Important: 10.7.4 and Sandy Bridge CPU PM Temporary Fix

[LosChristianos]

How do you guys make an SSDT file? Is there a guide somewhere?

Thanks

I did a little write up a few days ago and forgot to mention it here :? :
viewtopic.php?f=14&t=60363

Take a bow that's a great guide.

 

[racerx90]

I tired out all the SSDTs on this form. I tried them on 10.7.3 before moving to 10.7.4 to make sure it works. I tired (Greggen, Racerx, bassguitarist) ....etc...I changed generate p states = no; generate c states = yes and dropssdt to yes.

Will 12,1 definition I only get 2 P states 16,37. Seems like as if the SSDT is not working at all because you should get two states with dropssdt=yes. Any advice?

Also when I tired defintion Mac Pro 5,1. My system crashed on startup. Any reason for this?

You should use a MacPro3,1 system definition on that motherboard as you'll get 8 P-States and won't have the audio pop issue with the RTL 889 chip (that you will with the iMac12,1.) Although, even with iMac12,1 system definition you should still have gotten 5 P-States (this is what I was able to get using my SSDT.aml that I previously posted.)

When you had MSRDumper.kext loaded, did you put the processor under VARYING loads? It sometimes takes a while and a bit of playing around with the stress on the system, but you should have gotten 5 P-States even with iMac12,1 defined (this is what I was able to get with the same setup during my testing.)

As for the crash with MacPro5,1, it's because of the nehalem appletymcedriver.kext driver for Nehalem ECC RAM support. Just delete that kext from /S/L/E if you want to use that system definition (although, you'll still only get 5 P-States with MacPro5,1 based on my testing.)

Finally, are you OC'ing your chip? If you are, than the SSDT.aml I posted won't work for you. That is meant for stock speed i5-2500k with GA-Z68MA-D2H-B3 motherboard, but I guarantee it works if you have this exact setup.

 

[FishCow]

Hi,

I have the same problem as two people already mentioned.

I followed titsmgee's guide to extract and patch my SSDT.

But it looks completely different. My CPU is a i5-2500K.

DefinitionBlock ("SSDT.aml", "SSDT", 1, "AMICPU", "PROC", 0x00000001)
{
    External (\_PR_.OSC_, IntObj)
    External (\_PR_.PDC_, MethodObj)    // 1 Arguments
    Scope (\_PR)
    {
        Processor (P000, 0x01, 0x00000410, 0x06)
        {
            Method (_PDC, 1, NotSerialized)
            {
                \_PR.PDC (Arg0)
            }
            Method (_OSC, 4, NotSerialized)
            {
                Return (\_PR.OSC)
                Arg1
                Arg2
                Arg3
            }
        }
        Processor (P001, 0x02, 0x00000410, 0x06)
        {
            Method (_PDC, 1, NotSerialized)
            {
                \_PR.PDC (Arg0)
            }
            Method (_OSC, 4, NotSerialized)
            {
                Return (\_PR.OSC)
                Arg1
                Arg2
                Arg3
            }
        }
        Processor (P002, 0x03, 0x00000410, 0x06)
        {
            Method (_PDC, 1, NotSerialized)
            {
                \_PR.PDC (Arg0)
            }
            Method (_OSC, 4, NotSerialized)
            {
                Return (\_PR.OSC)
                Arg1
                Arg2
                Arg3
            }
        }
        Processor (P003, 0x04, 0x00000410, 0x06)
        {
            Method (_PDC, 1, NotSerialized)
            {
                \_PR.PDC (Arg0)
            }
            Method (_OSC, 4, NotSerialized)
            {
                Return (\_PR.OSC)
                Arg1
                Arg2
                Arg3
            }
        }
    }
}

Yeah - this is what my extracted SSDT for 2600k looks like as well.
No Method (_PSS, 0, NotSerialized) defined anywhere.

 

[voongoto]

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

Thanks for the feedback- would be interested in anyone's test results- keep them coming.

I deleted cstates and added ssdt and it works fine for me.

intel i2600k, gygabyte z68 motherboard.

 

[geigi]

Hi,

I have the same problem as two people already mentioned.

I followed titsmgee's guide to extract and patch my SSDT.

But it looks completely different. My CPU is a i5-2500K.

DefinitionBlock ("SSDT.aml", "SSDT", 1, "AMICPU", "PROC", 0x00000001)
{
    External (\_PR_.OSC_, IntObj)
    External (\_PR_.PDC_, MethodObj)    // 1 Arguments
    Scope (\_PR)
    {
        Processor (P000, 0x01, 0x00000410, 0x06)
        {
            Method (_PDC, 1, NotSerialized)
            {
                \_PR.PDC (Arg0)
            }
            Method (_OSC, 4, NotSerialized)
            {
                Return (\_PR.OSC)
                Arg1
                Arg2
                Arg3
            }
        }
        Processor (P001, 0x02, 0x00000410, 0x06)
        {
            Method (_PDC, 1, NotSerialized)
            {
                \_PR.PDC (Arg0)
            }
            Method (_OSC, 4, NotSerialized)
            {
                Return (\_PR.OSC)
                Arg1
                Arg2
                Arg3
            }
        }
        Processor (P002, 0x03, 0x00000410, 0x06)
        {
            Method (_PDC, 1, NotSerialized)
            {
                \_PR.PDC (Arg0)
            }
            Method (_OSC, 4, NotSerialized)
            {
                Return (\_PR.OSC)
                Arg1
                Arg2
                Arg3
            }
        }
        Processor (P003, 0x04, 0x00000410, 0x06)
        {
            Method (_PDC, 1, NotSerialized)
            {
                \_PR.PDC (Arg0)
            }
            Method (_OSC, 4, NotSerialized)
            {
                Return (\_PR.OSC)
                Arg1
                Arg2
                Arg3
            }
        }
    }
}

Yeah - this is what my extracted SSDT for 2600k looks like as well.
No Method (_PSS, 0, NotSerialized) defined anywhere.

I see you also have a P8P67, maybe it's because of the motherboard?

 

[racerx90]

I see you also have a P8P67, maybe it's because of the motherboard?

It sure looks like ASUS didn't supply one for that motherboard. Out of curiosity, have you tried also extracting it using a Linux Live CD in addition to what AIDA64 already gave you? (I don't expect it to be any different, but it's quick to extract so it might be worth a try.)

Have you tried the generic one? It might work for you since there's nothing other than bank definitions in that SSDT anyway.

 

[xxmacmanxx]

Hi,

first of all, I do not have an ASUS board, I have a MSI H67MA-E35 (B3) motherboard.

I didn't try to extract it in Ubuntu, but I'm going to use your SSDT now and see what happens.

 

[steveotron]

Running geekbench wont determine anything about stepping, it will make your CPU run at maximum clock because that is what it is designed to do.

Ah, it looks like my multiplier was stuck on 34 because I had NullCPUPowerManagement on for some reason. Upon taking it off, my multiplier is stuck at 16. I've put NullCPUPowerManagement back on for now, and I'm looking forward to the solution from Tony and everyone else. Thanks in advance.

 

[100khz]

i7 2600k gigabyte z68xp-ud3 board.. updated to 10.7.4

geekbench score got cut in half.. around 5000

after applying the fix to the plist file.. as suggested in this thread..

my geekbench score has doubled to 11502.. 64 bit geekbench test

but i checked just before updating to 10.7.4 yesterday.. my geekbench score was above 12000..

why do you think the geekbench score is not as high as it was with 10.7.3?

any suggestions?

i am running the test when no other application is on.. immediately after i boot the system..

 

[masi0]

MacMan: using your fix did the trick on 2600k, thanks