Physical vs Virtual Performance

[VinceP]

As suggested by many VMware users, we are using Passmark's Performance tests to compare physical vs virtual machines CPU/RAM performances.

As expected, PassMark reports the VM running at about 90% of native.

But there are 2 exceptions:
"CPU - Physics" test which is 115% higher on physical.
"MEMORY - Large RAM" test which is 149% higher on physical.

Can someone please explain why there is such a difference for these 2 tests?

Thanks!

[David (PassMark)]

The large RAM test depends on how much free RAM is available. It might just be a simple case that the physical machine has more free RAM available than the VM.

The physics test runs a physics simulation on the CPU. You can see this in action visually in the Visualized physics test, under the Advanced menu. I don't know why this test (when run without the 3D graphics componet) would be slower on a VM.

[whiz]

I'll write here since it's the same topic, but different tests.
On my host OS (Win 2k8 64bit) I have 50 MB/s in HDD Sequential Read. On some guest OSes I get 170 MB/s (in some it's 30 MB/s) in HDD SR test. Can anyone explain the difference ?