My New Toy

rscosworth

rscosworth

rscosworth

Associate
Joined
18 Oct 2002
Posts
1,704
Location
Glasgow
Guys,

Just thought i'd have a bit of an internet willy waving session.

My new toy just arrived.

Lenovo Thinkstation S20
Intel Xeon W3540 Quad Core 2.93
8GB RAM
2 x 250gb 7200 HDD
NVIDIA Quadro FX-1800 768MB PCIe
Vista Business 64

Just loaded it up, nice machine, i cant actually hear it which is a real bonus as i thought it would sound like a helicoptor.

Just thought id share.....
 
I dont game :-( i have no idea what im going to do with it in all honesty.

Next week i have a Thinkstation d20 arriving, Xeon X5570, 8gb, 147gb SAS and a Quadra FX-4800 1.5gb card.....
 
Any chance of selling the fx-4800 on at far below retail? I'd love to have a proper quadro card instead of the modded 8800 I'm running now. Cad work for me, I don't game either.

Could run folding on it while you wait to think of a use? Internet seems to think the quadro cards will fold better than their equivalents, but it's pretty hard to find an example of anyone who's tested it. I found no significant gains in folding after getting my card to work with quadro drivers, but it did very good things for most other benchmarks
 
Wasn't a completely serious request :)

Even if you don't fold for that long I'd quite like to know what sort of ppd they throw out. It turns out it's really hard to gauge how close in performance to a fx3700 my card is without actually having the more expensive card available
 
Here's a better guide than I could write http://forums.overclockers.co.uk/showthread.php?t=17538642

However I will give it a go. Here is the file http://www.stanford.edu/group/pandegroup/folding/release/[email protected]

You unzip it somewhere like c:\folding. Make a shortcut to the main file, and append -configonly to the command the shortcut runs by going into properties of the shortcut.

When you run the shortcut, it'll ask you for a name, and which team you want to be, 10 is ocuk. The rest you can leave on defaults, if you follow the configuration through you can set it to run whenever windows boots. Otherwise you run it using the original file, not the configuration shortcut. IT should turn up in task manager, download a small file, and proceed to simulate how the proteins might fold. It's in the interests of medical research / beating other teams :)

Thats enough to run it, but not monitor it. http://fahmon.net/downloads/current.php?windows will give you an installer for fahmon, which monitors folding at home. You run that, choose to add client and navigate to your c:\folding\gpu folder. It'll then track it for you, giving ppd estimates in the bottom right corner.

Pretty sure on all of the above. Basic idea is to have computers doing something constructive rather than sitting idle. As a side effect, the two computers listed above would do pretty well in terms of the leader table here, which is always nice :)
 
Yep, it'll do just fine under vista :)

If anything it's easier, graphics card folding under linux is quite a mission really. Your cpu can do this too if you find yourself reasonably enthusiastic. Good way to spend an afternoon, stanford might do useful things with the data. I believe the folding at home project now qualifies as one of the worlds biggest supercomputers which is also pretty cool. Introduce yourself a few forums down under distributed computing, astonishing how much enthusiasm you'll find.

I wouldn't be running it without the people there :)
 
Ok, ive set this up but it sits on Completed 0 out of 1500000, and in FahMon it doesnt seem to move and the PPD doesnt change from 0.00 on the bottom right hand.
 
That's fun. I take it you now have a terminal open with the client running inside, and fahmon open as well. If it's got that far, then it looks like it has downloaded work and just hasn't started yet. Give it 10 minutes or so, and if its still not running it may mean their server is misbehaving.

You've passed my knowledge if it refuses to start sadly, I'd have to guess you need to change a setting to get it running with vista. These guys will have the answer though http://forums.overclockers.co.uk/forumdisplay.php?f=39

Frustrating that I can't help further
 
Success. Yeah, it would be. Probably at quite a low percentage. Its running on your graphics card, assuming you're running the client I had in mind.

The cpu clients either run on one core, leading to needing to run many separate copies of them. Or there is a SMP multi threaded one, which gets many more points but might not be working well. Linux is far superior for cpu folding, windows for gpu folding
 
Depends a bit where the comma is, if that's 316 it probably means you're running the cpu client rather than gpu.


Ahh... i may have given you the wrong link :s

And indeed I have. http://www.stanford.edu/group/pandegroup/folding/release/Folding@home-Win32-GPU_Vista-623.zip if you still trust me after that :)
Process is exactly the same as before, except say yes to changing advanced options and set machine id to 2
At which point you'll be running a cpu client on one core, and a gpu one on the gfx card. Shortly after that you'll notice why most people fold on graphics cards

Sorry again, I should've checked the link
 
Last edited:
Sorry mate, was away from the computer for a while. That's interesting, I presume you're running nvidia's latest driver?

I've just checked, the quadro 1800 is roughly equivalent to a 9600gt. That makes it recent enough hardware to be cuda enabled, so all should be good. I get that complaint when I try to fold using linux, but all behaves itself under windows. Very strange
 
Yep, tried with the standard Lenovo drivers and nvidia drivers, i also done a clean install of vista 64 (not preload) and used the nvidia drivers to no avail!
 
I'm out :( hopefully one of the distributed computing guys will have an idea. I'd have recommended reinstall and trying with the vista quadro drivers. Hope to see a thread in the distributed forum soon :)

Thanks for the effort you've gone to so far man, no worries if you've had enough
 
You must log in or register to reply here.
Back
Top Bottom