Here is an up-to-date VB image:

Virtualbox Folding Image

You can use the VB SATA controller with the image for optimal performance.

I posted this in my thread in foldingforum, but thought I would post here as well in case anyone is interested. I have been working on a new folding appliance that uses the PXE functionality of the network controller to download a bootable folding image off a centralized server via TFTP. Folding data is stored via NFS shares on a server. This makes it possible to run diskless folding nodes. Also, the image supports serial console to run headless as well.

PXE Boot Image Linux  

It may still have some bugs though since I barely put this together over the last couple weeks but I have two diskless and soon to be headless systems both running A2 bigadv at the moment. One has submitted three bigadv work units so far. The Linux server setup at the moment requires some Linux knowledge to setup the necessary services but I have provided sample configs.

TheCrazyCanuck

More info, in particular what is PXE?

PXE makes it possible for a network controller to download and boot a bootloader from a TFTP server. The controller receives an IP from a DHCP server. The DHCP server also hands out the TFTP server address and location of the bootloader in the TFTP directory structure. The bootloader configuration file has the location of the kernel and system init image. With the combination of DHCP and TFTP, the system is able to boot up an operating system that is located on a remote server.

I find it to be a pain to setup a new disk and OS for every system I add. My systems are open-air and there really is not much room for a hard disk which is why I was using CF cards before. The CF cards wear out over time though. With this, I can roll out diskless systems with enabling PXE in BIOS and adjusting boot order.

kougar

Can VMware Workstation 7.1 be "modified" to use more than eight cores? Given the history of VMware software to be unlocked for more cores than it officially allows I'm curious if anyone has tried yet? I've forgotten most of the ini file edits for past modifications, but I'll give it a shot anyway.

 

I have not tried it but the option in the VMX file should be this:

numvcpus = "8"

kougar 
Regarding VirtualBox, the last time I tried the program the TPF increased by roughly 6-7 minutes for a Core i7 920, versus what I was getting with VMware. If they have fixed the performance issues it would be great to hear!

I saw this to the last time I tried folding in VB, but that was quite a while ago. A new version was just recently released since 3.2. The new one is 3.2.4.

Updated or not, even with the new 3.2.4 version of Virtualbox system responsiveness was just as bad as it was 6 months ago. CPU usage would ping-pong erratically between 80% and 100% across all cores, and even typing text into any program would experience lag. I had this problem before with Virtualbox, but never with VMware.

Not sure if it was a configuration issue on my part, but using SMP 12 Virtualbox could not fold a TPF in less than 1 hour for an A2 work unit. I've had better results when testing a Core i7 920 using Virtualbox... is anyone else using VB on a 980X and having issues with it?

TheSynthisterOne

Hey guys, first post here. I've folded plenty of 1920 WU's using -smp7 configurations. I recently picked up more ram and a new mobo. I finished my first -smp 8 WU yesterday(took about 4 days). Fahmon showed my credit for the WU around 25k. Everything great so far, however my new WU is only using 50% of my ram and the ppd dropped. Anyone know why the dip in ram and PPD?

Thanks in advanced,
Justin

if its taking more then 3days 20hrs be better off doing norm A3 as you can do 4 of them an day 
@3.8ghz aprox 
A3 Norm 16K PPD @ 6hrs per project
A3 bigadv is 20K PPD @ 3.2 days
A2 25k PPD @ 2.2 days (but these are stopping soon)

its the new A3 core its 20-30% slower when doing the same work, you need to get your CPU to 3.5 or higher (yours is at 3.7) that should be taking 48-50mins per frame, also do not exit A3 bigadv core (norm core is OK) at any time as it has an Very high chance of trashing the checkpoint (and backups do not work), the New A3 cores only need about 2gb of ram 

use HFM.net ( http://code.google.com/p/hfm-net/ ) as fahmon seems to have stopped developing it and still has the problem with it self (config gets messed up)

TheSynthisterOne

Sorry that I didn't mention that it wasn't folding 24/7. I let the PC Fold about half the day. I was getting around 39 minutes per frame, I think... I can't recall. How would I terminate the unit I'm folding now and get a new one? I don't know why anythings changed. It was working fine but after fetching the new unit its using half the ram when it was maxed on the first unit. Its still us

leexgx

TheSynthisterOne

Hey guys, first post here. I've folded plenty of 1920 WU's using -smp7 configurations. I recently picked up more ram and a new mobo. I finished my first -smp 8 WU yesterday(took about 4 days). Fahmon showed my credit for the WU around 25k. Everything great so far, however my new WU is only using 50% of my ram and the ppd dropped. Anyone know why the dip in ram and PPD?

Thanks in advanced,
Justin

if its taking more then 3days 20hrs be better off doing norm A3 as you can do 4 of them an day 
@3.8ghz aprox 
A3 Norm 16K PPD @ 6hrs per project
A3 bigadv is 20K PPD @ 3.2 days
A2 25k PPD @ 2.2 days (but these are stopping soon)

its the new A3 core its 20-30% slower when doing the same work, you need to get your CPU to 3.5 or higher (yours is at 3.7) that should be taking 48-50mins per frame, also do not exit A3 bigadv core (norm core is OK) at any time as it has an Very high chance of trashing the checkpoint (and backups do not work), the New A3 cores only need about 2gb of ram 

use HFM.net ( http://code.google.com/p/hfm-net/ ) as fahmon seems to have stopped developing it and still has the problem with it self (config gets messed up)

ing 100% of all 8 cpu's. 

A3 bigadv only needs about 2-3gb of ram, A2 Bigadv uses around 4-6gb

you would be better off doing Norm A3 folding as you not make the 4 day deadline to get good points out of Bigadv (due to how much time it takes to do an Bigadv A3 now) A3 norm takes about 6hrs each and is better suited for none 24hr running and works very well under winSMP (no need for VMware)

if you use HFM.net it only uses Deadline for bonus PPD that is 4 days if that date does pass some one els will get your work unit meaning that 2 users will be doing the same work unit so it be wasting project time 

at the moment there is still A2 Bgadv work units left but your not running it 24/7 (dedicated setup or even if you use it like i do) your lucky you even make the final deadline with A3 bigadv work

on that system you be helping stanford if you was Not doing bigadv on that system use the -advmethods option as work unit turn around is 5-6hrs per work unit and you can do it under WinSMP no need for VMware

Hey LR, a friend has a rig that consist of an SR-2, 2X X5677's, and 4 GTX480's. He wants to be able to fold his cards using GPU3 and try to run more than 8 threads on VMWare for bigadv. Can he run 2 instance of VMWare on the same rig since he has enough RAM to support it? Has this been done before? Thanks.

bogmali

Hey LR, a friend has a rig that consist of an SR-2, 2X X5677's, and 4 GTX480's. He wants to be able to fold his cards using GPU3 and try to run more than 8 threads on VMWare for bigadv. Can he run 2 instance of VMWare on the same rig since he has enough RAM to support it? Has this been done before? Thanks.

Sure. You can run multiple instances on the same rig. Just copy the image to separate directories and load them both up in VMware. You could even try and run three of them to cover 24-threads if there is enough RAM.

There are requirements for bonus points. You need to have completed 10 work units. You also need to use a passkey and need to have completed 80% of the work units by the deadline. If you fall below 80%, you will no longer get bonus points. Also, if you plan to run the SMP2 work units, you may be better off using the native Windows client rather than VMware. The performance should be a bit better this way.

Have you been completing the work units by the preferred deadline which is three days for most projects? You can also post on foldingforum.org. Some of the admins/moderators there that can do a search for your recent completed work units to see if something might be wrong.

Also what project # have you been running?

With 6014 in VMware, I normally see around 3:30-3:45 frame times with -smp 8 and 3.8 GHz. Provided you still meet the bonus point requirements and are running the work units without stopping in between, then you should be getting some bonus points for each one.

TheSynthisterOne

I switched to -advanmethods and now I'm only getting 486 credit per WU. With a PPD of 1779. I'm not going to let my i7 run at 82* C hours on end for less points then I get out of my gtx 260. I got a big boy computer so I want to get big boy points. I may have finished one unit a tad bit late, but that shouldn't mean that Stanford wont send me the right units. I was utilizing all my ram and CPU, scoring tons of points. Now it reads that the core is Unknown. I got the FAH image from the first post in this forum. Is there anything more up to date? I'd like to start getting some bonus even if it means more dedication to folding. Minimum benchmarks on the frames are 3mn 55s. 

Use HFM.NET http://code.google.com/p/hfm-net/ and enable bonus option it Tell you your PPD per date (FAHmon is no longer supported as its not been maintained and does not support kFactor)

Ten come back to me with the points per day (its should be about 3500 With Kfactor enabled (yours is off on HFM.net) per unit 14-17k per day aprox depending on system speed and project norm take about 5-6hrs to do each)  

if you do not meat the bonus dead line that is Classed as an not within the deadline and will result in not counting against the 80% rule (you still get points but Only with A3 big 8k as they have lowered it form A2 big 25k base points but put the kFactor up to 26.4)

...............................
tested your not using HFM.net or you not ticked the Calculate Bonus Credit and PPD option in HFM.net (remember to save the config and then use auto load and safe so HFM.net loads/save your settings), i am getting th dredded 6701 work units at the moment and they do suck as they are only giving me 11-12K PPD (6K 12 hrs ish)but the norm ones are around 15-16k PPD (3.5k or more but 5hrs )

update 2
and your last post Back to normal will only last for an short time as your Working on an A2 big project again, they be running out soon

TheSynthisterOne

I installed HFM.NET. Easy and sweet program. Right now I'm running an A3 Unit, project P2684. The credit for the unit is at 61,015. My PPD is 17,883 on the CPU alone. This is really great. Now I'm the fastest folder on my team! Hopefully I'll get some of the big A2 units before they end. Anyone have an idea if were talking days, weeks or months before we run out?

I am not sure on the timeline for A2's to reach end-of-life, but I still have been getting A2 work units via my Linux systems. I have one running 2681 currently.