Folding@home 3850 AGP

[GFreeman]

Hii all! I've downloaded the folding@home GPU2 client for ATI to do some folding. In my family I have some folks with cancer and I figured I want to donate my spare processing power to folding@home. I am pretty new when it comes to folding. It took me about 4 hours I think to complete a WU. I'm folding at a stable 864/2200 with the 3850 AGP. Temps stay at 82c with 99% GPU usage all the time..

http://img62.imageshack.us/img62/5793/folding.png

Today I've completed 2 WU's and gotten 511 points: http://fah-web.stanford.edu/cgi-bin/main.py?qtype=userpage&username=gfreeman

Is that about normal for my folding performance? Also I was wondering what team you TG guys fold in. I'd like to contribute some points ofcourse :) At the moment I'm still in team 0.

[TheCrimsonStar]

what the heck is this exactly? I'm lost...lol

[marfig]

Is that about normal for my folding performance?

For that card, I think so.

A few months ago I was folding with GTX 560 Ti at a 1200 score per WU. But that depends on the project. Each WU would take me roughly 4 hours, IIRC, for a daily total of up to 8k points per day. I suppose your card should indeed put you more or less where you are.

With the summer I stopped and will only resume when the cold season comes back. Your card is an old one, I gather. Should have been through a lot and some of its lifetime has already been expended. I suggest you do the same and wait for cooler days. Those 82c will drop a few degrees and you won't tax it so much.

@Crimson
Check http://en.wikipedia.org/wiki/Folding@home and http://folding.stanford.edu/

[GFreeman]

Thanks for the advice. Well my room temps are fine. I've placed an extra fan so the airflow inside the case is better. I'm giving folding a go, we'll see. I'm not too worried about wearing out the parts. If so then it's time for something new ;) I've had this PC for quite some time really. Anyways I've got my CPU and GPU folding now. CPU priority set to idle with 95% usage on each core. I did that so the GPU folding keeps going quick as well.

http://img806.imageshack.us/img806/8135/fold.jpg

[Kougar]

Hey GFreeman! I fold for EVGA, so your ATI card might revolt if you try to join them... :D In all seriousness, lots of sites have dedicated folding teams... but if you want suggestions, if you already browse w7forums they can always use extra folders. :)

For that card, I think so.

I have to underscore this. AMD GPUs just do not fold as quickly as NVIDIA cards, but the 3000 series wasn't particularly fast to begin with. For that card the performance is fine, but you will be burning a lot of power there for what it can do.

I only bring this up because if you plan to make this into a dedicated folding box it's not going to deliver much for all the power it eats. If electricity rates aren't a major problem then that's fine and ignore what I said. I'm always happy to see new folders. :)

You can use http://fahmon.net/ to monitor your clients, it will calculate PPD and give you info at a glance for both CPU and GPU clients. Only so you have a point of comparison, my GPU can deliver ~15,000PPD at its current speed... granted its a high end card, but NVIDIA cards tend to fold twice as fast as some AMD cards. I'm not aware of any NVIDIA AGP cards that can fold though, the ~8000 series support was being deprecated last I heard.

[GFreeman]

My goodness.. that's exactly what my findings were the other day.. I've tried using this older Dell 2,8GHz HT Prescott rig with a PCI-E slot with the nVidia GT 420 that was in my parents PC. They now have a ATI 6850 so my lil bro can do some gaming once in a while..

This GT 420 doesn't touch the gaming power of my ATI 3850 card.. I've reseated the GPU cooler and put some Artic silver on.. The PCB on the GT 420 is so small I couldn't put better cooling on it. But better cooling paste helped the temps. I've also been fiddling with the fan speed using MSI Afterburner and then some overclocking ofcourse. Temp stays at a solid 57c.

Long story short.. My ATI 3850 puts out 1704 PPD, while this nVidia GT 420 puts out 3081 PPD. It's almost twice as fast folding. CUDA must be some smart technology nVidia developed.

http://img69.imageshack.us/img69/9561/foldingq.jpg

[marfig]

Long story short.. My ATI 3850 puts out 1704 PPD, while this nVidia GT 420 puts out 3081 PPD. It's almost twice as fast folding. CUDA must be some smart technology nVidia developed.


It certainly is. But OpenCL isn't behind. What you are noticing instead is CUDA optimizations in the F@H client, whereas there's no such thing for OpenCL code. So, CUDA runs better because the client developers highly optimized the code and made better use of the technology. As opposed to OpenCL which they haven't.

That's thankfully changing. OpenCL is finally being addressed by the developers. See here (http://www.hardwarecanucks.com/forum/hardware-canucks-reviews/42471-f-h-core-16-taking-amd-gpu-folding-next-level.html).

[GFreeman]

Thanks awfully for that post! I've downloaded that new GPU client and am folding with that now ;) We'll see how that one does.

[Kougar]

That is exactly why I bring it up. I never want to be a killjoy... but seriously, for the energy costs involved with the OC'd 3850, you'd be better off with that low end NVIDIA card. As you found it it delivers almost twice the performance, yet I'm sure it uses much less power at the same time! The same goes for Pentium 4's...

The issues over ATI's sub-par folding has been a long and hotly debated one. I'm going to avoid it entirely, but I will say at least one part of the problem trickles down to the actual hardware silicon... NVIDIA devoted more of their GPU core real estate to general purpose computations than AMD did... this was why the NVIDIA 400 series dies were so much larger than AMD's, yet AMD's were pretty darn close in game performance with their smaller cores.

As Marfig's link points out, the Brook coding was another. Thanks for that link Marfig! I hadn't seen that yet... I can't wait to see what differences that brings. Any sort of performance increase would be very welcome for AMD/ATI users.

[GFreeman]

I agree with you, for folding I'm better off with the low end nVidia card. I mean I've read it's using around 50watts in 3d mode! I have it overclocked which shows a nice performance gain folding. Overclocked it possibly loads at 65 watts or so I reckon. The load wattage of the 3850 isn't too bad if you look in the reviews, but I run a rather big overclock on it. With the GPU folding my PSU reads about 200 watts load. With both CPU and GPU it reads about 240-250 watts of load.

[GFreeman]

It makes you wonder.. Would you benefit more from a low wattage system with low/average performance or a high performance rig which eats some more juice? For 24/7 folding I guess the load wattage system. Well when a low end card beats a midrange card in watts and performance the choice obviously is easy made..

[GFreeman]

That is exactly why I bring it up. I never want to be a killjoy... but seriously, for the energy costs involved with the OC'd 3850, you'd be better off with that low end NVIDIA card. As you found it it delivers almost twice the performance, yet I'm sure it uses much less power at the same time! The same goes for Pentium 4's...

The issues over ATI's sub-par folding has been a long and hotly debated one. I'm going to avoid it entirely, but I will say at least one part of the problem trickles down to the actual hardware silicon... NVIDIA devoted more of their GPU core real estate to general purpose computations than AMD did... this was why the NVIDIA 400 series dies were so much larger than AMD's, yet AMD's were pretty darn close in game performance with their smaller cores.

As Marfig's link points out, the Brook coding was another. Thanks for that link Marfig! I hadn't seen that yet... I can't wait to see what differences that brings. Any sort of performance increase would be very welcome for AMD/ATI users.

For a load wattage rig a Dothan chip with the converter and with DDR2 and PCI-E and the GT 420.. Switching to the lowest multi and letting the GPU fold that would consume almost nothing right.. ;)

[GFreeman]

How about this for a folding machine. I was thinking of the low end core i3 with the GT 425M. It's really low power and get some folding done okay I guess. Okay it's little and it's not your average powerhouse.. but for consistent folding :D It has some other cool features as well ofcourse.

http://www.asrock.com/nettop/overview.asp?Model=Vision%203D%20Series

[GFreeman]

I have just bought it ;) That little thing is going to be my folding dedicated system :)

I've found some power consumption figures... http://hothardware.com/Reviews/ASRock-Vision-3D-137B-HTPC-with-NVIDIA-3DTV-Play/?page=6

[marfig]

I find really awesome that you are putting this much effort into the folding@home project.

I think the machine is excellent for your requirements. Power consumption is awesome. The graphics card is listed as consuming around 35W. So it's possible you can reduce power consumption a bit further by not making use of things you don't need, like Wifi (assuming you don't need it) and audio.

[GFreeman]

Thank you! I will use it for movies and music as well :) It's a nice quiet system and it has a blueray player. I've seen the bios in the reviews, and it's quite diverse with settings. I also have the possibility to play with ratio's, multi's voltages and such. So that could be a nice way to reduce the load wattage even further as well ;)

Secretly I like a quiet system as well haha! My AMD rig isn't that quiet, but I will keep using it for a little gaming.

[GFreeman]

I've received my little Asrock folding system today http://www.asrock.com/nettop/overvie...%203D%20Series

Surpisingly I was able to overclock the GPU quite a bit! :D only sacrifising 4-5 degrees more on full load compared to the stock GPU clocks.

I've ran 3dmark06 on the nVidia GT 425M to see how it compares to my overclocked ATI 3850. The 3850 is faster, but it's not far off to be honest ;)

Here's the score with the 3850 (9314): http://3dmark.com/3dm06/15946967
Here's the score with the GT 425M (8288): http://3dmark.com/3dm06/15989378

Here's a pic of my overclock.
http://img148.imageshack.us/img148/3863/gpuzd.gif

It folds much quicker then the 3850 and much more power efficient ;)

http://img836.imageshack.us/img836/5793/folding.png

[GFreeman]

This thing is even not so bad gaming :) I'm running Dirt 2 at 1650x1024 with high detail and 4x AA settings.

Below my overclocking settings :)

http://img94.imageshack.us/img94/1949/27459131.png

<?xml version="1.0" encoding="UTF-8" ?>
<!-- Benchmark Mode Results -->
<!-- author: Douwe -->
<!-- date: 23:39:20 on 25/08/2011 -->
<!-- machine: Foldinghome -->
<benchmark>
<average min_fps="19.134340" av_fps="23.631243" min_fps_ms="52.262058" av_fps_ms="42.316860" />
<track>
<settings name="battersea" route="route_1">
<car name="sti" />
<car name="bmw" />
<car name="e9r" />
<car name="cr4" />
<car name="sor" />
<car name="sti" />
<car name="350" />
<car name="mer" />
</settings>
<results samples="2090" min_fps="19.134340" av_fps="23.631243" min_fps_ms="52.262058" av_fps_ms="42.316860" />
</track>
</benchmark>

Not a bad benchmark score right? :) It's playable and it isn't far off with my ATI 3850.. Well that plays with the ultra preset, this is too shoppy but still :)

[Rob Williams]

That is exactly why I bring it up. I never want to be a killjoy... but seriously, for the energy costs involved with the OC'd 3850, you'd be better off with that low end NVIDIA card.

I found this out the hard way. When the HD 4870 X2 dual-GPU card was new, I had to go on a business trip, so I set it up to Fold while I was gone. During the trip I got to talking to a friend about Folding, and I told him about what I did. He told me to call home and get someone to turn off the PC because I was effectively drawing so much power for so little gain.

Is actual progress being made here on the AMD side, or is it going to be pro-NVIDIA for a while? It's worth noting that I met the creator of the Folding project once in San Jose with NVIDIA, so to see the company excel at this is no surprise. But is it because of this partnership, or the fact that CUDA is just so damn efficient?

[DarkStarr]

Its due to the optimizations. Look at any other similar project and you will see AMD rips holes in Nvidia every time. Milkyway@home - AMD, Bitcoin mining - AMD, Anything@home except F@H - AMD, GPU password cracking - AMD. Basically they dont have it optimized for AMD, probably due to Nvidia but that is supposed to be changing but who knows.

[Kougar]

Gratz on those performance numbers GFreeman, I would'nt have thought a GT 425 would even give you that much in games!

Rob, DarkStarr, it's a tough call to answer. I'm sure some part of it is optimizations... but CUDA is just more efficient at this specific type of workload. Even taking into effect the doubling of performance for AMD cards... reported CPU usage is still ~25% for AMD folders. Guess how much CPU overhead is required for Folding on my GTX 480 @ 15,000 PPD? Less than 1%. Even if I drop down to a Quadcore system, it still measures the CPU processing overhead at <1% for the GPU client. Basically the entire Folding@home project can remain self-contained inside the GPU, with none of it getting outsourced to the CPU for computations it can't handle.

A very, very long time back NVIDIA's GPU folding used to eat ~10% of a four-thread capable system, so I am sure AMD's 25% processing overhead can be dropped significantly. But it has a huge way to go to still be comparable to NVIDIA, even assuming the PPD comparisons were equally matched, which they still are not.

[GFreeman]

Thanks Kougar! That is thanks to the overclock I run on it. It's overclocked to 37% :) I've also been playing with my fan speed settings so it runs 80 degrees tops when folding now. For folding I'm running this system at the lowest possible multiplier and vcore. According to AIDA64 it's using 6 watts running@700 Mhz with the lowest memory devider :) That's when the load is only on the GPU folding.

Here's some AIDA64 benchmark. I'm suprised what this little puppy puts out with a little tweaking / overclocking. I've included a zip file (contains PDF benchmark report).

[GFreeman]

Gratz on those performance numbers GFreeman, I would'nt have thought a GT 425 would even give you that much in games!

Rob, DarkStarr, it's a tough call to answer. I'm sure some part of it is optimizations... but CUDA is just more efficient at this specific type of workload. Even taking into effect the doubling of performance for AMD cards... reported CPU usage is still ~25% for AMD folders. Guess how much CPU overhead is required for Folding on my GTX 480 @ 15,000 PPD? Less than 1%. Even if I drop down to a Quadcore system, it still measures the CPU processing overhead at <1% for the GPU client. Basically the entire Folding@home project can remain self-contained inside the GPU, with none of it getting outsourced to the CPU for computations it can't handle.

A very, very long time back NVIDIA's GPU folding used to eat ~10% of a four-thread capable system, so I am sure AMD's 25% processing overhead can be dropped significantly. But it has a huge way to go to still be comparable to NVIDIA, even assuming the PPD comparisons were equally matched, which they still are not.

What I'm about to say now is how I feel about it and not factual so feel free to correct me haha! I have a feeling that nVidia get more out of there cards using technologies like CUDA.. If you look at the AMD cards it looks like they are about raw power. They have a lot more shaders if you compare them to the nVidia cards. Ofcourse nVidia is using a much different architecture compared to the AMD cards. It seems that AMD should even get a lot more performance out of there cards I guess..

[Kougar]

What I'm about to say now is how I feel about it and not factual so feel free to correct me haha! I have a feeling that nVidia get more out of there cards using technologies like CUDA.. If you look at the AMD cards it looks like they are about raw power. They have a lot more shaders if you compare them to the nVidia cards. Ofcourse nVidia is using a much different architecture compared to the AMD cards. It seems that AMD should even get a lot more performance out of there cards I guess..

In a nutshell, that's true. The important thing for everyone to remember is that Fermi was the first GPU architecture where NVIDIA combined their gamer & Tesla/Quadro market designs into a single core, a single product.

Fermi had to deliver all the GP-GPU and other workstation workloads, while at the same time being a powerful desktop GPU for gamers. Obviously in trying to be both, some compromises got made... but the point still stands Fermi was built for GP-GPU. AMD's designs are built almost entirely with gaming in mind. Until AMD decides to increase their core-size by incorporating additional GP-GPU hardware, there's always going to be a difference.

[DarkStarr]

See however you look at it its not CUDA, cuda is ~= to OpenCL which is what AMD uses now. Then take a look at all other distributed computing, AMD pwns typically several times over. I think they are not really trying because OCL is pretty similar to CUDA so therefore they could easily transfer optimizations that do the same thing on both cards and start on specific AMD ones. They have not. They implemented a BAD version of OCL, its stupid. I used a OCL bench tool and my 5750 scored higher than my 275 yet folding, my 275s each get ~9k while the 5750 doesn't even do half that.

The combined score (each 275 did about 50%) for the Nvidia cards was C1753.
The ATI alone was C2772.8.
This was done with the 275s highly OCed and the 5750 at scock 700/1150, I could easily push it to 870/1370.

So in essence my 5750 should kick ass in folding but it doesn't. Wonder why?

[GFreeman]

Thanks for the detailed explanation Kougar! Well I think AMD cards should do much better as they now score as well.. Let's hope there will be better future support for the AMD cards too..

Anyways, I have been fiddling with my bios settings and got a better overclock out of the CPU with adding a tiny little vcore :) The CPU can be overclocked more, but then I have to settle with a lower memory devider and I don't like that ;) I had to sacrifise a few Mhz on the ram, but it's still running @1400Mhz at CAS8. I'm happy with the performance it's much better then I expected. No throttling, it ran at 3150Mhz all the way :D This is with stock cooling and this machine is tiny. It runs a little bit hotter on full load, but not by much.

http://img836.imageshack.us/img836/2396/97196853.png

[GFreeman]

That makes quite a difference to the score in 3dmark06 :)

Here's 8819: http://3dmark.com/3dm06/16010672

[marfig]

Umm... the CPU will not affect your folding performance, GFreeman. F@H GPU client doesn't make use of GPU-CPU multithreading facilities in CUDA. It makes exclusive use of the GPU.

[GFreeman]

I know that :P When folding I'm only running@700Mhz with the lowest vcore setting ;) This configuration is when I use it for gaming every now and then. This system has become my main rig. I prefer it to the older X2 actually..

[GFreeman]

It's actually folding for 12/14 hours a day average. It folds when I'm at work. It's a little too noisy for night folding :) I have one room all in it at the moment. It is silent, but not silent enough for that purpose haha! Still looking for an apartment. If I have a seperate bedroom then it will fold at night as well.

[marfig]

Oh gotcha. Didn't realize you were thinking of a different purpose for it.

[GFreeman]

I have become a little too fond of it as a system too haha! But that doesn't mean it won't be folding. It will for the most part. This thing has 3d vision as well.. and a great 7.1 soundcard with digital output and USB 3.0.. Some features I really like. I've allready watched Avatar in 3D with it that was cool! I'm thinking making it my primair system for watching TV on the internet as well. A company is working at making TV channels online available in Holland. That would be fun having one multimedia system for such wide use. It even has a remote!

[GFreeman]

Windows is now reporting these scores with the i3 370m clocked at 3150 Mhz 1.176v of vcore :) Not so bad for this little HTPC. It has folded about 45 WU's so far.

http://img202.imageshack.us/img202/1084/scorekg.png

[DarkStarr]

I reran the OpenCL bench on my new 480 and got

GTX 480 C4281.1 at Stock on it.
GTX 275's (SLI) C1753.
ATI 5750 C2772.8.

So from those results their is no reason the 5750 should be folding that low.