heavily degraded SMP performance with with OS X host

[Bayard Bell]

I´ve been talking with MichalN and klaus-vb about this on #vbox-dev. Here´s the data so far. I´ve been using a kmk of VBox-4.0.8_OSE as a benchmark, compiling on 64-bit OpenIndiana (Solaris) guests. The first guest, indianapolis, is a 4 vCPU/4GB system. The second guest, bloomington, is a 2vCPU/2GB system. Here are the build times with an OS X host running 10.6.7 Server with all current patches:

first run result on blooomington

real       27:26.03
user       31:39.80
sys        11:24.31

second run results on bloomington

real       37:08.96
user       36:17.43
sys        17:06.48

first run result on indianapolis

real       58:50.34
user     1:29:07.86
sys      1:21:58.30

second run result on indianapolis

real     1:03:52.51
user     1:41:24.83
sys      1:47:03.53

Running on the same hardware with fully updated Ubuntu 11.04, here are the results for the same benchmarks:

first run on bloomington:

real       23:54.55
user       31:10.77
sys         9:59.77

second run:

real       27:18.57
user       34:35.85
sys        14:20.94

first run on indianapolis

real       19:02.80
user       41:20.67
sys        21:43.91

second run:

real       19:00.23
user       42:22.75
sys        24:10.66

[Bayard Bell]

VBox.log for indianapolis running under Ubuntu

[Bayard Bell]

VBox.log for bloomington running under Ubuntu

[michaln]

Can we have a VM for testing please? Is there any particular reason why the actual VM would need to be different, rather than simply changing the configuration to add/remove CPUs and memory?

[Frank Mehnert]

No response, closing. Changing the chipset type from ICH9 to PIIX3 could also fix your performance problems.