Quote :

"Anyone folded a T9500 yet? I'm getting a laptop with one, curious how many points they get."

Here is a very limited report on the T9300 I found over at foldingforum.org

Project---CPU Speed----PPD-----PPD/Core---PPD/GHz
-2653------2500MHz---1561.24---780.62------312.25



[Edited on May 6, 2008 at 4:07 PM. Reason : quote]

5/6/2008 3:59:20 PM

cool, thx



Support Cancer Research and the NCSU Wolfpack With Your Computer's Idle Time!

What Is It?
Essentially Folding@Home uses your unused CPU cycles to model how proteins fold in the human body. The misfolding of proteins is often involved in well known diseases, such as Alzheimer’s, Mad Cow (BSE), CJD, ALS, and Parkinson’s disease. Because this endeavor requires massive amounts of computing power, their project uses spare CPU time on computers that would otherwise sit idle in homes and businesses across the world. Here at North Carolina State University, we have formed this team in order to donate our spare CPU time to help this important endeavor.

The North Carolina State University Folding@Home Team (Team # 59) was one of the first teams to join the Folding@Home project, and was the very first University-based team to join. We are currently ranked in the Top 100 teams in the world and are ramping up production in an effort to make it into the Top 50. We are all dedicated to doing our part to help end some of these terrible diseases, and we would appreciate the support of anyone who would like to help.

All it takes is the installation of a small piece of software and your computer will start completing Work Units for the project and the team. The Folding@Home process runs at the lowest possible priority so you can run it 24X7 without slowing down your computer or interfering with your everyday computing tasks. If you have a Sony PS3 Game Console you can also run Folding@Home on it with great results.

How can I Help?
Download and install the software appropriate for your OS (Windows/Linux/Mac). For dual/quad core computers, be sure you install the SMP client. It will get you and the NCSU team far more points than the traditional client. Whichever client you choose, we recommend running the Folding@Home client as a service. The client will run in the background without notification, and will release CPU resources as needed since it runs at the lowest priority. The NCSU Team Number is 59.

Visit the team homepage for detailed installation instructions

Windows Computers
We recommend using the "text-only console” client for most windows computers. If you have a dual/quad core processor, be sure to install and run the SMP client to get you and the team far more points than the standard console client In general, its best to stay away from the graphical clients. If you have a Pentium 4 with Hyperthreading, its best to turn off Hyperthreading in the BIOS if you know how. Feel free to post in this thread with any questions.

Mac and Linux Computers
There are both "text-only" consoles and SMP clients for Linux and Macintosh computers too! If you have a dual/quad/etc core processor be sure to run the SMP client as you'll get far more points than the text-only client. But the regular console clients also help out too if you only have single-core processors.

Playstation 3
PS3s currently provide over 60% of the processing power in the Folding@home network and are capable of producing many points for the team. If you have PS3 system version 1.6 or later, you will see a Folding@Home icon in the Network column of the XMB (PS3 Networking menu). Just click on the icon and that's it. If you don't have 1.6 or later, perform a system upgrade. Make sure to use Team #59!

The more computers and PS3s you run the clients on, the more points your username and the NCSU team earns! Install as many clients as you can and watch your username/team progress at the stats page


Links

Team Links
Official NCSU F@H Team Website - Tell your friends people. It's ncsufolding.com
NCSU F@H Team Discussion - This page.
NCSU F@H Team Facebook Group - Join our facebook group!

F@H Links
F@H Project Homepage - Official Stanford project webpage
F@H Official Forums - Official Project forums. Full of great information about the project and the computer/ps3 clients
F@H Wikipedia Article - General information about the F@H project
F@H Wiki Site - Excellent Wiki site with tons of helpful information about the project. Geared towards F@H users like us.

Stats
Extreme Overclocking NCSU Team Stats Page - The best F@H stats site
Stanford Stats Page - Official Stanford stats site
Monitor your production - Extremely useful little utility. Monitor the point production of local/network F@H clients
Compare your production with other users - Use information from above to compare your points production with other folders

5/6/2008 4:00:17 PM

what kind of points can i expect with Core 2 Extreme X9000?

5/6/2008 6:02:52 PM

^ also i should note its been overclocked to 3.0 from 2.8 stock but the bios doesnt let me go 2x grrr

5/6/2008 6:54:41 PM

i can't find any numbers thru google...

check here though: http://fahinfo.org/

5/7/2008 9:14:07 AM

bump

5/8/2008 12:52:25 PM

uh o j33pownr is catching up to me

5/9/2008 8:58:23 AM

our team is looking pretty solid... if we can keep up the 70k ppd, we'll move into 81st in 3 months (we already are back up to 84

5/9/2008 10:56:47 AM

and THAT would be our highest ever right?

5/9/2008 11:04:29 AM

*sigh* okay fine

i finally installed fah onto my computers

one on my desktop (athlon 64 4000+)
one on my macbook pro (core duo 2.16ghz) w/ smp client
one on my home mega server (ibm x345, 2 xeons @ ~2ghz) w/ smp client

will install on my other stuff later, i'm lazy


i'll install it on a bunch of my stuff at work
might consider putting it on our esx farm
(each machine has 16 logical processors, we have 15 of these boxes)

so with 247 processors folding, will i win??!?!

5/9/2008 11:38:12 AM

holy crap man, welcome!

Only problem with the virtual machines is you need two cores to run the smp client. does each of those processors have 2+ cores? i'm thinking not since you said 'logical.'

where do you get the number 247 from? but sure, if you ran 247 processes you'd jump to the top of the team pretty quickly. assuming those processes are mostly SMP clients, thats about 60 smp clients, which at a conservative 700PPD each would yield 42,000 PPD So yeah, you'd win

5/9/2008 11:50:02 AM

there are 4 physical processors in each box
each is a quad core opteron

will the smp client be able to use all 16 logical ones on every box?

we're only utilizing ~50% of our available processing power for VMs right now, so they're just sucking up power as is.

15 * 16 = 240 + 1 (athlon) + 2 (macbook pro) + 4 (x345) = 247

5/9/2008 11:52:24 AM

NICE, we could use your help!

(if you have a PS3 this also works on PS3's)

i'm sure we'll see you jump to the top quickly!

i'm not to sure how it will work on a machine with 16 processors.... you may need to run more than one SMP client. (like ~4 SMP clients per machine), set the -local flag on each & config each client with a different machine ID

[Edited on May 9, 2008 at 12:29 PM. Reason : .]

5/9/2008 12:26:04 PM

noob question: what does PPD stand for?

5/9/2008 12:28:46 PM

points per day

rule of thumb = 1 SMP client per 4 cores

[Edited on May 9, 2008 at 12:30 PM. Reason : .]

5/9/2008 12:29:39 PM

ah so the 1 client won't thread more than 4 cores?

and is there an smp client that works with RHEL (well, it's vmware esx server which is essentially rhel)

5/9/2008 12:56:34 PM

Quote :

"there are 4 physical processors in each box each is a quad core opteron will the smp client be able to use all 16 logical ones on every box?"

so you have 4 physical servers, with 16 cores on each physical server.

so you have 15 VMs you're responsible for? Which servers are they running on? How many cores do you have allocated to each virtual machine? what OS are the VMs running?

and i really doubt you'll be able to run any clients on the ESX level (on the entire machine)...but i've been wrong before

5/9/2008 12:59:05 PM

^^not that i'm aware of... one client only launches four processes and distributes the cpu load between the four... even with the Q6600 Quad-core processor people have gotten more PPD running two SMP clients rather than one... you may just have to test it... but I can guarantee you that one client won't use all your processing power.

folding & rhel
http://ubuntu-tutorials.com/2007/05/23/folding-home-installation-configuration-automation-ubuntu-debian-rhel-fc/

[Edited on May 9, 2008 at 1:00 PM. Reason : .]

5/9/2008 12:59:17 PM

Excellent script here for installing FAH on Linux box. Totally automated with a few prompts, and is even nice enough to assist with running on boot.

http://ra.vendomar.ee/%7Eivo/

It's called "finstall".

5/9/2008 3:18:37 PM

Quote :

"so you have 4 physical servers, with 16 cores on each physical server. so you have 15 VMs you're responsible for? Which servers are they running on? How many cores do you have allocated to each virtual machine? what OS are the VMs running? and i really doubt you'll be able to run any clients on the ESX level (on the entire machine)...but i've been wrong before"

no

15 physical servers with 16 cores on each physical server.

i'm responsible for the 15 physical servers.

i'm not talking about running the client on the VMs, i'm talking about running it on the esx hosts. (on a side note, i'm responsible for just under a thousand individual VMs)

i'm pretty sure i can get the smp client working under esx, it's just RHEL... i'll try.

5/9/2008 4:31:02 PM

SMP Help: I'm trying to install the SMP beta client on a Windows Vista machine. I run the installer, then run install.bat as admin, then I tried running fah.exe with admin privledges in the compatability mode and I get an error:

Quote :

"Unable to connect to 'Dick-PC:8676, sock error: generic socket failure, error stack: MPIDU_Sock_postconnect(1228): unable to connect to Dick-PC on port 8676, exhausted all endpoints (errno -1) MPIDU_Sock_post_connect(1275): unable to connect to Dick-PC on port 8676, No connection could be made because the target machine actively refused it. (errno 10061) [21:18:50] Error running mpiexec. Please check your configuration and re-run install.bat if necessary"

I've been waiting forever for this thing to go out of Beta, but who knows when that will happen.

5/10/2008 5:25:12 PM

^turn off UAC, it hates the smp client

5/10/2008 5:30:02 PM

yeah get rid of UAC

i'd make sure youre running the 5.91 version

are you running any firewalls besides windows firewall? id so disable, also disable antivirus for testing if none of the above works. the account you're using in the install.bat is an admin account right?

also try installing as a service like the ncsufolding site talks about

5/10/2008 8:39:56 PM

I'm closer, but now I get the following unknown error

Quote :

"[14:21:54] Client-core communications error: ERROR 0x1 [14:21:54] Deleting current work unit & continuing... Folding@Home Client Shutdown at user request. Folding@Home Client Shutdown."

5/11/2008 10:35:10 AM

http://fahwiki.net/index.php/Error_0x0_and_0x1

5/11/2008 11:50:47 AM

^^ did you let it keep running? i'm thinking it may of worked it self out. if not delete the work folder, and the FahCorea1 file, and restart the client (or the service)

5/11/2008 12:10:13 PM

<----- Broke Top 100 for first time

5/11/2008 3:57:21 PM

I have no idea what's happening. I set it up as a service as the website explained, but after a reboot there was no CPU usage and the FAH log file did not update. I only had a limited amount of time to set it up anyway, and now I don't have access to the machine. I'll just keep an eye out for changes in my PPD.

5/11/2008 6:12:36 PM

go to services.msc and see if the service is running. i doubt your PPD are going to change if you didn't see the CPU getting used. check the log file and see if it says anything

5/12/2008 8:41:49 AM

damn the last 4 days have been kinda crappy

5/12/2008 11:14:41 AM

just noticed i have a new 1275pt work unit which is giving me only 350PPD on a coreduo machine

is it just me or are they clocking down the number of points produced by SMP clients?

5/12/2008 5:15:48 PM

I was getting that junk 2 weeks ago but im back to 1760 pt WUs on both c2d machines.

5/12/2008 8:43:20 PM

well im about to turn in this crap 1275pt workunit, hopefully i'll get a 1760 again...

5/13/2008 9:51:52 AM

is there a way to tell the program to dump the current work unit and get a new one? i left my power cord at home yesterday and i know i've missed the deadline for this current unit, but it's still chugging through it...the deadline's passed, though

5/13/2008 10:03:07 AM

gg on the first few workunits MagicalLeoplurodon

plz to do many more

5/13/2008 10:28:12 AM

^^as long as it's chugging, don't stop it... i've submitted many WU's after the deadline and gotten points for it.

if you stop it and restart the client, it will recognize the expired WU, delete it, and fetch a new one.

or you can stop the client, delete the queue.dat file and the work folder.

5/13/2008 12:39:07 PM

Q6600 for $180
http://microcenter.com/single_product_results.phtml?product_id=0287043

5/14/2008 2:01:26 PM

finally got my Q6600 back up

5/17/2008 7:03:46 PM

nice, we needs it

5/17/2008 7:34:32 PM

We just lost 20 active users. I'm guessing this is due to summer break.

5/17/2008 8:03:01 PM

Just added my new Q9450 on board, looks to be humming through quite nicely. Plus I went through all my LogMeIn pc's and checked them all and ended up added 2 more to my name.

BTW: got the proc to 3.1GHz stable, with the ram at 1590MHz. Im pissed, the ram wouldnt clock stable over that. Plus the C2Quad was approaching 65 degrees C at 3.2Ghz, so damn I gota run it at 3.1GHz.



Also, what do yall suggest? Large Work Units (>5MB) or Normal?

[Edited on May 18, 2008 at 11:58 PM. Reason : .]

5/18/2008 11:50:38 PM

that should put out some nice points

5/18/2008 11:53:16 PM

Well, according to the future calculation on the Q9450 box, it said 3300 PPD.

5/18/2008 11:55:26 PM

damn, is that with two smp clients or just one?

5/19/2008 12:06:00 AM

On my Q9450 box, it has one SMP client that is running 4 cores/threads.

Gimme a few days to get a real average on PPD. That was generated by FahMon.

5/19/2008 12:42:03 AM

well that should be pretty damn accurate as long as you don't run some video encoding at the same time.

if it gets that much with only one smp client, i bet it gets well over 4k with two. if you wanna try it out and need help setting it up let us know.

i'm running two on my Q6600 box and went from 2200PPD (running 1) to 3300 PPD with the right workunits.

5/19/2008 1:11:25 AM

So what exactly do I need to do in order to set up 2 SMP clients?

5/19/2008 10:12:37 AM

Quote :

"i finally installed fah onto my computers one on my desktop (athlon 64 4000+) one on my macbook pro (core duo 2.16ghz) w/ smp client one on my home mega server (ibm x345, 2 xeons @ ~2ghz) w/ smp client will install on my other stuff later, i'm lazy i'll install it on a bunch of my stuff at work might consider putting it on our esx farm (each machine has 16 logical processors, we have 15 of these boxes) so with 247 processors folding, will i win??!?!"

what happened to ya evan? i thought you were gonna jump to the top


^

in order to do it cleanly, i would wait for your current work unit to finish, then uninstall your current client.

then install one smp client to a folder outside of your program files folder (i hear using that folder can cause trouble for dual smps, havent seen it though)

in the install options, the only thing you do different go into the advanced options (when its asking you all the questions), and set the machine ID of one of the clients to 2 (anything other than 1).

then install another client to a second folder, advanced options and set the machine id to 3.

thats the only difference. i'd set both up using the service option and using the installation instructions here http://www.ncsufolding.com/?page_id=4

if you have any questions/problems let us know

5/19/2008 2:57:05 PM

^all you need is two separate installs in two separate folders, use the -local flag, and different machine ID's

5/19/2008 5:22:30 PM

i dont think you even have to set the local flag. im 99% sure it does that itself when you set the machine ids

^ which is basically what i said...i just like to give more details in case they're not familiar with f@h

5/19/2008 8:02:57 PM