Jump to content

SpeedStepper (now supports Mountain Lion 10.8.3)


flAked
 Share

574 posts in this topic

Recommended Posts

Hi All. published a new update 11с35, AICPUPM version 167.0.0, the patcher does not work

possible think of something more versatile than a patch, for example by editing a dsdt?

I would have done it already, if it were possible. But once the BIOS locks that register you can't change it without a general protection fault. The only other solution is to mod the BIOS itself, so it won't lock the register in the first place.

 

I did a BIOS patch and it seems to work, but it might get overwritten if you change certain BIOS settings. Besides, modding a BIOS is way more tedious and nerve-wrecking without the official tools for it. If I do something wrong, you can't even start your machine.

 

Using a patcher for AICPUPM is way more safer and relaxed and that's what I'm sticking to at the moment.

Just remember to copy NullCPU after any update and see if AICPUPM was changed.

Link to comment
Share on other sites

Studing the Windows/Linux's stepper program could we try to rebuild the SMC_PlatformPlugin plist to get more pstates working?

It should be easy to modify if we knew the format that is being used for the stepper programs. This is of course not published openly, so it would require a lot of work to figure it out.

 

Besides, I believe that 9 P-States are enough and that more states won't give more power-savings or better stepping. If you put the current states on a graph you would see that most of the time only 3-4 states are being used anyways.

Link to comment
Share on other sites

First of all, thank you. Really.

 

...just wondering: why, under sl, msrdumper shows me a good number of p-states while, under lion, just two (16 and 38, with a default 2600K)?

I mean, it's perfectly acceptable... but it's a little bit unexpected.

Link to comment
Share on other sites

This has been covered quite a bit. Under Lion the stepper program changed to only use minimal P-States.

 

It has been submitted to Apple QA, because this also results in hotter iMac's.

 

I heard that this wasn't fixed on the upcoming update, yet.

 

You can get more P-States by using MacBookPro8,3 as the model identifier.

Link to comment
Share on other sites

Hi there flAked, I have a Gigabyte P67A-UD3-B3 + i7 2600k and since the Lion update I've noticed some of the problems mentioned here in my system as well.

My PStates dropped from 9 to 6 (2 normal + 4 turbo), and I finally did what you suggested and copied the stepping data from the MacBookPro8,3 (from Lion) to the iMac12,2 plist and that brought back all the PStates I had with Snow Leopard.

My turbo states are set to 45-42 in the bios and my SSDT is modded to respect those multipliers. If I instead use Chameleon's P&C generators and delete the DropSSDT line I don't get the 45 turbo, but the rest is quite similar. So apparently Chameleon still don't generate the correct P states for Sandy Bridge systems, although it's very close and still useful for those that don't like to tinker with the system.

Before:
MSRDumper PStatesReached: 16 27 42 43 44 45

After:
MSRDumper PStatesReached: 16 26 30 35 40 42 43 44 45 

Chameleon's:
MSRDumper PStatesReached: 16 27 31 36 41 42 43 44

 

Another thing I've noticed is that I haven't seen the line about turbo ratios (AppleIntelCPUPowerManagement: Turbo Ratios 1234) since I upgraded to a Sandy Bridge system. Seems like Lion don't manage my PStates but the Bios does.

Link to comment
Share on other sites

Another thing I've noticed is that I haven't seen the line about turbo ratios (AppleIntelCPUPowerManagement: Turbo Ratios 1234) since I upgraded to a Sandy Bridge system. Seems like Lion don't manage my PStates but the Bios does.

I'm not sure at this point, it could be that they just disabled the output for it. I'll need to start analyzing the other registers in Lion to see if anything changed in that regard.

 

I always had the Turbo Ratio output on my SandyBridge 10.6 system, but not on 10.7.

 

One nice test you could do is booting with NullCPU and loading up MSRDumper and see what that will give you in terms of P-States.

Link to comment
Share on other sites

Hey the patcher works, i've managed to boot without the nullcpu.kext

My problem is that sometimes at the boot my mackintosh has random KP with CPU(0), CPU(1) or CPU(3), but after some hard reset it'll work...

with nocpukext work with no problem.

 

I have no SDDT tables, only DSDT without any CPU definition (just generic scope).

 

GeneratePstate and C state is done by chimera.

 

any suggestion?

Link to comment
Share on other sites

Hi Flaked,

have a trouble to enable speedstep and remove the nullcpupowermanagement from my E\E folder

i use DSDT autopatcher to patch my zotac H67-ITX-C-E dsdt but unfortunately,

i have a kernel panic (intelcpupowermanagement.kext) using the original dsdt patched with dsdt auto patcher

i checked , in HPET, method STA return to 0X0F and the CRS Method return to CRS as mentionned as the HPET fix

for that i can remove the nullcpupowermanagement.kext in the extra\extension folder ( that work fine if i have the nulcpupowermanement.kext)

i use your method to patch the AICPM.kext succesfully but always have the kernel panic with AICPM .kext crash

 

i am under 10.6.8

have you some clues to help me ?

thanks a lot

post-316806-1313258524_thumb.jpg

Link to comment
Share on other sites

If you still have kexts in /Extra, migrate them to /S/L/E and only use one kextcache, especially for Lion.

 

On KP, is it actually AICPUPM that is crashing?

 

Yes. actually it's kinda funny because it happens randomly at startup. it has like a 20% of chance.

What is the vantage of moving all kext from /extra to /s/l/e?

Link to comment
Share on other sites

Yes. actually it's kinda funny because it happens randomly at startup. it has like a 20% of chance.

What is the vantage of moving all kext from /extra to /s/l/e?

I had a similar issue where the kextcache in /Extra was changed occasionally by a failsafe usb boot.

 

Once upon a time there was the idea of an vanilla osx installation, keeping all additional kext in /Extra or even in the hidden EFI partition.

It was never a good idea in my opinion as it caused trouble with certain kext and the additional kextcache. /S/L/E is the place for kernel extensions and there is already a kextcache used by the system, so why not use the common standard and remove one more cause for errors?

 

The only benefit I see in using /Extra/Extensions is a better overview of the additional kexts, but then again I'm only using the one essential + sound and ethernet kexts, that I would also install in /S/L/E just like any 3rd party driver.

 

 

 

i am under 10.6.8

have you some clues to help me ?

thanks a lot

Your KP shows ECX=e2, that means that the patching was not successful (wrong version of speedstepper used) or (more likely) that your kextcache is not up-to-date.

 

See above and migrate your kexts to /S/L/E to reduce caching-errors.

Link to comment
Share on other sites

Hi,

you were right, i copy the AICPM.kext in a folder on the desktop, remove the AICPM.kext on the S/L/E folder , then i applyied the patch on the one in the desktop folder and then drop the patched AICPM.kext in kextutility.

After reboot, it works.Great thank you.

But just a question, why do you give us a script, an AICPM patched kext is not easier? or does the script make specifics user motherboard modification (so the AICPM patched kext isn't the same for all users (with the same mac os x version ex:10.6.8 i mean) ?

thank you again for resolving my kernel panic!

bye

Link to comment
Share on other sites

Glad I could help guys!

 

About patcher vs. prepatched kext: I simply won't redistribute non-opensource Apple kernel extensions, ever. Patched or unpatched. What people do with this program is their own responsibility.

Link to comment
Share on other sites

dear @flAked I used an Asus p8h67-m LX and worked like a charm. No more cpunull. Running 10.7.

One question, 10.7.1 has been released some minutes ago, does this means I will have to wait 'till you release a new propper speedstepper script for the new AICPUPM.kext and so on?

(yes, of course, I could paste the patched 10.7.0 AICPUPM.kext before restarting the 10.7.1 update).

thanks in advance for your help mate.

:)

Link to comment
Share on other sites

If the AICPUPM version changed in 10.7.1 then yes, a new version of SpeedStepper is needed.

 

Wait for the update until I release a new version or use NullCPU in the meantime.

 

EDIT: The 10.7.1 update doesn't include a new version of AICPUPM, update should run fine.

Link to comment
Share on other sites

Hey flAked, I tried to use SpeedStepper on my laptop with Sandy Bridge (as suggested by madl0n, who provided me with an edited dsdt). The patching seems to work fine, but I get a kernel panic.

 

What can I do?

 

I tried to use this guide: http://www.tonymacx86.com/viewtopic.php?f=...=169929#p169929 changing the IOPlatformPluginFamily.kext but nothing has changed.

 

I also tried to use this: http://www.insanelymac.com/forum/index.php...p;#entry1687122 putting the ssdt in the same directory of my dsdt (renaming it as "ssdt.aml" since I'm using Chameleon (RC5, last version compiled today from trunk)), editing my smbios.plist to identify my laptop as a MBP8,3 and editing the IOPlatformPluginFamily.kext with the new plist file found in that post.

 

I'm a bit newbie in all this stuff, so I probably did something wrong, but I still don't know what.

 

The only other kext I installed (in S/L/E) is FakeSMC.

 

I also updated kext cache multiple times with no result.

 

Please help me!

 

I attached a picture of my KP.

 

Thanks in advance

post-420698-1313609615_thumb.jpg

Link to comment
Share on other sites

 Share

×
×
  • Create New...