMP client I definitely recommend getting a passkey and adding that to your SMP client config. Here are the requirements for getting the quick return bonus

1. Run the latest SMP client (v6.29 or above).

2. Configure that client with a passkey.

3. Complete 10 a2 and/or a3 work units...

  • a. within their preferred deadlines
  • b. using the same passkey and fah user name combination
  • c. on one or more of your systems.

4. Successfully return >=80% of assigned [bonus] WU's.

source: http://foldingforum.org/viewtopic.php?f=58&t=13160

The key thing with getting the quick return bonus on the SMP client though is that you need to return a WU before its preferred deadline to get the bonus. The earlier you return the completed WU, the higher the bonus.

To illustrate, the project 6701 WU my SMP client is working on right now has a base value of 921 points. According to FahMon 2.3.99.3 if I turned in this WU right now I would get 5,127 points for it. However FahMon projects my current SMP WU won't be completed for another 12 hours. At that point I expect it to be worth somewhere between 2,500-3,500 points.

 

So let me know if you have any more questions Qui-Gon. That goes for anybody else on the team as well. :) As I said, I found running the SMP and GPU client on my main rig to be pretty straightforward. The only funky thing about the console for SMP is I have to use Ctrl+C to gracefully exit the SMP client and even then it doesn't seem very graceful as the client reports it was improperly shut down the next time it starts up. Maybe there is a better way to exit the SMP client but I haven't discovered it as yet.

Link to comment
Share on other sites

Phew. Months pass quickly. Amazing how far LFN has come, almost within the top 500, incredible stuff guys.

 

Once I'm done playing Assassin's Creed Brotherhood, I may let my PS3 do some folding again. Only during weekends, though. Keep up the amazing work!

Link to comment
Share on other sites

Nice to see you Miltiades. We'd love to have your PS3 rejoin the "fold." yeah, that was a really lame line

 

In other folding news, our team just passed the 11 million points milestone. I've wondered how many points could be considered a completed folding simulation run but I'm guessing that's not really feasible to calculate with all the diversity of projects Pande Group has put out.

 

Also, I felt it important to point out that Professor Pande has posted on his blog that the Stanford University network will be undergoing maintenance during the winter break from December 18 through January 2 so there is likely to be intermittent but brief issues with connectivity to folding@home servers. But the biggest outage looks to be this Thursday, December 16. It sounds like a large number of F@H servers will be down while the room they reside in is serviced. Work units may be in short supply that day so heads up.

 

Fold on, Team LFN! :rock:

Link to comment
Share on other sites

  • 2 weeks later...

A big shout out to Qui-Gon Glenn for hitting the [highlight]1 million points[/highlight] milestone and in record time. I don't think anybody else that has done this much folding has done it as quickly. Most impressive. :vader3:

 

As a matter of fact, these young ladies were so impressed that they felt compelled to write words of inspiration on their bodies and present them thus though they were a bit timid and decided to don boxes on their heads. :blush2:

[align=center](This BBCode requires its accompanying plugin to work properly.)[/align]

 

If anybody has thought I've been too "points-centric" over the past while I apologize. My participation in the Folding@home project really stems from a desire to make the world a better place by donating my computing resources to a group that is working to find ways to combat Alzheimer's and other chronic diseases. Points are just a way to measure that contribution. Go Team LFN! :rock:

Link to comment
Share on other sites

Go Team LFN indeed!! I do not accuse you of being points-centric, I simply accuse you of maximizing your hardware and blasting through WUs like a blowtorch!

 

I think congrats are indeed belonging to you, as you have surpassed our beloved ChAiNz' points total to become our new folding leader! And it is not like he has been slouching!

 

I think we all feel the same way, Char Ell, in that we are trying to make a difference in this small way. It is the coolest thing I am involved with, which doesn't say much about my social life, but does present to you that I too am proud to be able to offer my systems to a greater cause.

 

To that end, it is great to see Miltiades drop by the group! Glad that you might put your PS3 to some protein crunching soon :)

 

Also, Char Ell, thank you for your earlier post here about using the SMP client and the GPU3 client simultaneously. I am not doing it yet, but it is something I am going to do. The computer is built like a brick s**thouse, so I might as well make it work :)

Link to comment
Share on other sites

  • 3 weeks later...

Hey Gang - a quick note!

 

gratz to Char Ell and ChAiNz.2da for crossing over the 3 million point mark!! Puts you both near or better than the top 5K participants, which out of 1.5 million, is pretty danged awesome!

 

Kudos to Miltiades! Making that PS3 sweat a little lately, eh?

 

Our team is about to crack the 500 mark in rankings. I have been looking forward to this since I started... it has actually been my folding goal to this point :) I am psyched that we will be accessible from the All-Teams list with one click of the mouse.

 

Since that goal is nearly accomplished, and my personal folding goals were mostly reached when I started, it is time for me to make a new goal. I am going to try and inspire our current team-members to fold again! This will be done by PM.... hopefully I am not a bother, but the contribution to medicine is potentially significant, as I think we all agree!

Link to comment
Share on other sites

Hah, congrats to the big boys! I'm eagerly awaiting our entry in the top 500, it's gonna be awesome.

 

And yes, my PS3 is running from the moment I'm getting out of bed to the moment I'm getting back in it. :) Although to be honest, my desktop was folding pretty much that whole time as well. PS3 WU's are worth only 250 points, which is a shame. And it means my 8800GT did most of the work. But I'm happy, I got that 250,000 milestone.

Link to comment
Share on other sites

A question for those who use their PS3 for folding: what's the performance of your PS3's? Mine's about 0.0712/frame or 242ns/day. I'm not sure what it all means, and if this is related to the WU's, but I'd like something to compare it to.

 

Also, is there any advantage in using the cpu client over the gpu client if I have a Core 2 Duo processor? Or is the 8800GT a better choice for folding anyway?

Link to comment
Share on other sites

Not sure how many CUDA cores the 8800GT has... I think that it will not fold so much, as I had a BFG GT210 (newer card, not as high end as 88xx was at its release, but pretty much equivalent) and it would not fold with GPU client. So, I guess your C2D will do as well or better, especially if you use the SMP client.

 

The first couple months I folder, was all on standard client CPU. You can see my points in the months of June and July of 2010, that I was folding CPU only. That will give you an idea of standard CPU folding for a C2D (6400, 2.13ghz). I think SMP might double that output, possibly more, but Char Ell would know more about that than I :)

Link to comment
Share on other sites

On my PS3 Slim, a project 4045 WU is getting .1494s/frame and 115.49ns/day. I'm not sure how or even if the s/frame and ns/day stats translate into points production. But I've no idea what my PS3's points per day production currently is because as far as I know there is no way of discerning PS3 work units from the GPU and SMP work units my other F@H clients are turning in. I know last winter my PS3 was generally producing about 1,000 points per day but I determined that by shutting down all my other F@H clients for a few days so I knew my stats were only PS3 WU's.

 

My guess is an NVIDIA 8800GT should produce 3,500-4,500 points per day. With 112 stream processors it's no longer the folding PPD king it used to be, that's for sure ;), but it can still make a significant contribution. :nod:

 

A big shout out to ChAiNz for surpassing the 3,000,000 points milestone. You just keep on truckin' bro! :thumbs1:

 

Congratulations to JediExile on surpassing 2,500 points. :thumbsup:

I'm quite happy to see the uniprocessor folders like JediExile, chi7891, and Negative Sun continue to contribute to our team's collective folding efforts. Keep up the great work!

Link to comment
Share on other sites

I'm quite happy to see the uniprocessor folders like JediExile, chi7891, and Negative Sun continue to contribute to our team's collective folding efforts. Keep up the great work!

 

Hear Hear!! Together, we are knocking on the door of #499 in team rankings :D This will happen by late tomorrow, :clap2:

 

Congratulations to all TeamLFN folders, active or "on standby", for reaching a significant Team Milestone!!

Link to comment
Share on other sites

^^^ Yeah, you do have a point there... I guess I follow the extremeoverclocking accounting exclusively now :) I almost forget the client's links to team and personal standings.

 

So, I stand corrected! However, we are now 498 on the EO ranks, and shortly will be in the 480's, which by then surely will have us in the Stanford top 500.

 

Either way, I feel like wearing a party hat!

 

-------------------------------------------

 

How do any of you folders, when talking to people about folding for the first time, how do you tell them about folding? How do you "explain" what is going on? I am no scientist, and I barely grasp what it is we are contributing... and most of my friends and family look at me as "that guy" :p So at times, I struggle mightily to explain what it is, and I would like to do so better, as many people I talk to have perfectly good computers laying around doing nothing while powered on 24/7/365. These people should all be folding, on the Uniprocessor client at least, and more robust clients as their hardware allows. As we see in our team, uniprocessor client results count just the same as any other WU! And any contribution is potentially a significant one :)

 

Just curious what ya'll thought about that... or if it even comes up in your RL discussions?!?

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.


×
×
  • Create New...