Guide: X79 OS X Controlled SpeedStep (CPU Power Management)

[bigkahuna360]

Guide: Asus X79 OS X Controlled SpeedStep

Looks good to me. You have c2, c3 & c6 residencies and "AICPUPMI: CPU P-States [ 12 22 (27) 32 39 44 ]" shows you reaching 12, 22, 27, 32, 39 & 44. If you want to go deeper than what I've outlined, contact PikerAlpha

Okay.

 

[kundica]

Guide: Asus X79 OS X Controlled SpeedStep

So MacPro6,1 has become very problematic with my GTX 970 and 10.10.2 on my R4E w/3930k. I noticed Andrew suggests using MacPro5,1 and uses stingas patch(but places dummy in fakesmc), however, this results in very poor stepping/high temps for me.

Is it necessary to use MacPro6,1 with this method? I haven't messed around with my OSX install for some time so I'm a bit rusty atm.

 

[MK500]

Guide: Asus X79 OS X Controlled SpeedStep

I am running Shilohh's modified http://www.tonymacx86.com/attachmen...os-x-controlled-speedstep-r4e_patched4206.zip BIOS and have verified it is unlocked. I enabled everything on Advanced\ CPU Configuration\ CPU Power Management Configuration. I also show a C7 that isn't on Shilohh's picture; so I enabled that. I don't have the "limit" option either. These must be differences on the 3930k vs. his 4930k.

Because I have Sandy Bridge (3930k) I made some changes to the example ssdtPRGen based on:

       -cpu type [0/1/2/3]
          0 = Sandy Bridge
          1 = Ivy Bridge
          2 = Haswell
          3 = Broadwell

I selected 0 for Sandy Bridge.

and

       -workarounds for Ivy Bridge [0/1/2/3]
          0 = no workarounds
          1 = inject extra (turbo) P-State at the top with maximum (turbo) frequency + 1 MHz
          2 = inject extra P-States at the bottom
          3 = both

I selected 0 for no Ivy Bridge workarounds; since I do not have Ivy Bridge.

When I run this:

sudo /Users/administrator/Downloads/ssdtPRGen.sh-master/ssdtPRGen.sh -c 0 -w 0 -turbo 4200

I get this:

ssdtPRGen.sh v0.9 Copyright (c) 2011-2012 by † RevoGirl
             v6.6 Copyright (c) 2013 by † Jeroen
             v15.6 Copyright (c) 2013-2015 by Pike R. Alpha
-----------------------------------------------------------
Bugs > https://github.com/Piker-Alpha/ssdtPRGen.sh/issues <

gDataPath: /Users/administrator/Library/ssdtPRGen/Data
Override value: (-c) CPU type, now using: Sandy Bridge!
Override value: (-w) Ivy Bridge workarounds, now set to: 0!
Override value: CPU type changed, now using: Ivy Bridge!
Override value: (-turbo) maximum (turbo) frequency, now using: 4200 MHz!

System information: Mac OS X 10.10.2 (14C109)
Brandstring 'Intel(R) Core(TM) i7-3930K CPU @ 3.20GHz'


Warning: No ACPI Processor declarations found in the DSDT!
	 Using assumed Scope (\_SB) {}

Generating ssdt.dsl for a 'MacPro6,1' with board-id [Mac-F60DEB81FF30ACF6]
Ivy Bridge Core i7-3930K processor [0x206D7] setup [0x0a01]
With a maximum TDP of 130 Watt, as specified by Intel
Number logical CPU's: 12 (Core Frequency: 3200 MHz)
Number of Turbo States: 10 (3300-4200 MHz)
Number of P-States: 31 (1200-4200 MHz)
Injected C-States for P000 (C1,C3,C6)
Injected C-States for P001 (C1,C3,C6)

Intel ACPI Component Architecture
ASL Optimizing Compiler version 20140926-64 [Nov  6 2014]
Copyright (c) 2000 - 2014 Intel Corporation

ASL Input:     /Users/administrator/Library/ssdtPRGen/ssdt.dsl - 362 lines, 10545 bytes, 95 keywords
AML Output:    /Users/administrator/Library/ssdtPRGen/ssdt.aml - 2521 bytes, 40 named objects, 55 executable opcodes

Compilation complete. 0 Errors, 0 Warnings, 0 Remarks, 0 Optimizations

My concern is:

Override value: CPU type changed, now using: Ivy Bridge!

I also tried running without the -w flag at all:

sudo /Users/administrator/Downloads/ssdtPRGen.sh-master/ssdtPRGen.sh -c 0 -turbo 4200

and got this:

ssdtPRGen.sh v0.9 Copyright (c) 2011-2012 by † RevoGirl
             v6.6 Copyright (c) 2013 by † Jeroen
             v15.6 Copyright (c) 2013-2015 by Pike R. Alpha
-----------------------------------------------------------
Bugs > https://github.com/Piker-Alpha/ssdtPRGen.sh/issues <

gDataPath: /Users/administrator/Library/ssdtPRGen/Data
Override value: (-c) CPU type, now using: Sandy Bridge!
Override value: (-turbo) maximum (turbo) frequency, now using: 4200 MHz!

System information: Mac OS X 10.10.2 (14C109)
Brandstring 'Intel(R) Core(TM) i7-3930K CPU @ 3.20GHz'


Warning: No ACPI Processor declarations found in the DSDT!
	 Using assumed Scope (\_SB) {}

Generating ssdt.dsl for a 'MacPro6,1' with board-id [Mac-F60DEB81FF30ACF6]
Sandy Bridge Core i7-3930K processor [0x206D7] setup [0x0a01]
With a maximum TDP of 130 Watt, as specified by Intel
Number logical CPU's: 12 (Core Frequency: 3200 MHz)
Number of Turbo States: 10 (3300-4200 MHz)
Number of P-States: 31 (1200-4200 MHz)
Injected C-States for P000 (C1,C3,C6,C7)

Error: board-id [Mac-F60DEB81FF30ACF6] not supported by Sandy Bridge – check SMBIOS data / use the -c option

Do you want to continue (y/n)? y
Warning: 'system-type' may be set improperly (3 instead of 2)

Intel ACPI Component Architecture
ASL Optimizing Compiler version 20140926-64 [Nov  6 2014]
Copyright (c) 2000 - 2014 Intel Corporation

ASL Input:     /Users/administrator/Library/ssdtPRGen/ssdt.dsl - 278 lines, 8142 bytes, 64 keywords
AML Output:    /Users/administrator/Library/ssdtPRGen/ssdt.aml - 2090 bytes, 27 named objects, 37 executable opcodes

Compilation complete. 0 Errors, 0 Warnings, 0 Remarks, 0 Optimizations

This concerns me:

Error: board-id [Mac-F60DEB81FF30ACF6] not supported by Sandy Bridge

I'm nervous to reboot with SSDT generated from either of the above two methods; since the last time I did I ended up with major stability issues (which ended up corrupting a bunch of files in my file system). Is there a way I can "preview" the SSDT before running it? I think the minimum voltage was just too low last time and was causing my CPU to have issues.

EDIT: I tried a reboot with just the config.plist changes and no SSDT. This seemed to do the same thing as when I tried putting an SSDT in. It brings my CPU voltage down to the .8x v range. As soon as this happens I'm getting kicked back to the login window repeatedly until I can get some load running....then it will jump to 1.12v and up...and stabilize. Unfortunately it "crashed" out to login enough that it finally seems to have killed itself again. I can no longer boot (ending up at slash icon).

I should probably just stop messing with this and do another clean install without SSDT and just using my standard stable 4.2 overclock via BIOS. I've done the full 10.10 install enough times now that I'm really good at it

Note: I did try to remove X79 kext; but that seems to also give me the slash.

EDIT2: After some more SSDT trials I gave up. I just did a clean install and locked it to 4.2Ghz at 1.26v in BIOS. Got a solid 12 hours of Prime95 with no errors. Machine seems great now and temps never get above 61c. My geek bench is 3065/22480; so not nearly as fast as Shilohh's 4156/25795. That is more difference than I would expect from the difference between 4.2 and 4.5Ghz peak...maybe the 4930k is just more efficient.

Has anyone else gotten full speedstep working on the 3930k on Rampage IV Extreme and Yosemite? Which BIOS? What speeds and temps are you getting? How many hours of Prime95 have you run (stability test)?

 

[shilohh]

Guide: Asus X79 OS X Controlled SpeedStep

Because I have Sandy Bridge (3930k) I made some changes to the example ssdtPRGen based on:

I selected 0 for Sandy Bridge.

and

I selected 0 for no Ivy Bridge workarounds; since I do not have Ivy Bridge.

- You should create the SSDT exactly as the guide dictates. I know that you have SandyB-E and not IvyB-E but os X uses them the same way as they are both X-79. I feel like PikeRAlpha's labeling could have been a little less confusing for this script but he states he's never run his X79 system as a hack and always seems to focus his work on lower end systems. The script is referring to "Sandy Bridge" and not "Sandy Bridge-E". Both SB-E and IB-E function like the standard Ivy Bridge. You need the extra states added by the -w 3 or your CPU won't step. You also need the code added by -c 1.

This concerns me:

Error: board-id [Mac-F60DEB81FF30ACF6] not supported by Sandy Bridge

-Yes this is true and thats why you should create the SSDT as the guide says.

I'm nervous to reboot with SSDT generated from either of the above two methods; since the last time I did I ended up with major stability issues (which ended up corrupting a bunch of files in my file system). Is there a way I can "preview" the SSDT before running it? I think the minimum voltage was just too low last time and was causing my CPU to have issues.

EDIT: I tried a reboot with just the config.plist changes and no SSDT. This seemed to do the same thing as when I tried putting an SSDT in. It brings my CPU voltage down to the .8x v range. As soon as this happens I'm getting kicked back to the login window repeatedly until I can get some load running....then it will jump to 1.12v and up...and stabilize. Unfortunately it "crashed" out to login enough that it finally seems to have killed itself again. I can no longer boot (ending up at slash icon).

I should probably just stop messing with this and do another clean install without SSDT and just using my standard stable 4.2 overclock via BIOS. I've done the full 10.10 install enough times now that I'm really good at it

Note: I did try to remove X79 kext; but that seems to also give me the slash.

EDIT2: After some more SSDT trials I gave up. I just did a clean install and locked it to 4.2Ghz at 1.26v in BIOS. Got a solid 12 hours of Prime95 with no errors. Machine seems great now and temps never get above 61c. My geek bench is 3065/22480; so not nearly as fast as Shilohh's 4156/25795. That is more difference than I would expect from the difference between 4.2 and 4.5Ghz peak...maybe the 4930k is just more efficient.

Has anyone else gotten full speedstep working on the 3930k on Rampage IV Extreme and Yosemite? Which BIOS? What speeds and temps are you getting? How many hours of Prime95 have you run (stability test)?

I did have full speedstep running on my 3930K when I wrote this guide. Back then we were using a patched AppleIntelCPUPowerManagement kext binary inside of a dummy kext so that the patched binary wouldn't be overwritten by OS updates. The patched binary was from 10.9.1 and ceased working with newer versions of OS X. Thats why we have clover auto patch the binary of the OEM kext at boot now, which I haven't tested as I no longer run a 3930K. I have heard many success stories so you shouldn't get discouraged. I think the issues you were facing were more to do with incorrect SSDT but could be because of under voltage stability which you could fix in your overlock settings. I currently use offset mode which I believe cause much less wear on the cpu as it drops the voltages at idle and scales them up under load. It sounds like you were observing this behavior from the optimized defaults and the stock offset voltage may be a little low for OS X speed step. I also use manual Ai Overclock Tuner instead of XMP as the XMP sets vccsa much higher than Intel or Asus recommend going. GSkill will warranty your ram but not your CPU so I decided to err on the safe side with VCCSA and set all the other voltages and timings based on the XMP profile. You can see all my settings from my 4.5ghz offset overclock from my 3930K on 4206 below.



My 4930K 4.4Ghz Offset OC Temps/Voltages:

Idle

Load

 

[MK500]

Guide: Asus X79 OS X Controlled SpeedStep

I have heard many success stories so you shouldn't get discouraged. I think the issues you were facing were more to do with incorrect SSDT but could be because of under voltage stability which you could fix in your overlock settings. I currently use offset mode which I believe cause much less wear on the cpu as it drops the voltages at idle and scales them up under load. It sounds like you were observing this behavior from the optimized defaults and the stock offset voltage may be a little low for OS X speed step.

My SpeedStep is alive! New geekbench is 3861/23233. Thanks so much for taking the time to write up this response Shilohh! It gave me the confidence to fix my issues and get this working.

I am using the exact settings you gave in the BIOS photos. My VCCSA is now WAY lower than before; which is amazing. I did all my original VCCSA testing on an early BIOS and could never get anything much below what G.Skill recommended without panics. Now I'm idling around 1.1v and peaking around 1.1375v.

I'm concerned about my VCORE though. It idles nicely around 1.0v with brief drops as low as .74v, but at full load (Prime95) it sits quite a lot at 1.35v; which is max safe range. I normally run this CPU as high as 1.27v; but haven't really been over 1.3v much ever. I guess it's a bit of a tradeoff. Much better VCCSA for a little scary VCORE. Maybe I shouldn't worry about the 1.35v as it will rarely be there during normal use (as nothing stresses quite as hard as Prime95).

How do you have your CPU fan set? I think I may have to tweak my settings there. I'm currently hitting 71c max according to HWMonitor; at the same time ROG OC KEY is showing 58c. I'm not sure how much I trust HWMonitor; as the numbers seem a bit off from ROG OC KEY often.

Oh; for power utilization...total system watts:

Before:
270 idle / 350 load
Now:
150 idle / 415 load

*Note that this is not loading GPU...I'm sure quite a bit higher if I also loaded the 680 at same time.

Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: v3.3 Copyright © 2012-2014 Pike R. Alpha. All rights reserved
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: logMSRs............................: 1
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: logIGPU............................: 0
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: logCStates.........................: 1
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: logIPGStyle........................: 0
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: MWAIT C-States.....................: 135456
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: MSR_CORE_THREAD_COUNT......(0x35)  : 0x6000C
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: MSR_PLATFORM_INFO..........(0xCE)  : 0xC0070012000
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: MSR_PMG_CST_CONFIG_CONTROL.(0xE2)  : 0x1E000403
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: MSR_PMG_IO_CAPTURE_BASE....(0xE4)  : 0x20414
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: IA32_MPERF.................(0xE7)  : 0xF0DCEAEC62
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: IA32_APERF.................(0xE8)  : 0x1316083115F
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: MSR_FLEX_RATIO.............(0x194) : 0xE0000
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: MSR_IA32_PERF_STATUS.......(0x198) : 0x2AC100002B00
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: MSR_IA32_PERF_CONTROL......(0x199) : 0x2B00
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: IA32_CLOCK_MODULATION......(0x19A) : 0x0
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: IA32_THERM_STATUS..........(0x19C) : 0x88170008
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: IA32_MISC_ENABLES..........(0x1A0) : 0x850089
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: MSR_MISC_PWR_MGMT..........(0x1AA) : 0x400001
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: MSR_TURBO_RATIO_LIMIT......(0x1AD) : 0x393939393939
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: IA32_ENERGY_PERF_BIAS......(0x1B0) : 0x0
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: MSR_POWER_CTL..............(0x1FC) : 0x2104005F
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: MSR_RAPL_POWER_UNIT........(0x606) : 0xA1003
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: MSR_PKG_POWER_LIMIT........(0x610) : 0x8006FFFF00C8FFFF
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: MSR_PKG_ENERGY_STATUS......(0x611) : 0xD8B54E86
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: MSR_PKGC3_IRTL.............(0x60a) : 0x8850
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: MSR_PKGC6_IRTL.............(0x60b) : 0x8868
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: MSR_PKGC7_IRTL.............(0x60c) : 0x886D
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: MSR_PP0_CURRENT_CONFIG.....(0x601) : 0x80001FFF
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: MSR_PP0_POWER_LIMIT........(0x638) : 0x80000000
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: MSR_PP0_ENERGY_STATUS......(0x639) : 0xB0A725CF
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: MSR_PP0_POLICY.............(0x63a) : 0x0
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: MSR_PKG_C2_RESIDENCY.......(0x60d) : 0x1440E6D820
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: MSR_PKG_C3_RESIDENCY.......(0x3f8) : 0x160BAE60
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: MSR_PKG_C6_RESIDENCY.......(0x3f9) : 0x1F162AB600
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: MSR_PKG_C7_RESIDENCY.......(0x3fa) : 0x0
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: CPU Low Frequency Mode.............: 1200 MHz
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: CPU Maximum non-Turbo Frequency....: 3200 MHz
Apr  6 11:04:15 MKMacPro kernel[0]: AICPUPMI: CPU Maximum Turbo Frequency........: 5700 MHz
Apr  6 11:04:16 MKMacPro kernel[0]: AICPUPMI: CPU P-States [ (43) ]
Apr  6 11:04:16 MKMacPro kernel[0]: AICPUPMI: CPU C3-Cores [ 0 1 5 8 10 11 ]
Apr  6 11:04:16 MKMacPro kernel[0]: AICPUPMI: CPU C6-Cores [ 0 1 4 5 7 ]
Apr  6 11:04:16 MKMacPro kernel[0]: AICPUPMI: CPU C3-Cores [ 0 1 2 3 5 7 8 10 11 ]
Apr  6 11:04:16 MKMacPro kernel[0]: AICPUPMI: CPU C6-Cores [ 0 1 4 5 6 7 8 ]
Apr  6 11:04:17 MKMacPro kernel[0]: AICPUPMI: CPU C3-Cores [ 0 1 2 3 4 5 7 8 10 11 ]
Apr  6 11:04:17 MKMacPro kernel[0]: AICPUPMI: CPU C6-Cores [ 0 1 3 4 5 6 7 8 11 ]
Apr  6 11:04:17 MKMacPro kernel[0]: AICPUPMI: CPU C3-Cores [ 0 1 2 3 4 5 6 7 8 9 10 11 ]
Apr  6 11:04:17 MKMacPro kernel[0]: AICPUPMI: CPU C6-Cores [ 0 1 3 4 5 6 7 8 9 10 11 ]
Apr  6 11:04:18 MKMacPro kernel[0]: AICPUPMI: CPU C6-Cores [ 0 1 2 3 4 5 6 7 8 9 10 11 ]
Apr  6 11:07:35 MKMacPro kernel[0]: AICPUPMI: CPU P-States [ (12) 43 ]
Apr  6 11:07:43 MKMacPro kernel[0]: AICPUPMI: CPU P-States [ 12 (22) 43 ]
Apr  6 11:07:59 MKMacPro kernel[0]: AICPUPMI: CPU P-States [ 12 22 (38) 43 ]
Apr  6 11:08:09 MKMacPro kernel[0]: AICPUPMI: CPU P-States [ 12 22 (27) 38 43 ]
Apr  6 11:09:43 MKMacPro kernel[0]: AICPUPMI: CPU P-States [ 12 22 27 (32) 38 43 ]

 

[MK500]

Guide: Asus X79 OS X Controlled SpeedStep

I'm concerned about my VCORE though. It idles nicely around 1.0v with brief drops as low as .74v, but at full load (Prime95) it sits quite a lot at 1.35v; which is max safe range. I normally run this CPU as high as 1.27v; but haven't really been over 1.3v much ever. I guess it's a bit of a tradeoff. Much better VCCSA for a little scary VCORE. Maybe I shouldn't worry about the 1.35v as it will rarely be there during normal use (as nothing stresses quite as hard as Prime95).

I think I may have a solution! I found that if I set the Turbo Ratio (by all cores) to 42 in BIOS it will max at 4.2Ghz. So that is fine. Then I set the Offset Mode to (-) and the Offset to .020.

Now my CPU will scale up to a max of 4.2Ghz at right around the old 1.26v I usually run (it occasionally goes a bit higher). But I'm no longer seeing the 1.35v+ numbers that were causing me problems (getting too hot; 77c+).

The only problem is that I'm concerned that at the lowest speed step; the volts are now dropping too far. As I can't find a way to set a "minimum"; I'm wondering if it is possible to delete one or two steps off the bottom end of my SSDT. Is there a guide for manually editing the SSDT? I should probably also modify the top step to be 42 instead of 43 also.

By the way; making these changes is effecting my geek bench score very little...so I still get the benefits of much faster single core performance even with the overall limit set to 4.2Ghz :-D

Thanks again for all the help! It's great to see this working so well.

 

[MK500]

Guide: Asus X79 OS X Controlled SpeedStep

The only problem is that I'm concerned that at the lowest speed step; the volts are now dropping too far. As I can't find a way to set a "minimum"; I'm wondering if it is possible to delete one or two steps off the bottom end of my SSDT. Is there a guide for manually editing the SSDT? I should probably also modify the top step to be 42 instead of 43 also.

Yeah; I got 13 hours of Prime95 with zero errors at (-).02.....but as soon as I quit Prime95 and did some work I got a lockup. I'm assuming the voltage dropped too far (looked like it was around .7v VCORE when it locked). For now I pushed it back up to (+).01 as that doesn't cause temperature problems unless running Prime95.

Researching manual SSDT editing to remove those lower speed steps. I think I probably don't need to step down below 2Ghz or so. SSDTs are complicated.

I'm a bit stumped as to how I can check stability at low v. I can do something like prime95 for max v; but how do I know I'm not getting errors when I step down to lower speeds (and lower v)?

Maybe it seems like I'm overthinking this; but I need to know there aren't errors creeping in. For those doing graphical or rendering work I suppose a pixels out of place every thousand frames isn't the end of the world...but I do databases and code where a misplaced bit could be devastating. A lot of little memory or calculation errors can creep in before you get an actual crash or lockup.

 

[shilohh]

Guide: Asus X79 OS X Controlled SpeedStep

Setting up offset can be tricky. I feel like the only way to get it set right is narrow down your settings, load optimized defaults, clear cmos and reenter everything manually again. Then stability test again. I feel like it holds on to old stuff sometimes.

I deffinetly found that using negative (-) offset values was not stable at idle. Getting the balance right is key. I find that I need 0.8-1.0 vcore at idle. You can play with load line calibration to bring your full load vcore up or down a little. I have found that nothing stresses the cpu as hard as prime so geekbench's stress test might be a better indicator of real world heat scenarios. Id try to find your min stable idle vcore. Then shoot for the highest stable multiplier. If your not comfortable with the max vcore, try lowering the multiplier. Use load line calibration if you need a little more at full load. Even if your max vcore is 1.3, I think it's better than runnin 1.27 24/7.

I I think my current min is .9ish and max is 1.29ish.

About ssdt editing. You can create a ssdt for a 5ghz max turbo and it will still only reach 4.2 if that's what you set in the bios. The generic one I distribute in the Eric build is created for max turbo of 4.5 but people only reach the turbo they set in their bios as long as it's below 4.5. Also I wouldn't mess with the lower values manually. You could try disabling c1e in the bios though. Or use pikes script ant play with the -w numbers.

 

[jbauman5]

Guide: Asus X79 OS X Controlled SpeedStep

Hi all,

I have a fully working x79-up4 with a 4930k and am thinking of putting in a 2680v2. I know the CPU is supported by the motherboard but I'm curious if anyone knows what else I may have to adjust to make things work. I know I'll have to edit my voodooTSCSync from 11 to 19 for the 10 core, but is that it? I am on 10.10.4.

Thanks,

jb

 

[shilohh]

Guide: Asus X79 OS X Controlled SpeedStep

From experience I'd say that editin the logical CPU count in voodootscsync is not actually necessary but it won't hurt to set it like the developer recomends. Off the top of my head, I'd say you'll need to delete your CPU ssdt and dsdt if you have one, install the CPU and create a new CPU ssdt and dsdt (if you use a dsdt). I can't think of anything else that may be necessary at the moment.