• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Possible Alternative to Linux VM?

W

Wheresatom

[H]ard|Gawd
Joined
Mar 20, 2007
Messages
1,390
I just saw this post in lifehacker and the first thing I thought was how cool it would be to be able to run Linux SMP in that (having never tried anything with virtual machines) and then use vista for my GPU2.

The article talks about a way to pretty much run Ubuntu inside Windows as if it were just another program. I might try this out later tonight or tomorrow but if anyone can give a good reason it wont work, I wont bother.

Does Linux still give better ppd for equivalent harware?

 
This is very interesting indeed. If someone wanted to make a version with F@H SMP and then host that file that would be even better (for us truly lazy people).

I'm going to download to give it a whirl. I am curious about running multiplie instances of this though. It is not mentioned in the article. I am also curious about "copying"/"cloning" so we don't have to do the F@H SMP install in each and every instance that we want to create.

EDIT: downloaded, but can't seem to get it to run in Vista Ultimate x64. No error messages so I can't even debug it :(

 
Last edited:
LinSMP requires 64-bit Linux to run. I was unable to find any documentation regarding which version this uses, but I think it's pretty likely that it's the 32-bit version, for the sake of compatibility. If someone does get this to work well though, I would be very interested, since at the moment I can only fold on two of my cores with the free version of VMWare.
 
I'm using the free version of VMware and I have 2 instances open, each running on 2 cores. What issue are you having with running a second instance?
 
Logan321 said:
I'm using the free version of VMware and I have 2 instances open, each running on 2 cores. What issue are you having with running a second instance?
Click to expand...
The ATI GPU client requires one full core, and it would be somewhat pointless to run a second VM on the one remaining core. So it isn't that I can't run two instances, it's that I can't run one instance while taking advantage of my three available cores.
 
Zero82z said:
The ATI GPU client requires one full core, and it would be somewhat pointless to run a second VM on the one remaining core. So it isn't that I can't run two instances, it's that I can't run one instance while taking advantage of my three available cores.
Click to expand...

I'm down to 10% or less of one core using CAT 9.3, and the BROOK_YIELD, CAL_NO_FLUSH, and FLUSH_INTERVAL environment variables
 
[H]ugh_Freak;1033928209 said:
I'm down to 10% or less of one core using CAT 9.3, and the BROOK_YIELD, CAL_NO_FLUSH, and FLUSH_INTERVAL environment variables
Click to expand...
I haven't had a chance to play around with that stuff yet.
 
Zero82z said:
The ATI GPU client requires one full core, and it would be somewhat pointless to run a second VM on the one remaining core. So it isn't that I can't run two instances, it's that I can't run one instance while taking advantage of my three available cores.
Click to expand...

Ahh... right. I remember that from my X1950XT. Was a pain in the ass. I think it was one of the reasons I switched back to nVidia after that. Getting 12k ppd with q6600 @ 3.6GHz and GTX 260 @ 700/1500/1200.

You could setup two VM that either overlap on a single or on all 3 unused cores then you can limit it from accessing the last one of 'em by editing the folding.vmx file with notepad, and adding to the bottom:

priority.ungrabbed = "idle"
processor0.use = "FALSE"

You can set the "0" to whichever core you want ignored, or add more than one processor#.use line.

Here's the url that I got mine setup from... http://www.evga.com/forums/tm.asp?m=100498063&mpage=1&key=&#100498063

Does it significantly affect the GPU client to be floating, and not locked to a specific core?
 
Zero82z said:
LinSMP requires 64-bit Linux to run. I was unable to find any documentation regarding which version this uses, but I think it's pretty likely that it's the 32-bit version, for the sake of compatibility. If someone does get this to work well though, I would be very interested, since at the moment I can only fold on two of my cores with the free version of VMWare.
Click to expand...

This is exactly what I was thinking, there have been a few projects like this in the past (coLinux, andLinux for example) and they've all been 32-bit. And since the newest LinSMP client is a 64-bit app now, it appears now more than ever that a 32-bit Linux client is never going to happen.
 
Logan321 said:
You could setup two VM that either overlap on a single or on all 3 unused cores then you can limit it from accessing the last one of 'em by editing the folding.vmx file with notepad, and adding to the bottom:

priority.ungrabbed = "idle"
processor0.use = "FALSE"

You can set the "0" to whichever core you want ignored, or add more than one processor#.use line.

Here's the url that I got mine setup from... http://www.evga.com/forums/tm.asp?m=100498063&mpage=1&key=�
Click to expand...
Hmm, I might give that a shot. I'm not sure how that would affect my folding performance though, considering both clients will be sharing a core.
Logan321 said:
Does it significantly affect the GPU client to be floating, and not locked to a specific core?
Click to expand...
Yes. I had to lock it to a single core, because otherwise it would get stifled.
 
Try this program: WinAFC

I use it on all of my boxes and it works perfectly. EVERY TIME!!!!

I set the configuration as necessary and I get maximum PpD from WCG & F@H GPU2.
 
omega-x said:
is there a way to automate this?

every time i set affinity, it resets at the next WU.
Click to expand...

It doesn't automate it in the way that one of those other programs might (I haven't used them) but in the task manager you have to set the affinity of the Folding@home.exe process and not just the FahCore_11.exe. As long as you don't restart F@H via the system tray every WU will still be set to which ever core you set Folding@home.exe to.







 
Thanks for that tip jimbowdang :) I was wondering why my affinity kept on resetting itself after every WU... Does it still keep the affinity if you were to shut down the client, and restart it, or does it only last as long as the client is running?
 
I did some more poking around on the link in the OP, and this thing is built on top of coLinux, which is 32-bit only. So no LinSMP goodness here for now.
 
alan2308 said:
I did some more poking around on the link in the OP, and this thing is built on top of coLinux, which is 32-bit only. So no LinSMP goodness here for now.
Click to expand...

I looked on the SourceForge site and there is a request for x64 support.
Maybe that will happen soon.

In the long run, SMP2 may be out best hope.

Begin the "Waiting Game" ....... ---> NOW
 
Lol, you are right EA. The SMP client is one I am still interested in because of it's versatility, but Stanford seems busy with the GPU right now.

Thanks for your research Alan. You are definitely the man to go to with Linux questions around here.
 
Wheresatom said:
Lol, you are right EA. The SMP client is one I am still interested in because of it's versatility, but Stanford seems busy with the GPU right now.
Click to expand...

You have to remember that they have been working on the SMP2 for a while as well.
We just don't have any details about it. Would also explain why there is no Linux x86 version of SMP.

We shall just hope
 
You must log in or register to reply here.
Back
Top