Jump to content

Testers needed: CPU Power Management for SB and IB Xeon(s) or i7-39xx on X79 or C60x chipset


  • Please log in to reply
556 replies to this topic

#501
xelanaiznac

xelanaiznac

    InsanelyMac Geek

  • Members
  • PipPipPipPip
  • 211 posts
  • Gender:Male

hi rampage. here a ssdt for the 4960x (made with -w 3) at stock freq, turbo 4000

https://dl.dropboxus.../ssdt-1.aml.dmg

 

done using vanilla appleintelcpupowermanagement

hi rampage, here the 4960x with turbo 4500

https://dl.dropboxus...5 4690x.aml.zip



#502
stinga11

stinga11

    InsanelyMac Geek

  • Members
  • PipPipPip
  • 105 posts

In the second beta of 10.9.3 the appleintelcpupowermanagement changed a little bit but the omni patch for sandy bridge-E still working.

 

842f01
to
852f01
 
 
3e753a
to
3e9090
 
 
845f01
to
855f01
 
 
7410b9
to
eb10b9
 
 
7507b9
to
eb07b9
 
 
fc0274
to
fc02eb
 
 
017458
to
01eb58


#503
fabiosun

fabiosun

    InsanelyMac Geek

  • Members
  • PipPipPip
  • 103 posts
  • Gender:Male

updated to OS X 10.9.3 (13D17)

 

New AICPUPM.kext (v.216.100.1 previous was v216.0.0)

 

Pros

Now IPG and HW monitors see OOTB low steps (cpu 2690 V2)

 

Cons

AICPUPMI now freezes my system

Greater wattage and temperature shown

with kextstat seems that AICPUPM is not loaded

 

I have to do others tests but probably I will revert to first 10.9.3 beta or better to 10.9.2

 

added

to check if mach_kernel is the same of previous versions


Edited by fabiosun, 16 March 2014 - 06:21 AM.


#504
Rampage Dev

Rampage Dev

    InsanelyMac Deity

  • Local Moderators
  • 2,021 posts
  • Gender:Male

updated to OS X 10.9.3 (13D17)

 

New AICPUPM.kext (v.216.100.1 previous was v216.0.0)

 

Pros

Now IPG and HW monitors see OOTB low steps (cpu 2690 V2)

 

Cons

AICPUPMI now freezes my system

Greater wattage and temperature shown

with kextstat seems that AICPUPM is not loaded

 

I have to do others tests but probably I will revert to first 10.9.3 beta or better to 10.9.2

 

added

to check if mach_kernel is the same of previous versions

 

Try roll back as well. 



#505
fabiosun

fabiosun

    InsanelyMac Geek

  • Members
  • PipPipPip
  • 103 posts
  • Gender:Male

reverted to previous AICPUPM , AICPUPMCLIENT and mach_kernel

IPG and HWSensors does not show anymore low steps

temps and wattage are ok

 

I have a noob consideration...

with latest 10.9.3 beta seems that part of extra SSDT (i am using only Pike's ssdt) is not seen, I am refer in particular to turbo states and PM

X86platformplugin family is always in IOREG, but it seems like xcpm is used by default

 

?



#506
omni

omni

    InsanelyMac Sage

  • Members
  • PipPipPipPipPip
  • 286 posts
  • Gender:Male

It's nice that you guys are beta testing Apple's unfinished code, but perhaps we should wait till they come out officially with a final build before trying to get it working?



#507
fabiosun

fabiosun

    InsanelyMac Geek

  • Members
  • PipPipPip
  • 103 posts
  • Gender:Male

Hi omni,

I don't understand your statement here because forum like IM here helps to try different things before other people fixes bugs or similar stuff

 

For me, hackintosh is a funny hobby, so I try all I can..and then if things go wrong I revert it to the previous step without complain, but only trying to understand better accordly to my knowledge.

And then, people trying things in this post or in other are not beta tester?

 

Also.. I don't understand rampagedev like :-) or maybe yes? :-P

 

To clarify better:

rampage, without this thread, maybe your fakesmc test.kext solution should have had a different time release.. or not?

 

 

It's nice that you guys are beta testing Apple's unfinished code, but perhaps we should wait till they come out officially with a final build before trying to get it working?



#508
magnifico

magnifico

    InsanelyMac Deity

  • Donators
  • 2,478 posts
  • Gender:Male

To clarify better:
rampage, without this thread, maybe your fakesmc test.kext solution should have had a different time release.. or not?

In fact, he says to do not use mac mini StepContextDict, and then with this kext it emulates, and chickens bite

#509
Rampage Dev

Rampage Dev

    InsanelyMac Deity

  • Local Moderators
  • 2,021 posts
  • Gender:Male

Hi omni,

I don't understand your statement here because forum like IM here helps to try different things before other people fixes bugs or similar stuff

 

For me, hackintosh is a funny hobby, so I try all I can..and then if things go wrong I revert it to the previous step without complain, but only trying to understand better accordly to my knowledge.

And then, people trying things in this post or in other are not beta tester?

 

Also.. I don't understand rampagedev like :-) or maybe yes? :-P

 

To clarify better:

rampage, without this thread, maybe your fakesmc test.kext solution should have had a different time release.. or not?

 

Apple blowup 10.9.3 to bring out issues. Let it get to a higher build number before we begin to really dig in. Also the Test.kext injects code that corrects the issue with the lower speed step not being achieved. Using the Mac Mini SMBIOS loaded other items which in turn caused issues. This is the safe way to implement the required changes. 



#510
magnifico

magnifico

    InsanelyMac Deity

  • Donators
  • 2,478 posts
  • Gender:Male

Apple blowup 10.9.3 to bring out issues. Let it get to a higher build number before we begin to really dig in. Also the Test.kext injects code that corrects the issue with the lower speed step not being achieved. Using the Mac Mini SMBIOS loaded other items which in turn caused issues. This is the safe way to implement the required changes.


what kind of problem they had users who have used SMBIOS Mac mini, you make me understand?

#511
fabiosun

fabiosun

    InsanelyMac Geek

  • Members
  • PipPipPip
  • 103 posts
  • Gender:Male
Apple people are doing their job as usual.

If you read carefully this thread from start, you can see that ,in my case , I have tryied different Smbios.plist to understand why lower steps were not printed by IPG and Hw tools when in AICPUPMI were present and when also I could see correct temps and watts, in idle and in full load
Then, always in this thread ,people discover which was enough only to change MacMini 6.2 StepContentDict in x86platformplugin(MacPro6.1 plist), and Stinga11 creates a dummy kext to achieve same results after every OS X upgrade.

Then you repack all of these in your dmg.

After many tests i decided to use Macpro6.1 and vanilla fakeSMC because Macmini6.2 smbios (StepContntDict)( or in my rig also iMac14.2) produce lower steps in IPG and HW Sensors output but some minor problem are here overall with high OC with 3930k or 4930k CPU
My Xeon V2 instead has no problem but I prefer always regular Macpro6.1 smbios.plist because temps and watts are fine and I think that the lack of lower steps in IPG or HW could be only a cosmetic issue that programmers have to solve if they are interested to.
Patching fakesmc with MacMini6.2 StepContentDict is not the right way to achieve our goals IMHO.


Apple blowup 10.9.3 to bring out issues. Let it get to a higher build number before we begin to really dig in. Also the Test.kext injects code that corrects the issue with the lower speed step not being achieved. Using the Mac Mini SMBIOS loaded other items which in turn caused issues. This is the safe way to implement the required changes.



#512
Rampage Dev

Rampage Dev

    InsanelyMac Deity

  • Local Moderators
  • 2,021 posts
  • Gender:Male
It is the right way. I did not patch fakesmc at all for pm. I just put the file in there so people would not worry about it.

#513
magnifico

magnifico

    InsanelyMac Deity

  • Donators
  • 2,478 posts
  • Gender:Male

It is the right way. I did not patch fakesmc at all for pm. I just put the file in there so people would not worry about it.



Sure ? You are convinced that is the right way? I do not think .... try to ask someone who has the overclocked cpu ... and tell him to launch an prime 95 mac .... and see if they are stable overclock .... andrew listen with my 3930 k overclocked to 4700 emulation with the Mac mini restarts .... I had the time ... when I was working with fotoshop ... ahahahahah .... please do not say nonsense .... for me is' just aesthetics. I tell you ... and also the reason ... on asus R4 I measured the voltage tester .... and it 'all ok ... so for me it is' a mental masturbation ....

#514
Rampage Dev

Rampage Dev

    InsanelyMac Deity

  • Local Moderators
  • 2,021 posts
  • Gender:Male

Sure ? You are convinced that is the right way? I do not think .... try to ask someone who has the overclocked cpu ... and tell him to launch an prime 95 mac .... and see if they are stable overclock .... andrew listen with my 3930 k overclocked to 4700 emulation with the Mac mini restarts .... I had the time ... when I was working with fotoshop ... ahahahahah .... please do not say nonsense .... for me is' just aesthetics. I tell you ... and also the reason ... on asus R4 I measured the voltage tester .... and it 'all ok ... so for me it is' a mental masturbation ....

 

4.75 GHz here and stable. OCing issues can be due to a lot of factors. 



#515
magnifico

magnifico

    InsanelyMac Deity

  • Donators
  • 2,478 posts
  • Gender:Male

4.75 GHz here and stable. OCing issues can be due to a lot of factors.



I solved my problem ... eliminating emulation stepcontentdict Mac mini from x86 platform...lol

#516
shilohhh

shilohhh

    InsanelyMac Protégé

  • Members
  • PipPip
  • 88 posts
  • Gender:Male
  • Location:Texas
MM6,2 StepContextDict works beautifully for me.

#517
mr. west

mr. west

    InsanelyMac Protégé

  • Members
  • PipPip
  • 55 posts

Sabertooth X79/4930k user here.  I'm on 10.9.2 using Clover, although everything I describe has happened under Chameleon as well.

 

PM works fine for me using Vanilla AICPUPM and latest FakeSMC.  However I'm having stability issues.  Every once in a while it just randomly restarts.  The console log isn't much help.  A whole lot of mdworker errors.  It follows a pattern though.  Typically programs will start crashing, especially Chrome and finder.  And eventually the whole ship goes down.  This is the typical way.  Some times it is more sudden.

 

I've chased some red herrings when looking through the console logs, thinking it may be certain programs.  But in time all the crashes persist.  

 

I've run MEMTEST86 and everything is good there.

 

My current working theory is that with turbo and XMP and otherwise stock settings, perhaps the newly available speeds (3.9 ghz) are causing issues.  So I've upped the voltage, tweaked load line calibration, etc.  We'll see.  I've run overclocked and the issues seem to appear faster, so perhaps that is it.  My solutoin was always to go back to stock (or near) settings and run with no OC.  So maybe I just need more voltage etc. on stock settings.  I don't know.

 

It is frustrating.

 

Anyone else experiencing similar problems?  Besides the random restarts it all works great.



#518
joe75

joe75

    Renegade

  • Retired
  • 2,257 posts
  • Gender:Male
  • Location:Rochester, NY

Open a new topic, mr. west. Restarting without a panic is usually a hardware problem. I would suspect ram, oc, psu, can also happen from not having full graphics support. 



#519
mr. west

mr. west

    InsanelyMac Protégé

  • Members
  • PipPip
  • 55 posts

Open a new topic, mr. west. Restarting without a panic is usually a hardware problem. I would suspect ram, oc, psu, can also happen from not having full graphics support. 

 

I'll start another thread but this seemed like an appropriate place as everything is *rock solid* when not loading a SSDT for PM.  And has been for years (3820) and months (4930k pre-PM).  So I'm all ears as to what could be causing the issue, but it seems like it has to be PM related.  If I remove the SSDT for PM, everything is stable as can be.



#520
magnifico

magnifico

    InsanelyMac Deity

  • Donators
  • 2,478 posts
  • Gender:Male
@rampage dev



I know rampage that many factors help to achieve a good and stable OC, but before I was explain you that with same firmware settings I have random freeze if I use your test.kext (with inside MacMin6.2 stepcontentdict)
instead, with vanilla fakesmc and MacPro6.1 SMBios.plist all is perfect.
Do you understand me now?





0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users

© 2014 InsanelyMac  |   News  |   Forum  |   Downloads  |   OSx86 Wiki  |   Mac Netbook  |   PHP hosting by CatN  |   Designed by Ed Gain  |   Logo by irfan  |   Privacy Policy