Notfred's VM Appliance + VMWare Player Guide

PeddlerOfFlesh · 03-11-09, 02:27 PM

OK, this is my first attempt at a guide, so it may be missing stuff and I'm certain it will be ugly (at least to start). There seems to have been a lot of interest in this, but a lot of questions. Here goes.

How to set up Notfred's VMWare Virtual Appliance with VMWare Player.

1. Obtain and install VMWare Player. (VMWare) (Filehippo) Filehippo doesn't require registration. I think the only option during install is the install directory. Other than that, any good firewall will probably pop up asking if you want to allow the newly added virtual network cards to have access to your computer or LAN. Make sure you allow these.

2. Obtain the virtual appliance image. (link)

3. Open the zip containing folding.vmdk and folding.vmx. Extract these to whatever directories you want to use, say C:\folding\vm for example. If you have a quadcore and are planning to use 2 VMs with static IP addresses (suggested, for monitoring purposes), name it something like vm1 and make a copy of it and call the second directory vm2. If you are using DHCP you can copy it AFTER you make the configuration changes in step 8.

4. (Optional) for management purposes, you may want to change the name of the VM. Right click on the .vmx file and open it with wordpad. Where it says displayName = "Folding@Home", you can change what is in the quotes to whatever you want to help you keep track of which is which if you have a quadcore and are running 2 VMs. You need to do this for both directories.

5. You can now run the virtual machine by double clicking the .vmx file or by opening it through VMWare Player. (My virtual machines randomly stopped working by opening the .vmx file and I now have to run VMWare Player as Administrator, then open the virtual machine) At this point it will boot up and detect your architechture and launch the appropriate configuration. Assuming it has access to the internet¹, it will download what it needs and start folding (under notfred's username).

6. When the VM boots up, take note of the IP address given on the console. Now go to your web browser and go to http://<IP address of VM>. It will bring up the status page. Click "Reconfigure this host and any USB drives". This page is where you will configure your VM.

7. Configure it however you'd like. Remember to change the username and team! Here are some of the options that might not be obvious.

SMP Instance per - This is how many CPUs to use. I don't think there's any point in using more than 2, as VMWare Player won't use them.
Remote Reboot - This enables a link on the status page that will let you reboot the machine.
Install to USB drives on boot - When the VM boots it will try to install to a USB drive for use on another computer or on the same computer for diskless folding. I've never messed with this, so I don't know if it works.
Console blank time - Simply blanks the VM's window.
Windows Networking - This will enable samba file sharing on the VM, so you can browse its files from the computer it's running on.
Login shell - This enables you to login from the VM to run commands. It's very limited and you probably won't need it, but it doesn't hurt to enable it.

8. Click reconfigure. It will tell you it has to reboot. Let it do its thing. After a few seconds it will be back up and folding under your name. You can double check this on the console in the area marked in blue on the picture below.

9. Repeat 4 - 9 if you have another static IP VM to set up. If you used DHCP, now would be the time to make a copy of the first VM and get it up and running.

10. (Optional, but suggested) You may want to set the priority to vmware-vmx.exe to low. It may provide a slightly higher PPD yield from the GPU2 clients, especially if they require more than 1-2% CPU. It will also make it less likely to bog down the rest of your computer. If running GPU2 clients, you should also set their priority to slightly higher via the GPU2 client (not task manager) if you haven't already.

11. (Optional, but suggested) To avoid loss of data, you should set your virtual machines to go into a suspend state when closed, rather than powering down. To do this, go to "VMWare Player" at the top of the VMWare Player window, go to "Preferences..." and change the option of "When exiting:" to "Suspend the virtual machine" if the radio button is not already in that position. Now when you hit the X in the upper right hand corner it will take a minute to to save the machine's state to your hard drive for restoration later. Keep in mind that having them shut off for too long can make you miss the deadline! I have about a 12 hour window for having them shut off on my Q6600.

¹ If it doesn't have access to the internet, try switching between bridged and NAT and rebooting the VM. You can switch between bridged and NAT by right clicking the two computes in the bottom of the VMWare Player window. This is circled in red in the image below. Also make sure your firewall is allowing it out. It will probably ask you if you want to allow it's virtual network cards when you install VMWare Player.

Some visual clarification:

Monitoring with FahMon

1. Download Fahmon, if you don't have it already. You can get 2.3.99+ here.

2. In the main window right click "Add a client". Fill in the name as whatever you'd like to describe it. I find it helpful to use the name I gave it with displayName= in step 4.

3. In the location box you can either use the IP (if it's static) or the client's hostname. To find the host name you can either look at the console, or go to the client's built in webpage. This name will be something like FOLD-XXXX. In the location box enter \\<either the IP, or the hostname>\c\etc\folding\1\

4. Check the box that says "Client is on a Virtual Machine" and click OK.

5. (Optional) You may want to go to FahMon -> Preferences -> Monitoring and check the "Ignore Asynchronous Clocks" option if you are still having problems with it saying it is hung.

Common Problems

1. You get something like this:

Code:

hda1: rw=1, want 2094840, limit 2088387
attempt to access beyond end of device

This seems to happen to everybody. It doesn't seem to cause any real problems. notfred knows about this and may be working on a fix. A few people have said their PPD has gone down after it happened, but any ill effects are not known for sure.

2. You get an out of RAM error. This crashes your folding core and you have to reboot the VM. Unless you move up the RAM, this can happen again.

To increase the amount of RAM, go to VMWare Player on the title bar of the VMWare Player, then go to Troubleshoot, then Change Memory Allocation. At minimum I would use 864, but some people have said it requires more. It has also been reported that being a multiple of 32 helps. Then go to the configuration webpage and reboot the virtual machine. NOTE: this doesn't seem to actually use up more RAM. What it DOES seem to do is increase the size limit of the .vmem file in your folding directory. If your system doesn't have much RAM, don't be afraid to move this up.

3. You lose work when powering off your Virtual Machine.
[INDENT]P
OK, now the information for the guide is here. I just gotta look at other people's guides to get an idea how to pretty it up.

SniperXX · 03-11-09, 02:36 PM

I was just looking to start this on my gpu folding rigs.

Thanks.

Mustanley · 03-11-09, 02:52 PM

Seems easy enough... I think I will go ahead and switch from the winsmp client to this linux/vmware config. I know the points boost will be pretty significant on 3 Quads. Thanks for the guide

Jolly-Swagman · 03-11-09, 04:14 PM

Was just about to do a guide on this too,, Thanks PeddlerOfFlesh,

harlam357 · 03-11-09, 04:19 PM

Alas... I still can't monitor these rigs correctly with FahMon.

It just shows up as *Hung* even though it's still running.

Yes, I have "Ignore Async Clocks" checked.

I need a single app to monitor all my clients at a glance. Checking them all manually is a pain == not an option. I need to find the time (and energy) to build my own monitoring app. FahMon is beginning to be more of a headache than a friendly tool anymore.

If anyone can get FahMon to monitor these notfred virtual appliances correctly, I'm all ears.

Oh, and last but certainly not least...

@Peddler- very nice guide.

PeddlerOfFlesh · 03-11-09, 04:36 PM

Quote:

Originally Posted by harlam357

Alas... I still can't monitor these rigs correctly with FahMon.

It just shows up as *Hung* even though it's still running.

Yes, I have "Ignore Async Clocks" checked.

I need a single app to monitor all my clients at a glance. Checking them all manually is a pain == not an option. I need to find the time (and energy) to build my own monitoring app. FahMon is beginning to be more of a headache than a friendly tool anymore.

If anyone can get FahMon to monitor these notfred virtual appliances correctly, I'm all ears.

Oh, and last but certainly not least...

@Peddler- very nice guide.

Have you tried a SVN build? I used to have problems before I started using SVN builds. I have a build up here if you wanna try it. It adds the option to mark the machine as a virtual machine, which I think is what fixed it. It also adds HTTP monitoring, which works with notfreds.

Also, fahmon is open source. If you can improve it, I'm sure uncle_fungus would appreciated the help.

Jolly-Swagman · 03-11-09, 04:38 PM

Quote:

Originally Posted by harlam357

Alas... I still can't monitor these rigs correctly with FahMon.

It just shows up as *Hung* even though it's still running.

Yes, I have "Ignore Async Clocks" checked.

I need a single app to monitor all my clients at a glance. Checking them all manually is a pain == not an option. I need to find the time (and energy) to build my own monitoring app. FahMon is beginning to be more of a headache than a friendly tool anymore.

If anyone can get FahMon to monitor these notfred virtual appliances correctly, I'm all ears.

Point your notfreds Diskless or your Notfreds VM to EG- \\192.168.XX.XXX\C\etc\folding\1\

Yellow IP to that of the Machine client is set up for and you should be good to go

@PeddlerOfFlesh --- yeah I think that is also the version fahmon 2.3.4 I have too

harlam357 · 03-11-09, 05:32 PM

Quote:

Originally Posted by PeddlerOfFlesh

Have you tried a SVN build? I used to have problems before I started using SVN builds. I have a build up here if you wanna try it. It adds the option to mark the machine as a virtual machine, which I think is what fixed it. It also adds HTTP monitoring, which works with notfreds.

Also, fahmon is open source. If you can improve it, I'm sure uncle_fungus would appreciated the help.

Yes and no... I'm using an older v2.3.3svn build I got from a member of OCAU. He actually posted about it here many months ago and I picked it up then. The distribution v2.3.4 just sucks... so I've just stuck with my older svn version.

I'll definitely try the version you've got linked. If I can mark it as a VM and it be happy, then that will deter my future efforts for a bit... although I've already started work on the log parsing code.

As far as helping improve the current code... I'm just not a C++ guy. I'm a C# guy. Meaning, it's not that I'm C++ illiterate... I do some C++ for work purposes (very limited), but I don't really enjoy it very much- nor would I be happy coding C++ full time. So, if I'm going to take the time to work on a monitoring utility, then I'm going to use the language I enjoy. Blame it on my college education... when I went through school they shoved Java down our throats (~10 years ago)... so I took a liking to working in a managed memory environment.

Quote:

Originally Posted by Jolly-Swagman

Point your notfreds Diskless or your Notfreds VM to EG- \\192.168.XX.XXX\C\etc\folding\1\

Yellow IP to that of the Machine client is set up for and you should be good to go

@PeddlerOfFlesh --- yeah I think that is also the version fahmon 2.3.4 I have too

Well... that's how I have it setup now silly.

It must just work correctly with this newer svn version you guys have. Which is exactly what I'm looking for.

Now we'll just have to see if this version is good at handling 15~20 clients, as the distribution 2.3.4 just chokes on all my clients.

cuda340 · 03-11-09, 07:08 PM

If you have not, give Fahspy a go. Works here for monitoring diskless, vm & gpu folding. With Fahmon depending on the time of day i check vm's will show hung when they have not and yes, I have "Ignore Async Clocks" checked also.

HayesK · 03-11-09, 08:52 PM

Quote:

Originally Posted by harlam357

Now we'll just have to see if this version is good at handling 15~20 clients, as the distribution 2.3.4 just chokes on all my clients.

I have been running 2.3.4 svn rev 478 for about 2 weeks monitoring 40 instances on my home network.

MARCI · 03-12-09, 12:07 AM

I was gonna give this ago, unfortunately the image site seems to be down.

PeddlerOfFlesh · 03-12-09, 02:45 AM

Quote:

Originally Posted by MARCI

I was gonna give this ago, unfortunately the image site seems to be down.

Yes, it does seem to go down occasionally. I think it is just running from a home computer or something. It seems to be back up now.

MARCI · 03-12-09, 03:12 AM

Yeah, already downloaded and gotten it running..

jintatsu · 03-12-09, 08:55 AM

nice guide Peddler..

Edward2 · 03-12-09, 01:21 PM

Quote:

Originally Posted by jintatsu

nice guide Peddler..

I agree. Very nice job and hopefully this will encourage more people to try VM/SMP folding, since it gives more PPD. With MPC on our tails, we need all the points we can get.

the_cultie · 03-12-09, 02:25 PM

When I ever I finish a WU on one of my pc's I'm goin to give this a shot. More PPD is always a good thing

jintatsu · 03-12-09, 03:34 PM

Quote:

Originally Posted by Edward2

I agree. Very nice job and hopefully this will encourage more people to try VM/SMP folding, since it gives more PPD. With MPC on our tails, we need all the points we can get.

Yep hopefully most of us will use vm/smp folding. It would really help us keeping the number 3 spot.

Quote:

Originally Posted by the_cultie

When I ever I finish a WU on one of my pc's I'm goin to give this a shot. More PPD is always a good thing

More PPD is always a good thing..

I will try this also and compare it with my xubuntu VM after I finish the current WU.

Roisen · 03-12-09, 03:55 PM

This needs some glue

Edward2 · 03-12-09, 04:41 PM

Quote:

Originally Posted by jintatsu

I will try this also and compare it with my xubuntu VM after I finish the current WU.

Please do let us know how it compares. I'm curious myself.

PeddlerOfFlesh · 03-12-09, 05:53 PM

Quote:

Originally Posted by Edward2

Please do let us know how it compares. I'm curious myself.

Me too. I don't see how it'd make much of a difference production wise. Just with simplicity setting it up. I suppose not having the other background processes running in the VM could slightly improve performance.

I'm also interested in seeing how much of a production boost you guys are getting from setting them up. Mustanely, have you got yours set up? I'd like to see the difference with your 12 cores.

I wish I could do this on my Q8200

. Stupid lack of VT...

harlam357 · 03-12-09, 07:41 PM

Ok, I got the 2.3.4svn (you most graciously linked) installed today... and I'm happy to report that the notfred VM is being reported accurately.

I will probably setup a couple more notfred VMs now. The only thing I don't like (or haven't found yet) is how to setup the client to do -oneunit. How do I stop the little bugger not in the middle of a WU?

If there is no real way to do so... I think I'll keep a couple of Ubuntu VMs around to facilitate finishing WUs from notfred VMs. I guess I can go to the browser and do a 'Kill' on notfred's VM, then copy the WU to a box where I have full control over the client, finish it and send.

Also, what about the 'on completion' hang bug? I see there is a 'hang check' log file. But when a hang occurs what's the resolution? Again, kill the cores and copy the WU to a box (VM) where I have full control over the client? Then qfix, upload, etc?

It sure is quick to get setup with these virtual appliances... but I am beginning to see a few drawbacks.

PeddlerOfFlesh · 03-12-09, 07:54 PM

Quote:

Originally Posted by harlam357

Ok, I got the 2.3.4svn (you most graciously linked) installed today... and I'm happy to report that the notfred VM is being reported accurately.

I will probably setup a couple more notfred VMs now. The only thing I don't like (or haven't found yet) is how to setup the client to do -oneunit. How do I stop the little bugger not in the middle of a WU?

If there is no real way to do so... I think I'll keep a couple of Ubuntu VMs around to facilitate finishing WUs from notfred VMs. I guess I can go to the browser and do a 'Kill' on notfred's VM, then copy the WU to a box where I have full control over the client, finish it and send.

Also, what about the 'on completion' hang bug? I see there is a 'hang check' log file. But when a hang occurs what's the resolution? Again, kill the cores and copy the WU to a box (VM) where I have full control over the client? Then qfix, upload, etc?

It sure is quick to get setup with these virtual appliances... but I am beginning to see a few drawbacks.

You bring up good points. I haven't tried to do the -oneunit thing. That'd be nice. When I upgraded them last time, I just waited until the WU was finished. That'd be really annoying with a bunch of clients though.

I've never had the hanging on completion thing happen to me, so I'm not sure about that either. I believe in recent versions he made the file system writable via the windows networking. You MIGHT be able to copy qfix over like that and then run it from the login shell. Same for adding one unit, perhaps. I would try that, but I'm afraid the unit will finish in the middle of the night and I'll lose a bunch of folding time...

cw823 · 03-12-09, 08:06 PM

Quote:

Originally Posted by Roisen

This needs some glue

I copied it into the super duper folding@home install/setup thread earlier today.

harlam357 · 03-13-09, 09:52 AM

Quote:

Originally Posted by PeddlerOfFlesh

You bring up good points. I haven't tried to do the -oneunit thing. That'd be nice. When I upgraded them last time, I just waited until the WU was finished. That'd be really annoying with a bunch of clients though.

I've never had the hanging on completion thing happen to me, so I'm not sure about that either. I believe in recent versions he made the file system writable via the windows networking. You MIGHT be able to copy qfix over like that and then run it from the login shell. Same for adding one unit, perhaps. I would try that, but I'm afraid the unit will finish in the middle of the night and I'll lose a bunch of folding time...

I forgot about the login shell... it's new, at least since I last ran notfred's USB stick folding (which is pretty much the same thing as the virtual appliance, just packaged differently). Can one stop and restart the client cleanly from the logic shell?

What would be nice... is in the config web page, there was a place to add client arguments. So I could add the -oneunit flag, then restart the VM and let it re-fire the client with the flag. That would solve that problem. The hang on completion would still need addressed, but for the most part it's not a big issue... since it happens pretty rarely.

Thanks again for your help & the great instructions Peddler!

cw823 · 03-13-09, 10:11 AM

With our folding mentor program, we should probably have a list of people that are familiar with each setup, like a list of people that know GPU2, ubuntu VMs, etc...

harlam357 · 03-13-09, 10:31 AM

That's a fantastic idea!

Mustanley · 03-13-09, 02:48 PM

Quote:

Originally Posted by PeddlerOfFlesh

Mustanley, have you got yours set up? I'd like to see the difference with your 12 cores.

I wish I could do this on my Q8200

. Stupid lack of VT...

That's a project for this weekend. I'll be sure to get some before and after PPD totals from these systems and post them up.

the_cultie · 03-13-09, 04:22 PM

Just got this installed on my main rig, will probably install it on my second rig when it finish's it current WU. Interested in how it will preform.

PeddlerOfFlesh · 03-13-09, 06:33 PM

Quote:

Originally Posted by harlam357

I forgot about the login shell... it's new, at least since I last ran notfred's USB stick folding (which is pretty much the same thing as the virtual appliance, just packaged differently). Can one stop and restart the client cleanly from the logic shell?

What would be nice... is in the config web page, there was a place to add client arguments. So I could add the -oneunit flag, then restart the VM and let it re-fire the client with the flag. That would solve that problem. The hang on completion would still need addressed, but for the most part it's not a big issue... since it happens pretty rarely.

Thanks again for your help & the great instructions Peddler!

I'm not sure if you can do that from the login shell. I'm thinking probably not, though. You could probably kill -9, but that probably isn't clean. The way it starts up before the login shell makes me think it wouldn't work.

I agree that having a check box or something on the config page would be awesome. I assumed there were complications in getting that going. Not sure why else he wouldn't include them. I wish he had like a trac for it or something, where we could see what people have requested, known bugs, etc. I think with some scripting even a qfix button could be put into the page. I haven't really looked at the guts on this yet. For me it was just set it and forget it.

Let's not forget to thank notfred (even though he's not here, I don't think). Without his work building these, I wouldn't even be able to make this guide.

Mustanley · 03-14-09, 04:45 AM

So I converted one machine over to the notfred VM image. Boosted my PPD from SMP from 2600PPD to 5200PPD. Problem is my GTX 285 seems to be folding at 1/2 speed, generating only 4200PPD.

I tried changing the priority of the vmware-vmx.exe processes to "Low", bumped the FahCore_14.exe process of the GPU2 client to normal, but my GPU2 folding times are unchanged. Anybody know what I'm missing here?

*EDIT* If I prevent vmware from utilizing core 3, then the GPU2 client returns to normal production. How do I allow both VMs to utilize all four cores and still get normal GPU2 performance?

03-11-09, 02:27 PM	#1
PeddlerOfFlesh Member Join Date: Jan 2009	Notfred's VM Appliance + VMWare Player Guide OK, this is my first attempt at a guide, so it may be missing stuff and I'm certain it will be ugly (at least to start). There seems to have been a lot of interest in this, but a lot of questions. Here goes. How to set up Notfred's VMWare Virtual Appliance with VMWare Player. 1. Obtain and install VMWare Player. (VMWare) (Filehippo) Filehippo doesn't require registration. I think the only option during install is the install directory. Other than that, any good firewall will probably pop up asking if you want to allow the newly added virtual network cards to have access to your computer or LAN. Make sure you allow these. 2. Obtain the virtual appliance image. (link) 3. Open the zip containing folding.vmdk and folding.vmx. Extract these to whatever directories you want to use, say C:\folding\vm for example. If you have a quadcore and are planning to use 2 VMs with static IP addresses (suggested, for monitoring purposes), name it something like vm1 and make a copy of it and call the second directory vm2. If you are using DHCP you can copy it AFTER you make the configuration changes in step 8. 4. (Optional) for management purposes, you may want to change the name of the VM. Right click on the .vmx file and open it with wordpad. Where it says displayName = "Folding@Home", you can change what is in the quotes to whatever you want to help you keep track of which is which if you have a quadcore and are running 2 VMs. You need to do this for both directories. 5. You can now run the virtual machine by double clicking the .vmx file or by opening it through VMWare Player. (My virtual machines randomly stopped working by opening the .vmx file and I now have to run VMWare Player as Administrator, then open the virtual machine) At this point it will boot up and detect your architechture and launch the appropriate configuration. Assuming it has access to the internet¹, it will download what it needs and start folding (under notfred's username). 6. When the VM boots up, take note of the IP address given on the console. Now go to your web browser and go to http://<IP address of VM>. It will bring up the status page. Click "Reconfigure this host and any USB drives". This page is where you will configure your VM. 7. Configure it however you'd like. Remember to change the username and team! Here are some of the options that might not be obvious. SMP Instance per - This is how many CPUs to use. I don't think there's any point in using more than 2, as VMWare Player won't use them. Remote Reboot - This enables a link on the status page that will let you reboot the machine. Install to USB drives on boot - When the VM boots it will try to install to a USB drive for use on another computer or on the same computer for diskless folding. I've never messed with this, so I don't know if it works. Console blank time - Simply blanks the VM's window. Windows Networking - This will enable samba file sharing on the VM, so you can browse its files from the computer it's running on. Login shell - This enables you to login from the VM to run commands. It's very limited and you probably won't need it, but it doesn't hurt to enable it. 8. Click reconfigure. It will tell you it has to reboot. Let it do its thing. After a few seconds it will be back up and folding under your name. You can double check this on the console in the area marked in blue on the picture below. 9. Repeat 4 - 9 if you have another static IP VM to set up. If you used DHCP, now would be the time to make a copy of the first VM and get it up and running. 10. (Optional, but suggested) You may want to set the priority to vmware-vmx.exe to low. It may provide a slightly higher PPD yield from the GPU2 clients, especially if they require more than 1-2% CPU. It will also make it less likely to bog down the rest of your computer. If running GPU2 clients, you should also set their priority to slightly higher via the GPU2 client (not task manager) if you haven't already. 11. (Optional, but suggested) To avoid loss of data, you should set your virtual machines to go into a suspend state when closed, rather than powering down. To do this, go to "VMWare Player" at the top of the VMWare Player window, go to "Preferences..." and change the option of "When exiting:" to "Suspend the virtual machine" if the radio button is not already in that position. Now when you hit the X in the upper right hand corner it will take a minute to to save the machine's state to your hard drive for restoration later. Keep in mind that having them shut off for too long can make you miss the deadline! I have about a 12 hour window for having them shut off on my Q6600. ¹ If it doesn't have access to the internet, try switching between bridged and NAT and rebooting the VM. You can switch between bridged and NAT by right clicking the two computes in the bottom of the VMWare Player window. This is circled in red in the image below. Also make sure your firewall is allowing it out. It will probably ask you if you want to allow it's virtual network cards when you install VMWare Player. Some visual clarification: Monitoring with FahMon 1. Download Fahmon, if you don't have it already. You can get 2.3.99+ here. 2. In the main window right click "Add a client". Fill in the name as whatever you'd like to describe it. I find it helpful to use the name I gave it with displayName= in step 4. 3. In the location box you can either use the IP (if it's static) or the client's hostname. To find the host name you can either look at the console, or go to the client's built in webpage. This name will be something like FOLD-XXXX. In the location box enter \\<either the IP, or the hostname>\c\etc\folding\1\ 4. Check the box that says "Client is on a Virtual Machine" and click OK. 5. (Optional) You may want to go to FahMon -> Preferences -> Monitoring and check the "Ignore Asynchronous Clocks" option if you are still having problems with it saying it is hung. Common Problems 1. You get something like this: Code: hda1: rw=1, want 2094840, limit 2088387 attempt to access beyond end of device This seems to happen to everybody. It doesn't seem to cause any real problems. notfred knows about this and may be working on a fix. A few people have said their PPD has gone down after it happened, but any ill effects are not known for sure. 2. You get an out of RAM error. This crashes your folding core and you have to reboot the VM. Unless you move up the RAM, this can happen again. To increase the amount of RAM, go to VMWare Player on the title bar of the VMWare Player, then go to Troubleshoot, then Change Memory Allocation. At minimum I would use 864, but some people have said it requires more. It has also been reported that being a multiple of 32 helps. Then go to the configuration webpage and reboot the virtual machine. NOTE: this doesn't seem to actually use up more RAM. What it DOES seem to do is increase the size limit of the .vmem file in your folding directory. If your system doesn't have much RAM, don't be afraid to move this up. 3. You lose work when powering off your Virtual Machine. [INDENT]P OK, now the information for the guide is here. I just gotta look at other people's guides to get an idea how to pretty it up. __________________ Folding 24/7 with: Q6600 @ 2.88GHz + PNY 9800GT 1GB w/ shader @ 1728 + PNY 8800GT 512MB w/ shader @ 1728 Q8200 @ 3.16GHz + EVGA 8800GS w/ shader @ 1782 Radeon HD3650 @ stock Last edited by PeddlerOfFlesh; 05-14-09 at 06:53 PM. Reason: Added a step.

03-11-09, 02:36 PM	#2
SniperXX Member Join Date: Jul 2002 Location: Folding in Diamond Bar	I was just looking to start this on my gpu folding rigs. Thanks. __________________ i7 920 @ 4.2Ghz Asus Rampage II GENE 2x BFG GTX285 6GB Crucial Dominator 1600 Crucial 750TX Help find cures for cancer by running Folding @ Home. HEATWARE

03-11-09, 02:52 PM	#3
Mustanley Team 32 Folding Member Join Date: May 2003 Location: Asheville, NC (Blue Ridge Mtns)	Seems easy enough... I think I will go ahead and switch from the winsmp client to this linux/vmware config. I know the points boost will be pretty significant on 3 Quads. Thanks for the guide __________________ Q9550 on P5Q-Deluxe(460x8.5=3910MHz), 2x2GB GSkill 1066, 750GB WD Black HD, GTX285(702,1512,2592) Philips 16x DVD+RW D/L, Silverstone OP650, Samsung 1080P 37" LCD + Dell 2408WFP, MX1000 laser mouse, Saitek backlit keyboard, P182, Xigmatek S1283 Black Knight Q6600 on Asus P5Q-EM HDMI (400x9=3600MHz - 1.4V), 2X2GB Gskill 1066 , LG 6X Blueray, 500GB + 640GB WD HDs, XFX GTX-260 (192sp) , Antec Earthwatts 430, Samsung 1080P 50" Plasma, Antec 300 case, Xigmatek HDT-S1283 Q6600 on P5K Premium (400MHzx8), 2x2gb Kingston HyperX 1066, BFG 8800GTS-512, Antec Earthwatts 430, Dell 2408WFP, Xigmatek S1283 Asus eee PC 901 FOLDING slave of Team32 Please Fold!

03-11-09, 04:14 PM	#4
Jolly-Swagman Team 32 Folding Member Join Date: Sep 2007 Location: Folding@Home in Downunder Australia.Victoria	Was just about to do a guide on this too,, Thanks PeddlerOfFlesh, __________________ MyHeat I'm a Team 32 Folding Mentor - Ask me about Folding My Ultimate QUAD CORE PC CASE-CM690\|Q6600@3.2GHz(8x400)\|GAP35DS3L\|DDR2-4GigXMS2Pro6400/800-4-4-4-12\|Palit9800GTsonic-1GB\|HX520W\|DVDRWLiteonLS\|Sata160+250+320Gig\|OCZVen detta2+HR05NB\|BenQE2200HDWinXPProSP3(BetaTesting)Windows7-Ultimate-64bit <- Folding@Home-2xVM-SMP+GPUv2 My CORE 2 DUO PC E6550@3.0Ghz\|GAG31MS2L\|4GigXMS2Pro6400C5-800-5-5-5-18\|ION2-550W/\|Inno3D8800GT512MB+HR03GT\|ACFrezzer7Pro + NB-HSF\|Viewsonic-VA703B x2 OS's-Ubuntu8.04/WinXPPro<-Rosetta/Folding@Home-VM-SMP+GPUv2->Folding for my Parents for a Cure! The Milestone Thread 4.0 \| Smooth-n-Fold-Custom Project \|The Sled Acrylic Tech Bench

03-11-09, 04:19 PM	#5
harlam357 Senior Fold-a-holic Join Date: Sep 2004 Location: @Home... Folding@Home	Alas... I still can't monitor these rigs correctly with FahMon. It just shows up as Hung even though it's still running. Yes, I have "Ignore Async Clocks" checked. I need a single app to monitor all my clients at a glance. Checking them all manually is a pain == not an option. I need to find the time (and energy) to build my own monitoring app. FahMon is beginning to be more of a headache than a friendly tool anymore. If anyone can get FahMon to monitor these notfred virtual appliances correctly, I'm all ears. Oh, and last but certainly not least... @Peddler- very nice guide. __________________ Q9550(C1)@(450x8) 3.6GHz 1.28v \| Gigabyte EP45-UD3P v1.0 F8 bios \| 2x2gb G.Skill@DDR900 5-5-5-15 2T 2.05v \| EVGA GTX275 \| Antec P180 - Corsair 650TX \| WD740GD - WD3200AAKS \| Samsung 2433BW FOLD32: 2xGTX285@1548MHz, GTX275@1566MHz, 2xGTX260-216@1458MHz, Yorkfield Q9550@3.6GHz, 3xKentsfield Q6600@3.4~3.0GHz, Wolfdale E8400@3.0GHz HFM.NET - Client Monitoring Application for Folding@Home \| I'm a Team 32 Folding Mentor - Ask me about Folding \| Team 32 Folding@Home Intro \| Team 32 F@H FAQ \| Folding/Gaming Room Project \| H's HTPC Build \| Heat (39-0-0)

03-11-09, 07:08 PM	#9
cuda340 Member Join Date: May 2004 Location: Ontario, Canada	If you have not, give Fahspy a go. Works here for monitoring diskless, vm & gpu folding. With Fahmon depending on the time of day i check vm's will show hung when they have not and yes, I have "Ignore Async Clocks" checked also.

03-12-09, 12:07 AM	#11
MARCI Team 32 Folding Member Join Date: Jun 2008 Location: Folding for T32 in Atlanta	I was gonna give this ago, unfortunately the image site seems to be down. __________________ My Heat Daily Driver Rocketfish C2D E8400 @3.6 1.184vCore GIGABYTE EP45-DS3L 4 GB G.SKILL PC8500 MSI GTX260 (216) OC 896MB folding 24/7 CORSAIR CMPSU-1000HX MotherGoose: The answer is: you can't. You're essentially a dimension short. Do what thou wilt shall be the whole of the law

03-12-09, 03:12 AM	#13
MARCI Team 32 Folding Member Join Date: Jun 2008 Location: Folding for T32 in Atlanta	Yeah, already downloaded and gotten it running.. __________________ My Heat Daily Driver Rocketfish C2D E8400 @3.6 1.184vCore GIGABYTE EP45-DS3L 4 GB G.SKILL PC8500 MSI GTX260 (216) OC 896MB folding 24/7 CORSAIR CMPSU-1000HX MotherGoose: The answer is: you can't. You're essentially a dimension short. Do what thou wilt shall be the whole of the law

03-12-09, 08:55 AM	#14
jintatsu Team 32 Folding Member Join Date: Oct 2007 Location: Folding@Philippines	nice guide Peddler.. __________________ Intel C2Q Q9550 @ 3.82Ghz (1.3v) \| Thermalright Ultra 120 Extreme \| GSkill 2x2Gb ddr2-1066 Gigabyte GA-P35-DS3P v1.1 \| Thermalright HR-05 IFX \| MSI 8800GTS 512MB Corsair HX620 \| Cooler Master CM690 \| Creative Sound Blaster X-Fi Xtrememusic Fold for [Team 32] Use your idle computer to fold

03-12-09, 02:25 PM	#16
the_cultie Team 32 Folding Member Join Date: Feb 2005 Location: Folding in Co. Louth, Ireland and designing my folding reactor	When I ever I finish a WU on one of my pc's I'm goin to give this a shot. More PPD is always a good thing __________________ AMD Opteron 180 @ 2.784GHz (232x12),, DFI LanParty Ultra-D,, G.Skill 2GB (2x1024MB) DDR500 @ 225MHz,, XFX 8800GTS 512MB,, Corsair TX650W,, MCP655->MCR220->Apogee GT->MCW60->MicroRES,, 2x WD CaviarSE16 500GB--RAID0,, NEC DVD±RW,, Modded Jeantech Light Fusion Folding for Team32 Friends don't let friends quit Folding -ihrsetrdr TEAM 32 NEEDS YOUR HELP!!!! Join the rally!! Help us get back in the top 3!

03-12-09, 03:55 PM	#18
Roisen Team 32 Folding Member Join Date: Feb 2007 Location: Folding in Ames, IA	This needs some glue __________________ Folding for team 32:__ /¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯ __________________________/Heimdall\ \|E6400 @ 3.4Ghz \|\| 3GB 500mhz 5-4-4-12 \| \|2x 9600GT SLI c725, s1900, m1000 (riva) \| ¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯ Looking for a Q6xxx processor to adopt to replace my dual-core. PM me to make my dreams come true.

03-12-09, 07:41 PM	#21
harlam357 Senior Fold-a-holic Join Date: Sep 2004 Location: @Home... Folding@Home	Ok, I got the 2.3.4svn (you most graciously linked) installed today... and I'm happy to report that the notfred VM is being reported accurately. I will probably setup a couple more notfred VMs now. The only thing I don't like (or haven't found yet) is how to setup the client to do -oneunit. How do I stop the little bugger not in the middle of a WU? If there is no real way to do so... I think I'll keep a couple of Ubuntu VMs around to facilitate finishing WUs from notfred VMs. I guess I can go to the browser and do a 'Kill' on notfred's VM, then copy the WU to a box where I have full control over the client, finish it and send. Also, what about the 'on completion' hang bug? I see there is a 'hang check' log file. But when a hang occurs what's the resolution? Again, kill the cores and copy the WU to a box (VM) where I have full control over the client? Then qfix, upload, etc? It sure is quick to get setup with these virtual appliances... but I am beginning to see a few drawbacks. __________________ Q9550(C1)@(450x8) 3.6GHz 1.28v \| Gigabyte EP45-UD3P v1.0 F8 bios \| 2x2gb G.Skill@DDR900 5-5-5-15 2T 2.05v \| EVGA GTX275 \| Antec P180 - Corsair 650TX \| WD740GD - WD3200AAKS \| Samsung 2433BW FOLD32: 2xGTX285@1548MHz, GTX275@1566MHz, 2xGTX260-216@1458MHz, Yorkfield Q9550@3.6GHz, 3xKentsfield Q6600@3.4~3.0GHz, Wolfdale E8400@3.0GHz HFM.NET - Client Monitoring Application for Folding@Home \| I'm a Team 32 Folding Mentor - Ask me about Folding \| Team 32 Folding@Home Intro \| Team 32 F@H FAQ \| Folding/Gaming Room Project \| H's HTPC Build \| Heat (39-0-0)

03-13-09, 10:11 AM	#25
cw823 Short-Break Member Join Date: Dec 2000 Location: Lima, OH	With our folding mentor program, we should probably have a list of people that are familiar with each setup, like a list of people that know GPU2, ubuntu VMs, etc... __________________ taking suggestions for sig

03-13-09, 10:31 AM	#26
harlam357 Senior Fold-a-holic Join Date: Sep 2004 Location: @Home... Folding@Home	That's a fantastic idea! __________________ Q9550(C1)@(450x8) 3.6GHz 1.28v \| Gigabyte EP45-UD3P v1.0 F8 bios \| 2x2gb G.Skill@DDR900 5-5-5-15 2T 2.05v \| EVGA GTX275 \| Antec P180 - Corsair 650TX \| WD740GD - WD3200AAKS \| Samsung 2433BW FOLD32: 2xGTX285@1548MHz, GTX275@1566MHz, 2xGTX260-216@1458MHz, Yorkfield Q9550@3.6GHz, 3xKentsfield Q6600@3.4~3.0GHz, Wolfdale E8400@3.0GHz HFM.NET - Client Monitoring Application for Folding@Home \| I'm a Team 32 Folding Mentor - Ask me about Folding \| Team 32 Folding@Home Intro \| Team 32 F@H FAQ \| Folding/Gaming Room Project \| H's HTPC Build \| Heat (39-0-0)

03-13-09, 04:22 PM	#28
the_cultie Team 32 Folding Member Join Date: Feb 2005 Location: Folding in Co. Louth, Ireland and designing my folding reactor	Just got this installed on my main rig, will probably install it on my second rig when it finish's it current WU. Interested in how it will preform. __________________ AMD Opteron 180 @ 2.784GHz (232x12),, DFI LanParty Ultra-D,, G.Skill 2GB (2x1024MB) DDR500 @ 225MHz,, XFX 8800GTS 512MB,, Corsair TX650W,, MCP655->MCR220->Apogee GT->MCW60->MicroRES,, 2x WD CaviarSE16 500GB--RAID0,, NEC DVD±RW,, Modded Jeantech Light Fusion Folding for Team32 Friends don't let friends quit Folding -ihrsetrdr TEAM 32 NEEDS YOUR HELP!!!! Join the rally!! Help us get back in the top 3!

03-14-09, 04:45 AM	#30
Mustanley Team 32 Folding Member Join Date: May 2003 Location: Asheville, NC (Blue Ridge Mtns)	So I converted one machine over to the notfred VM image. Boosted my PPD from SMP from 2600PPD to 5200PPD. Problem is my GTX 285 seems to be folding at 1/2 speed, generating only 4200PPD. I tried changing the priority of the vmware-vmx.exe processes to "Low", bumped the FahCore_14.exe process of the GPU2 client to normal, but my GPU2 folding times are unchanged. Anybody know what I'm missing here? EDIT If I prevent vmware from utilizing core 3, then the GPU2 client returns to normal production. How do I allow both VMs to utilize all four cores and still get normal GPU2 performance? __________________ Q9550 on P5Q-Deluxe(460x8.5=3910MHz), 2x2GB GSkill 1066, 750GB WD Black HD, GTX285(702,1512,2592) Philips 16x DVD+RW D/L, Silverstone OP650, Samsung 1080P 37" LCD + Dell 2408WFP, MX1000 laser mouse, Saitek backlit keyboard, P182, Xigmatek S1283 Black Knight Q6600 on Asus P5Q-EM HDMI (400x9=3600MHz - 1.4V), 2X2GB Gskill 1066 , LG 6X Blueray, 500GB + 640GB WD HDs, XFX GTX-260 (192sp) , Antec Earthwatts 430, Samsung 1080P 50" Plasma, Antec 300 case, Xigmatek HDT-S1283 Q6600 on P5K Premium (400MHzx8), 2x2gb Kingston HyperX 1066, BFG 8800GTS-512, Antec Earthwatts 430, Dell 2408WFP, Xigmatek S1283 Asus eee PC 901 FOLDING slave of Team32 Please Fold! Last edited by Mustanley; 03-14-09 at 04:52 AM.