Jump to content

[Guide] Mac OSX 10.12 and 10.13 with X99 Broadwell-E family and Haswell-E family


nmano
 Share

1,954 posts in this topic

Recommended Posts

Forgot to remove it. And........ fixed.   :ninja:

 

Prefer to release .command over .sh, as it makes it more user-friendly--allow users to double click the file to execute. 

Seen.  :lol:

Yes. But I prefer to open a .command via Sublime Text.  ^_^

hi PMheart and mfc88

if you agree, of course

put MTH in app folder

Good job !

 

EDIT:

Some suggestions: 

1 - Check SIP status before verifying/repairing permissions, rebuilding caches, installing kexts, loading AppleIntelInfo

2 - Make `bdmesg` embedded in your app.

3 - Make a independent menu to choose English/French at a glance.

Edited by PMheart
  • Like 1
Link to comment
Share on other sites

 

Forgot to remove it. And........ fixed.   :ninja:

 

Prefer to release .command over .sh, as it makes it more user-friendly--allow users to double click the file to execute. 

 

http://www.insanelymac.com/forum/topic/313296-guide-mac-osx-1012-with-x99-broadwell-e-family-and-haswell-e-family/page-61?do=findComment&comment=2371965

 

 

Thanks a lot man... In "Package Power Time Window" under "Long Duration Package Power Limit" i've leave auto (why 127 in your setting?).

On your setup do you prefer OsxAptioFix2Drv-free2000.efi or OsxAptioFixDrv-64.efi? Thanks a lot again.

Link to comment
Share on other sites

Thanks a lot man... In "Package Power Time Window" under "Long Duration Package Power Limit" i've leave auto (why 127 in your setting?).

On your setup do you prefer OsxAptioFix2Drv-free2000.efi or OsxAptioFixDrv-64.efi? Thanks a lot again.

It basically means after hitting a TDP of 140watts for 127s it will start to decrease the core clocks to prevent excessive CPU voltage/heat. At a TDP of 168watts immediately decrease CPU clocks.

 

Been currently running PMheart's OsxAptioFixDrv-64.efi and it's been relatively stable for a few weeks. Used OsxAptioFix2Drv-free2000.efi prior to that and it's been relatively stable for a few months. 

 

Up to you on which to run (according to your mobo).

Link to comment
Share on other sites

@Fljagd @mfc

I noticed that in the results of your Intel Info it appears like this!

MSR_PMG_CST_CONFIG_CONTROL.......(0xE2)  : 0x8400

------------------------------------------

 - I/O MWAIT Redirection Enable......... : 1 (enabled, IO read of MSR(0xE4) mapped to MWAIT)

 - CFG Lock............................. : 1 (MSR locked until next reset)

 - C3 State Auto Demotion............... : 0 (disabled/unsupported)

 - C1 State Auto Demotion............... : 0 (disabled/unsupported)

 - C3 State Undemotion.................. : 0 (disabled/unsupported)

 - C1 State Undemotion.................. : 0 (disabled/unsupported)

 - Package C-State Auto Demotion........ : 0 (disabled/unsupported)

 - Package C-State Undemotion........... : 0 (disabled/unsupported)

 

 

I've been running Intel Info and in mine it appears like this with just tree XCPM patches. I tried only with two patches, but it seemed to me more stable with three. But with more time I still have to explore better with only two. Anyway, as it stands it seems perfect!

 

AppleIntelInfo.kext v2.0 Copyright © 2012-2016 Pike R. Alpha. All rights reserved

enableHWP................................: 0

Settings:

------------------------------------------

logMSRs..................................: 1

logIGPU..................................: 0

logCStates...............................: 1

logIPGStyle..............................: 1

InitialTSC...............................: 0x1695b0b1451 (47 MHz)

MWAIT C-States...........................: 8480

Processor Brandstring....................: Intel® Core i7-5820K CPU @ 3.30GHz

Processor Signature..................... : 0x306F2

------------------------------------------

 - Family............................... : 6

 - Stepping............................. : 2

 - Model................................ : 0x3F (63)

Model Specific Registers (MSRs)

------------------------------------------

MSR_CORE_THREAD_COUNT............(0x35)  : 0xFFFFFF80A80C5F00

------------------------------------------

 - Core Count........................... : 6

 - Thread Count......................... : 12

MSR_PLATFORM_INFO................(0xCE)  : 0x20080C3BF3812100

------------------------------------------

 - Maximum Non-Turbo Ratio.............. : 0x21 (3300 MHz)

 - Ratio Limit for Turbo Mode........... : 1 (programmable)

 - TDP Limit for Turbo Mode............. : 1 (programmable)

 - Low Power Mode Support............... : 1 (LPM supported)

 - Number of ConfigTDP Levels........... : 1 (additional TDP level(s) available)

 - Maximum Efficiency Ratio............. : 12

 - Minimum Operating Ratio.............. : 8

MSR_PMG_CST_CONFIG_CONTROL.......(0xE2)  : 0x1E000005

------------------------------------------

 - I/O MWAIT Redirection Enable......... : 0 (not enabled)

 - CFG Lock............................. : 0 (MSR not locked)

 - C3 State Auto Demotion............... : 1 (enabled)

 - C1 State Auto Demotion............... : 1 (enabled)

 - C3 State Undemotion.................. : 1 (enabled)

 - C1 State Undemotion.................. : 1 (enabled)

 - Package C-State Auto Demotion........ : 0 (disabled/unsupported)

 - Package C-State Undemotion........... : 0 (disabled/unsupported)

 

My question is: Which one of us is right and which one is wrong? 

And as no one has spoken, for me, the answer is: I am the one who is well and I'm going to have a gin to make the party  :thumbsup_anim: 

post-1313347-0-86961900-1488148184_thumb.png

Link to comment
Share on other sites

Thank  :wub:  :wink_anim:

Hi Flijagd.

 

I've optimized your M-T-H.  ^_^

 

New features:

1 - Check SIP status before some important verbs like installing kexts, repairing permissions.

2 - More precise code.

3 - Installing multi kexts simultaneously.

4 - We don't have to put this app into /Applications cuz I've got rid of absolute path.

 

Defects:

1 - NO French support. (Sorry. I don't understand French...)

2 - Potential bugs. Need further testing.

 

Please test and feedback !  :D

 

 

M-T-H.app.zip

  • Like 1
Link to comment
Share on other sites

Today with a little more time, I've been testing for a few hours and my system works perfectly just with two patches. Idle almost always at 1200 MHz, with all the Steps (P, C6 and C3). Test done with XMP Profile 1 enabled (prints attached). However, for better performance it seems necessary to use the Patch XCPM Performance Fix as well. I tested AppleIntelInfo Loader by mfc88 and PMheart and it worked, but interestingly I got more speedsteeps doing the test manually and did not have to wait that long...

The Intel Info Kext used in both modes was the last v2.4 and the Geek 4.0.3. Soon I will do some more tests to confirm if stability is maintained.

 

EDIT: Ok... I just did more two hours of testing with only two XCPM patches and it's perfect for me. Good Idle with low temperatures and consumption and sufficient performance for what I want, without overclocking and with stability. For now it will keep like this!

post-1313347-0-94105400-1488217299_thumb.png

post-1313347-0-04955200-1488217341_thumb.png

post-1313347-0-26090600-1488217503_thumb.png

post-1313347-0-91193700-1488224436_thumb.png

  • Like 1
Link to comment
Share on other sites

Hi Flijagd.

 

I've optimized your M-T-H.  ^_^

 

New features:

1 - Check SIP status before some important verbs like installing kexts, repairing permissions.

2 - More precise code.

3 - Installing multi kexts simultaneously.

4 - We don't have to put this app into /Applications cuz I've got rid of absolute path.

 

Defects:

1 - NO French support. (Sorry. I don't understand French...)

2 - Potential bugs. Need further testing.

 

Please test and feedback !  :D

 

 

attachicon.gifM-T-H.app.zip

Hello

I just added a line to facilitate the reading of the results

I like the principle of function

Thank you

Not serious for the French

Thank  :D

post-1181448-0-36254100-1488218343_thumb.png

  • Like 2
Link to comment
Share on other sites

Today with a little more time, I've been testing for a few hours and my system works perfectly just with two patches. Idle almost always at 1200 MHz, with all the Steps (P, C6 and C3). Test done with XMP Profile 1 enabled (prints attached). However, for better performance it seems necessary to use the Patch XCPM Performance Fix as well. I tested AppleIntelInfo Loader by mfc88 and PMheart and it worked, but interestingly I got more speedsteeps doing the test manually and did not have to wait that long...

The Intel Info Kext used in both modes was the last v2.4 and the Geek 4.0.3. Soon I will do some more tests to confirm if stability is maintained.

 

EDIT: Ok... I just did more two hours of testing with only two XCPM patches and it's perfect for me. Good Idle with low temperatures and consumption and sufficient performance for what I want, without overclocking and with stability. For now it will keep like this!

 

Hi!  I think it's too much power consumption (watt).      It must run between 2-5 watt at 1,2 frequency.    Istruggling with this myself to,to get the perfect power management, I obtained 7wat but the frequency isn't correct.

post-1219427-0-71105200-1488261241_thumb.png

  • Like 2
Link to comment
Share on other sites

Hi!  I think it's too much power consumption (watt).      It must run between 2-5 watt at 1,2 frequency.    Istruggling with this myself to,to get the perfect power management, I obtained 7wat but the frequency isn't correct.

I think the balance between the consumption of mine and the overall result is very good. In addition to having different processors, hardware and probably different Bios settings, we have to take into account the applications that are running in the background and that use the processor. In my case I have some installed and obviously interfere with the idle of the processor, consumption... etc. But as I said overall it's perfect for me, although I know it's always possible to improve anything else.

Now it's time to make the update for macOS 10.12.4 beta 4 (16E175b) :yes:

 

EDIT: Already done the update for 12.4 beta 4 (16E175b). Once again everything ok only with two XCPM patches! (Thanks PMheart) 

​Sygey02, about lowering the consumption in idle, if you want to give some tips, I can try ;) 

post-1313347-0-12090500-1488385626_thumb.png

post-1313347-0-60244600-1488385637_thumb.png

post-1313347-0-14356000-1488406015_thumb.png

  • Like 2
Link to comment
Share on other sites

Hi!

 

Anybody in this thread is running stable system with Above 4G enabled in BIOS?

 

Thanks.

My system is more than stable with 4G enabled. I use thunderbolt2 EX dual card with audio card and raid attached and since when i use 4genabled i've no more random memory allocation errors and peripherals work smoothly.

Link to comment
Share on other sites

My system is more than stable with 4G enabled. I use thunderbolt2 EX dual card with audio card and raid attached and since when i use 4genabled i've no more random memory allocation errors and peripherals work smoothly.

 

Cool!

I had to disabled it because it was random freezing my machine. Two days until I realized the problem was Above 4G -> Enable. It wasn't a problem when I was configuring my machine (short periods of time between reboots) because the freezing rarely happens before 30-60 minutes of use. It's also a fact that enabling it I lose sleep (machine freezes on wake up event).

 

Now It's a pain in the ass again to struggle with aptio memory allocation problems. Can help me find what is the difference between my conf and yours?

 

I'd like to check your bios configuration (you also have a thunderbolt 3 card, haven't you?) and efi folder if you don't mind. For BIOS, there is an option to save your config as a txt file in tools>oc profiles (or something similar).

 

Thank you very much.

Link to comment
Share on other sites

I think the balance between the consumption of mine and the overall result is very good. In addition to having different processors, hardware and probably different Bios settings, we have to take into account the applications that are running in the background and that use the processor. In my case I have some installed and obviously interfere with the idle of the processor, consumption... etc. But as I said overall it's perfect for me, although I know it's always possible to improve anything else.

Now it's time to make the update for macOS 10.12.4 beta 4 (16E175b) :yes:

 

EDIT: Already done the update for 12.4 beta 4 (16E175b). Once again everything ok only with two XCPM patches! (Thanks PMheart) 

​Sygey02, about lowering the consumption in idle, if you want to give some tips, I can try ;)

Frequency looks pretty good for idling. Not too many spikes/valleys with some flat spots. However, I'd be curious to see if you ran any stress tests what your MAX frequency would reach and whether or not it'd flat-line at max (good) or if it would spike at peak and then immediately valley,  then spike at peak and then valley...etc (bad). If it's 100%, it should stay at the peak.

 

On a side note, top tip for screenshots: If you want to take WINDOW screenshots, simply press Command+Shift+4 then press the Space Bar (cursor should turn from a crosshair to a camera)!

 

Window Screenshot (camera icon):

gYQaVkg.png

 

Snapshot (crosshair):

FmZGFjM.png

 

Or... a UEFI BIOS Screenshot (insert a *NTFS/FAT* formatted USB drive and press F12):

A1l8V2f.png

  • Like 1
Link to comment
Share on other sites

Cool!

I had to disabled it because it was random freezing my machine. Two days until I realized the problem was Above 4G -> Enable. It wasn't a problem when I was configuring my machine (short periods of time between reboots) because the freezing rarely happens before 30-60 minutes of use. It's also a fact that enabling it I lose sleep (machine freezes on wake up event).

 

Now It's a pain in the ass again to struggle with aptio memory allocation problems. Can help me find what is the difference between my conf and yours?

 

I'd like to check your bios configuration (you also have a thunderbolt 3 card, haven't you?) and efi folder if you don't mind. For BIOS, there is an option to save your config as a txt file in tools>oc profiles (or something similar).

 

Thank you very much.

Here is my efi and pictures of my TB settings. I've a thunderbolt card 3 too but i prefer to use TB2 because i've thunderbolt 2 audio interface so i prefer to use it with tb2 card instead of buying an adapter. Try to use my aptio drv with 4g enabled and check if you still have issues.

post-1194286-0-25432200-1488782054_thumb.jpg

post-1194286-0-31711000-1488782055_thumb.jpg

EFI-clover39xx.zip

  • Like 1
Link to comment
Share on other sites

Congratulations Fergarth!

I don't think you need CodecCommander and any modifications on AppleHDA. Please try the method on http://www.insanelymac.com/forum/topic/313296-guide-mac-osx-1012-with-x99-broadwell-e-family-and-haswell-e-family/?p=2371527

 

Note that you should keep /SLE/AppleHDA.kext vanilla, if you don't have one please try this: (Extracted from 10.12.4 16E163f, with version 279.54)

attachicon.gif10124_ 16E163f_279.54_AppleHDA.kext_vanilla.zip

 

So what you should do:

1 - Remove any modifications on AppleHDA, in other words, re-install my vanilla AppleHDA above.

2 - Remove CodecCommander, I don't think you need it.

3- Try the way I mentioned above (that link)

Hello my friend PMheart! :) 

Sorry for the late reply on this.

If I'm not wrong (and I'm not) my AppleHDA.kext is and has always been vanilla, because I've already made several updates to the Sierra and never lost the onboard sound, nor did I have to re-apply commands. I only use the HDA patches in the Clover Plist.

Regarding CodecCommander.kext, I do not really need it for the sound to work well. Just use it to "fix the no sound after sleep".

EDIT: But yes, when I have time, I have to try the method you suggested and if it works with "sound after sleep" I will use it.

Now with Sierra 12.4 beta 4 and last Clover r4035

Frequency looks pretty good for idling. Not too many spikes/valleys with some flat spots. However, I'd be curious to see if you ran any stress tests what your MAX frequency would reach and whether or not it'd flat-line at max (good) or if it would spike at peak and then immediately valley,  then spike at peak and then valley...etc (bad). If it's 100%, it should stay at the peak.

Hi friend mfc, thanks for the print tips.

As you well know, it is not easy to combine all this perfectly, but much we already have! As you know, currently I'm only using two XCPM patches, which seems to me to have a more perfect idle and overall I think works fine. I made a CPU stress test and the result is in the attached print.

I'm still waiting for a new 12.4 NV Web Drive :( 

Test done with Sierra 12.4 beta 4, without graphical acceleration and last Clover r4035

post-1313347-0-00802400-1488823023_thumb.png

  • Like 1
Link to comment
Share on other sites

 Share

×
×
  • Create New...