Folding Forum

Has anyone gotten this to work inside a VM? Virtualbox 3.0.2 supports up to 32 cores in the guest OS, but our team has had ZERO luck with core stability on the following distros installed inside VB: Ubuntu 9.04, Debian (rev?), Redhat, Slackware, and CentOS.

I have the same poor results with VirtualBox : it's awfully slow (compared to VMWare VMs, even to install or to load Linux), and SMP client isn't able to run on the VB's VMs

You should read this thread for instance : viewtopic.php?f=16&t=10449

Last I heard, Virtualbox didn't work with -smp 4 so it's pretty likely that a future upgrade will be required.

bruce wrote:Last I heard, Virtualbox didn't work with -smp 4 so it's pretty likely that a future upgrade will be required.

I downloaded the latest VirtualBox kit this weekend and tried to run my Ubuntu 8.04 VM with 4 VM CPUs (on my XP 64 running Phenom II x4 810) and had nothing but problems ;

- The VM would hang after about an hour
- The clock in the VM would slow down
- FAH died after about 30 minutes, the first time I tried to run the VM -- I didn't think to note the error msg

toTOW wrote:I have the same poor results with VirtualBox : it's awfully slow (compared to VMWare VMs, even to install or to load Linux), and SMP client isn't able to run on the VB's VMs

You should read this thread for instance : viewtopic.php?f=16&t=10449

Already read it, and the info in there is no longer applicable (at least on the 2 machines I tried).

VB 3.0.2 has corrected the slow SMP issues of VB 3.0.0, and I can fold using Notfred's distro just fine with frame times within 1% of the VMWare setup I was previously using.

It's possible there are still some issues w/ VB preventing bigadv to work properly, but SMP -8 or 2 x SMP -4 work just fine withing VB if using Notfred's VM or CD ISO.

bruce wrote:Last I heard, Virtualbox didn't work with -smp 4 so it's pretty likely that a future upgrade will be required.

Are you talking about a problem with the SMP client crashing? Or are you referring to the number of cores available to the guest OS (i.e. the client). The 2nd has been upgraded, and VB now allows more core access than any other VM setup.

brentpresley wrote:

bruce wrote:Last I heard, Virtualbox didn't work with -smp 4 so it's pretty likely that a future upgrade will be required.

Are you talking about a problem with the SMP client crashing? Or are you referring to the number of cores available to the guest OS (i.e. the client). The 2nd has been upgraded, and VB now allows more core access than any other VM setup.

I was talking about SMP crashing, but that information may also be obsolete since VB seems to be fixing problems rather rapidly.

SMP crashes for me in Ubuntu, but not in Notfred's modified Slackware distro.

From my perspective, it looks like some things have been fixed, but there are still some issues to be resolved.


Sigh, guess I'm going to have to try that eval of VMware vSphere after all (PITA to install).

I tried Ubuntu 9.04 Desktop with the SMP client in VBox 3.0.2 today. But after about 10 minutes, close to the time it would progress 1%, the VM would simply freeze, though the VM window itself was still working. I noticed that the VM's CPU usage in task manager dropped from 88-95% to 25% (I have a quad-core).

But apparently Notfred's folding VM works in VBox? How is the PPD in 4 processors? Roughly double the PPD earned in one VMware with 2 processors?