Important: 10.7.4 and Sandy Bridge CPU PM Temporary Fix

[SStreet]

Re: IMPORTANT: 10.7.4 and Unlocked Sandy Bridge CPUs [TEMP FIX]

OK, so the issue may be all Sandy Bridge i7s, and not just K models. Seems like we have a few reports now of the i7-2600 having the same issue. Could be due to Hyperthreading?

Can't you disable Hyperthreading on your bios and try out?

Tried that, no change in symtoms.

Also noticed this in my boot messages (dmesg):

P-State Stepper Error 18 at step 2 on CPU 0
P-State Stepper Error 18 at step 2 on CPU 1
P-State Stepper Error 18 at step 2 on CPU 4
P-State Stepper Error 18 at step 2 on CPU 6
P-State Stepper Error 18 at step 2 on CPU 5
P-State Stepper Error 18 at step 2 on CPU 7
P-State Stepper Error 18 at step 2 on CPU 3
P-State Stepper Error 18 at step 2 on CPU 2

Same with HT enable or disabled(just less CPUs)

 

[tonymacx86]

Re: IMPORTANT: 10.7.4 and Unlocked Sandy Bridge CPUs [TEMP FIX]

We could be looking at an issue that will be solved when we can improve the code in Chameleon/Chimera. MacMan's already done a ton of changes in Chimera 1.9.2, but there is more work to be done on the Sandy Bridge CPU code.

 

[H8TR]

Re: IMPORTANT: 10.7.4 and Unlocked Sandy Bridge CPUs [TEMP FIX]

It affecting my i5-2500 non K processor as well. 64 bit Geekbench before was 9850, now it's 4800.

 

[Mcfool123]

Re: IMPORTANT: 10.7.4 and Unlocked Sandy Bridge CPUs [TEMP FIX]

I restored and downloaded new chameleon to try and fix boot error but when fixing chameleon the boot to choose disk works but then when selecting lion drive it loads apple screen and then reboots. Trying to just back to 10.7.3 and thenwill move when an update is made cause changing plist no work here.

 

[Mieze]

Re: IMPORTANT: 10.7.4 and Unlocked Sandy Bridge CPUs [TEMP FIX]

OK, so the issue may be all Sandy Bridge i7s, and not just K models. Seems like we have a few reports now of the i7-2600 having the same issue. Could be due to Hyperthreading?

I've got a Celeron G530 which is affected too, although is doesn't support HT. It is incorrectly recognized as i7 by the OS. The posted fix didn't work for me. The only way to get the machine to boot was installing NullCPUPowerManagement.kext.

Mieze

 

[Ali]

Re: Important: 10.7.4 and i5/i7 Sandy Bridge CPUs Temporary Fix

Intel Core i7 i2600K?Gigabyte GA-P67A-UD5-B3

Updated 10.7.3 -> 10.7.4. using combo update.

Geekbench scores
10.7.3.: 12249
10.7.4.: 5789

P-State Stepper Error 18 at step 2 on CPU 0
P-State Stepper Error 18 at step 2 on CPU 1
P-State Stepper Error 18 at step 2 on CPU 4
P-State Stepper Error 18 at step 2 on CPU 3
P-State Stepper Error 18 at step 2 on CPU 2
P-State Stepper Error 18 at step 2 on CPU 6
P-State Stepper Error 18 at step 2 on CPU 7
P-State Stepper Error 18 at step 2 on CPU 5

 

[berniefer]

2500k on DSDT-GA-Z68M-D2H-F5.

after update 10.7.4 computer seemed slower than usual and geekbench scores were cut by half.

Edited the org. chameleon.Boot.plist file and scores on geekbench are back to normal. I did have to use multibeast and use audio rollback to get my audio running again.

Great work and thank you for all your help tonymac

 

[Inspector42]

Re: IMPORTANT: 10.7.4 and Unlocked Sandy Bridge CPUs [TEMP FIX]

OK, so the issue may be all Sandy Bridge i7s, and not just K models. Seems like we have a few reports now of the i7-2600 having the same issue. Could be due to Hyperthreading?

Can't you disable Hyperthreading on your bios and try out?

Tried that, no change in symtoms.

Also noticed this in my boot messages (dmesg):

P-State Stepper Error 18 at step 2 on CPU 0
P-State Stepper Error 18 at step 2 on CPU 1
P-State Stepper Error 18 at step 2 on CPU 4
P-State Stepper Error 18 at step 2 on CPU 6
P-State Stepper Error 18 at step 2 on CPU 5
P-State Stepper Error 18 at step 2 on CPU 7
P-State Stepper Error 18 at step 2 on CPU 3
P-State Stepper Error 18 at step 2 on CPU 2

Same with HT enable or disabled(just less CPUs)

I ran across a discussion about this message a few weeks ago here:
http://revogirl.wordpress.com/2012/04/23/aboutp-state-stepper-error-18-at-step/. It seems there is an issue with defining p-states with greater steps than 100 MHz, although from looking at my current system, I have some bigger gaps between states 16 and 33 and it worked in 10.7.3. So there might be a more sophisticated rule behind p-states and AppleCPUPowerManagement. I have now two systems (see signature). Both are used on a daily basis so I have to find a good time to shut them down for at least half a day to do some testing with 10.7.4.
I also recognized that on both systems only the bulk multiplier for all cores changes according to HWMonitor. Looking at the MSRDumper messages earlier in this thread, it seem that in 10.7.4. there are also individual frequencies assigned to cores, that could be an explanation, why the p-states don't work anymore.

 

[H8TR]

Would it make a difference that I'm using chameleon r1924 erMac trunk instead of Chimera 1.9.2?

 

[Rick5]

Would it make a difference that I'm using chameleon r1924 erMac trunk instead of Chimera 1.9.2?

nope