VirtualBox 4.2.8 r83876 is extremely slow on MacBookPro

[pwagle]

VirtualBox 4.2.8 r83876 running Windows 7 is extremely slow on Dec 2010 MacBookPro running MacOSX 10.8.2, and greatly bogs down the entire system:

Processor 2.66 GHz Intel Core i7 Memory 8 GB 1067 MHz DDR3

The behavior was better with only 4G, being only very unpleasant. Added 4G so I could run vboxes with 2G memory. Now the whole system is immediately unusable. So, I don't think its lack of memory.

This has been reported before, but everything was marked a duplicate of a bug that wasn't written by someone following the thread. I'm following this thread.

I changed the host and guest types since this might be specific.

Included is my VBox.log file.

[pwagle]

VBox log of Win7 boot, shutdown

[Frank Mehnert]

Is this the only VM you run at this time or are you running several VMs in parallel? Also, do any other CPU-hogging applications run in parallel to that VM?

[pwagle]

Only one VM at the time, the CPU monitor wasn't showing anything notable, even from the vbox apps. Just ran it again, and watched: brief spike (2 replots) to 104% of the 4 cpu's (it's supposed to use only 1 cpu), then down to 5% for the first 3/4 of the boot, then 10-15% cpu for the remainder. Disk was writing at about 5-10 MB/s for the first 3/4 of the boot, then dropped into the noise.

Nothing I could see was really competing for disk and cpu.

One person said that giving vbox an external disk sped it up. Haven't tried that yet, haven't had spare disks. I'll try harder if you need that benchmark.

[Frank Mehnert]

Any change with VBox 4.3.2?

[pwagle]

It is faster. I do not know windows well enough to know if its fast now, but its certainly usable now. Thanks!

What was the problem?

By the way, I had trouble getting it installed: the 4.2.28 instance I had running didn't think there was an update, and when I got it anyway from the website, the installer seemed to get confused by the fact that the apple app security system was rejecting its attempts to install kernel modules.

[Frank Mehnert]

Not sure what the problem was, therefore I was asking if there was any change. So nice to hear that there was an improvement and I would like to close this ticket.

Regarding the warning you saw: First, the update notification service was not yet engaged to report 4.3.2 as new available version. This is by intention as there are still some problems which we want to solve. Among these problems is also the signature for the kernel extensions. As long as the kexts are not signed by Oracle, OS X Mavericks will warn about unsigned kexts. We are working on a solution but this warning is not fatal.

[pwagle]

Ah ok.. Thanks!