Hey guys,
Wanted to throw in my experience with this since i did things a little differently, though the results seem similar.
I started with a i5 2500k GA-P67A-UD5-B3:
- running 10.6.7 MBP '11
- identified as MacBookPro8,1 in smbios
- latest chimera
- no NullCPUPM
- geekbenching high 11k
In order to move the system over to the iMac install, i spoofed it to look like an imac 12,1.
I did this simply by changing smbios entries and by modifying the systemversion.plist in /S/L/C to read 10.6.6 and Build 10J4026. I also installed NullCPUPM just in case.
At this point software update kicked in and notified me of an update for my imac to 10.6.7.
So i ran the software update version of the update and all went well.
On reboot everything was working except for sound, which for some reason the latest versions of AppleHDA dont seem to work with the typical patch we used to use, so i went back to 1.9.9 (latest is 2.0.5).
In anycase there were no issues at all. geekbench was hitting 12k and all the software that got updated is working, etc.
Then i thought it would be a good time to get rid of NullCPUPM again. It was working fine on the previous macbook8,1 setup without it. So i removed it and reboot.
No boot problems at all, but... indeed performance dropped to abysmal levels, 4k geekbench, etc.
So it seems there is an issue with the imac12,1 identifier and the latest applintelcpupm on our hacks. This is a real bummer. I am a bit torn weather to go back to macbook8,1 identifier or just use nullcpupm. For now i am using nullcpupm, but i am not sure there is any advantage to running the imac identifier despite it being closer to the real hardware.
Hope we can come up with something to explain this soon. seems it wants to run our hacks at lowest performance setting all the time. On the other hand with nullcpupm or another identifier we get turbo back but probably not speedstep. Could this be any issue of bootloader and needing to generate p/c states?
thanks,
g\