EnglishPaul

I am quite surprised at how confused I am getting while trying to select the right program to use. And suspecting that I am not alone, I hope that someone will take the time to explain the options and differences between them.


In my case I have a new Advent QC-6003. It is an Intel Q8200 (Core 2 Quad processor), with a Intel GMA X4500hd graphics built into the motherboard. And to further complicate things, I believe it runs vista at 64bit.

Aus_Karlos

System Tray Client:
Small icon on the system tray that displays how much of the WU is completed, can right click on the icon to view a nice 3D protein, view stats or exit.

Console:
Run in the background under services (automatic start is optional). No user interaction wont be displayed anywhere on your system tray or taskbar. Will be listed under Services in Task manager.

I would use the System Tray Client.

EnglishPaul

Thanks.

I had grasped that, but I'm sure its useful for others.
I thought that, to best use a quad core, I needed to do something more advanced.
What is an SMP client?
What is deino versus MPICH? (Their description in Wikipedia was like a unknown world to me)
Are there any special requirements with a 64bit system?
How do you choose a GPU program, and would one work with a motherboard based card?
Can you run GPU and SMP programs at the same time?

I know all these questions relate to the page of complex downloads, but I would have thought it possible to choose and use an advanced program without being a computer guru... or should they be left well alone?

Aus_Karlos

For Dual/Quad cores. Uses more system resources, PC may not respond as fast when wanting to do something. And you wont be able to play games while its running.

Nope.

Nothing lower than the X1600 for ATI and i think 7 series on Nvidia, but dont quote me on that. However running the GPU F@H will use all system resources, you will be waiting for 5mins for a web browser to load up.

Yes, however in advance options you need to set the applications "MachineID" differently . Example SMP is configured as MachineID 1 and the GPU is machineID 2. IDs must be unique.

EnglishPaul

Thanks. F@H made easy!

Now there is just the small question of which of the 2 SMP clients to use: the one for deino MPI or the one for MPICH?

Aus_Karlos

MPI is 32bit code only where as MPICH is 32 and 64-bit. Although the Deino MPI is much more stable than MPICH. There is more it than that though, im not to sure myself.
http://foldingforum.org/viewtopic.php?f=46&t=4993

mattlock

If you're looking for the best points production you only choice is the SMP client. The GPU2 client for Nvidia will only run on CUDA compatible cards , and will not out perform the SMP client on anything less than an 8800GT. Some onboard GFX are CUDA compatible. See the link below for a list of CUDA compatible devices.

http://www.nvidia.com/object/cuda_learn_products.html

EnglishPaul

Nope, great shock, my Intel graphics chip is not on the CUDA list!
It can do directX 10 though. Would GPU1 client work, and are they still available?


The MPICH client is living up to its reputation of being unstable, so it looks like I have a bit of work on my hands to try and find a stable way to run it.
Early attempts seem to show that it works ok without -smp at 100%, taking up only 25% of my processors capacity; but with -smp it falls over with a write error. And sometimes it then falls over in -configonly !
Anyway, I will let it try to finish this unit without -smp, and then try a few config adjustments.

Doh, I hoped this thread would be an easy description for people with fewer computer skills, but it seems that SMP clients are a fast lesson in dos windows and stuff!

Aus_Karlos

Only certified GPU cards can run the GPU client. I think for ATI its the X1600 and above. For Nvidia its the 6 or 7 series. And CUDA GPUs are 8,9 and 200.

I gave up on the SMP client and stuck with the system tray. i manly use my system for gaming but ive yet to turn off F@H even playing GTA IV. Last time i checked it was at 150/500. Came of GTA after 2hrs and it was on 200/500.

mattlock

GPU Client:

GPU and OS Support

Which cards are supported?

We now support several classes of GPU boards, including X1600, X1800, and X1900 series GPUs from ATI. At the launch, we supported X1900 series cards only. X1800 cards do not provide the performance seen in X1900's and so we strongly recommend X1900 class cards. X1900 and X1800 cards are actually quite different -- they have different processors (R520, R530 vs. the R580). The R580 in the X1900 makes a huge difference in performance -- its 48 pixel shaders are key, as we use pixel shaders for our computations. Also note that the card should have at least 512MB of RAM, otherwise the GPU client will put a huge load on the client machine (although we do note that the 256MB X1950Pro using PCIe does work reasonably well on current projects).

What about video cards with other (non-ATI) chipsets?

The R580 (in the X1900XT, etc.) performs particularly well for molecular dynamics, due to its 48 pixel shaders. Currently, other cards (such as those from nVidia and other ATI cards) do not perform well enough for our calculations as they have fewer pixel shaders. Also, nVidia cards in general have some technical limitations beyond the number of pixel shaders that makes them perform poorly in our calculations.

GPU2 Client for Nvidia:

Basic Requirements:

* a GeForce, Quadro, or Tesla card that supports CUDA (G80 or later for the most part)
* A CUDA capable driver, version 174.55 is recommended. Or 177.35 for GTX 2xx cards (you can download the driver for Win XP, Win XP 64 bit, Vista, and Vista 64 bit).
* Windows operating system (32 or 64 bit), XP or newer (better Vista and 64-bit support coming soon)

GPU2 Client for ATI:

Basic Requirements:

* 2xxx/3xxx ATI Video Card, or newer
* ATI Driver v8.1+, v8.3 or newer recommended
* AGP GPU aperture size in the BIOS must be set to 128 MB or larger
* Microsoft .NET Framework 2.0, with updates recommended
* Windows operating system, XP or newer