Join the SatelliteGuys Folding@Home Team!

Wow, I'm really holding up the rear-end, aren't I? Anyone have a recommendation on a GPU for most PPD/$ spent?
This is the card I got from NewEgg: Newegg.com - ZOTAC ZT-98PES3P-FCP GeForce 9800 GTX+ 512MB 256-bit GDDR3 PCI Express 2.0 x16 HDCP Ready SLI Supported Video Card - Desktop Graphics / Video Cards, but I see one from PNY with a slightly slower clock speed and a rebate to bring it down under $100: Newegg.com - PNY VCG98GTXPXPB GeForce 9800 GTX+ 512MB 256-bit GDDR3 PCI Express 2.0 x16 HDCP Ready SLI Supported Video Card - Desktop Graphics / Video Cards

The Zotac cost $130 and was good for 5 Kppd on my son's 2.8 GHz Dell 4700; call it 38.4 ppd/$. As long as you get a 8600 card or better, you should be happy.

One issue that came up with the nVidia card on my son's PC: the User Interface response time when Folding isn't very good. The ATI HD4850 was much more responsive, but it sucks down a PC core unless you tweak some environmental variables.

Well, the HP xw8200 arrived today, and it is going to need some tweaking to make the best use of the hardware, I think. I'm currently running the ATI HD4850 card and the SMP client, but I can tell that the two don't play well together. By itself, the HD4850 should be able to generate close to 3,400 ppd; the SMP client on a 3.2 GHz machine with two cores should be producing 2,400 ppd, but right now I'm seeing 2,900 ppd.
 
Last edited:
Hey guys

My production has went down recently because I went to put my 2 8800Gs's in 1 rig and got a system 32 error so I had to fix that and now everything is up and running again. (making room for another card in another rig:D)


As Plazz would say, Fold on :):up

Welcome back Scorp.!!!

I too will lose a few points today as I awoke to my home clients starving for units and not sending and or receiving. :eek: It appears I was about 2 weeks past due for a firmware update on my home router. :rolleyes: Seems all is back to normal now.

Hey Foxbat what operating system are you running on the new machine? Just curious if you ar running the ATI with environment variables? I had to run my 2 3850's with environment variables to get my processor back for the smp client.

WOW Voyager is smokin..!!!

Well congrats everyone and FOLD ON...!!! :D:up

Plazz
 
Hey Foxbat what operating system are you running on the new machine? Just curious if you ar running the ATI with environment variables? I had to run my 2 3850's with environment variables to get my processor back for the smp client.
I will rebuild this box from the ground up some time this weekend. Out of the box, it had Windows XP SP3, but there was a bunch of added froo-froo that I won't be using on this box, plus the 320 GB drive I can use in an external USB enclosure and take the 120 GB drive and use the drive that's currently in their for the system disk. But it will be running Win XP Pro SP3 when I get done.

I also noticed the BIOS is a bit behind, so I want to flash that to the latest version. I read the Folding Forum on configuring the ATI client with the Environmental Variables, so I set those this morning. I also have Hyperthreading turned on, but I've seen some sites on the Internet that suggest the HT in the Socket 604 Xeons is flawed and you're better off without it. Tweak-tweak-tweak, so many things to try! :cool:

Hey, I see the Team had our best day ever yesterday! 138,935 Points and 250 Work Units! Plus, we breezed past 41,000,000 total points! Keep up the good work, and Fold On!
 
After a night of running with the ATI-recommended System Environment variables (and some time lost because they DO require a reboot after changing them!) my ATI card is now Folding at the same level it was when it was in the Dell with the 2.8 GHz P4 w/HT. I guess the limit is the GPU and not the CPU. InCrease reports this GPU2 client is up to 2,500 ppd.

The plus of running the ATI GPU2 Client with the BROOK_YIELD set to "2" is the SMP client now is up to 760 ppd. Something tells me that this may be all that I can expect out of a Core_A1 SMP running on two 3.2 GHz Hyperthreaded Nocoma CPUs.
 
After a night of running with the ATI-recommended System Environment variables (and some time lost because they DO require a reboot after changing them!) my ATI card is now Folding at the same level it was when it was in the Dell with the 2.8 GHz P4 w/HT. I guess the limit is the GPU and not the CPU. InCrease reports this GPU2 client is up to 2,500 ppd.

The plus of running the ATI GPU2 Client with the BROOK_YIELD set to "2" is the SMP client now is up to 760 ppd. Something tells me that this may be all that I can expect out of a Core_A1 SMP running on two 3.2 GHz Hyperthreaded Nocoma CPUs.

Foxbat the way I run my environment variables was to make a small .bat file that I run before starting the GPU client. I found it made it much easier to change the variables and test. Yes I found also that with running the environment variables you never really get the full 100% out of the GPU. I did figure you might do a tad better than that on your smp client though.

Fold On..!!! Plazz
 
Foxbat said:
Hey, I see the Team had our best day ever yesterday! 138,935 Points and 250 Work Units! Plus, we breezed past 41,000,000 total points! Keep up the good work, and Fold On!
Looks like we blew those totals away....Yesterday we had 150,019 Points and 271 work units!
Well, another record-setting day! Sunday, May 3rd, 2009: 163,525 Points and 288 Work Units completed! Woo-Hoo!

Hey, Voyager6, every time I think I've got enough horsepower to catch up with your daily points production, somehow you manage to kick it up another notch. Now that I'm at 12K ppd, you're at 13.5K ppd!
 
Last edited:
Voyager6, you must have realigned the warp coils and flushed the deuterium conduits! You're flying now, so far over 14K points today! Have you fired up a new Folder?
Yep, you guessed it. I brought another 9800GT online last week. I will be installing one (or more) of the CPU clients today to make use of the quad core.:D I also did some tweaking on the other two folding machines. New CPU coolers and Riva Tuner help those machines run more efficiently.
 
Congrats to Scorpion.sting for reaching the 7 Million points mark and to Plazzman for reaching the 3 Million points mark. :up:up
 
I'm still debating whether it would be better to swap the two GPU clients. Even with the Environmental Variables, it seems there is a definite CPU hit when running the ATI GPU2 client. Since my son's Dell doesn't run anything but the GPU2 client, it can handle the extra CPU loading. The SMP client on the HP xw8200 definitely runs faster when the ATI client isn't running, so it would prefer the lower CPU requirements of the nVidia GPU2 client.

Doing that, however, requires major downtime on both systems as the P/S in the xw8200 may not have the right power connectors for the Zotac card I got (it has a funky 8-pin 12V connector instead of the usual 6-pin PCI-e connector). I'd rather not swap power supplies along with the video cards and I'm not really interested in buying yet another power supply! I think I can make it work, though, with the adapters that came with the ATI card and the nVidia card.

At least I figured out the Hyperthreading question: I saw SMP frame times of 31m27s with the HT on, and 45m12s with it off on the same Work Unit. HT stays on. The SMP client is still contributing some, but not much, to the cause.
 
I'm still debating whether it would be better to swap the two GPU clients. Even with the Environmental Variables, it seems there is a definite CPU hit when running the ATI GPU2 client. Since my son's Dell doesn't run anything but the GPU2 client, it can handle the extra CPU loading. The SMP client on the HP xw8200 definitely runs faster when the ATI client isn't running, so it would prefer the lower CPU requirements of the nVidia GPU2 client.

Doing that, however, requires major downtime on both systems as the P/S in the xw8200 may not have the right power connectors for the Zotac card I got (it has a funky 8-pin 12V connector instead of the usual 6-pin PCI-e connector). I'd rather not swap power supplies along with the video cards and I'm not really interested in buying yet another power supply! I think I can make it work, though, with the adapters that came with the ATI card and the nVidia card.

At least I figured out the Hyperthreading question: I saw SMP frame times of 31m27s with the HT on, and 45m12s with it off on the same Work Unit. HT stays on. The SMP client is still contributing some, but not much, to the cause.

Your right, ATI cards use a full core while Nvidia only uses like 2-3% of a core.
 
I have a few quick questions

I'm using an Intel E4500 core 2 duo that I'm currently running two of the regular CPU clients on. Would I be better off running the smp client instead since it's just a 2 core CPU?

Also I want to add a graphics card to run the gpu2 client, but didn't know which card I would be better off with.

The 8400GS and HD4350 are the same price, and a 9400GT and HD4550 are about $12 more.

Any suggestions on the graphics card would be appreciated.

Thanks.
 
I ordered the 9800gtx+ card:

Newegg.com - EVGA 512-P3-N871-AR GeForce 9800 GTX+ 512MB 256-bit GDDR3 PCI Express 2.0 x16 HDCP Ready SLI Supported Video Card - Desktop Graphics / Video Cards

I'm excited to connect it to make it work. I ended up ordering a PCI to molex adapter for my 2nd power supply because I don't have enough capacity on my main power supply to power the new GPU. I also ordered a HDMI to DVI adapter to try connecting my GPU to my home theater via HDMI. Hope it all works.

My PPD should be skyrocketing 'soon'.
 
I ran into a thread on the F@H forums where a guy though someone else was 'cheating' by getting around 1900 ppd (or more) on a quad core processor. I thought, "sounds about right since I only get about 500 ppd on my quad-core processor (q6700)". So I read the thread and the assumption was wrong. Then I investigated how to make my CPU more effective. (I've already given up my GPU because it can't complete a WU under the max time limit).

It took a while and lots of junk to read through, but I was able to download, install, setup & run the SMP client on Vistax64 using all 4 cores. Now my CPU is maxxed out now and I'm looking forward to seeing what ppd I get. The 1st WU I downloaded is SMP Gromacs at 1760 points.

Some of the things holding up the show were issues making my 64-bit Vista machine actually run the program, using 'run as admin' in the command prompt, following all the steps in the guide (Folding@home - WinSMPGuideMPICH), installing it to a non 'program files' directory, using the right username/password, and setting the config to 'big' instead of the normal size. Apparently there were no files to download for the normal/small range, and when I changed it to 'big' I got the SMP Gromacs right away and started folding.

I wish it was easier to sort through the sea of info regarding folding at home setup with all of the different options. It was like this was hidden for so long until I stumbled upon it. The SMP download location was not obvious and it wouldn't work for so long, that a normal user would probably give up and try the regular tray icon and miss out on maximizing their performance.

So far it looks like 12 minutes per percentage point - so it should take about 20 hours to complete the 1760 point WU. That's about 2100 ppd, which is 4 times the points I was getting before running my CPU at 50% load all the time.
 
bebop, when you say you gave up on the GPU client, do you mean the 9800GTX+ card you ordered? I can't imagine it would take too long to complete with your Q6700; my son's Dell 2.8 GHz P4 can crank out a nVidia WU in about 3 hours. According to the WU, it has a deadline of 6 days.