LGA 2011 Overclocking On A Hackintosh

About:

The following are suggested LGA 2011 Six Core Over-clocking settings for Mac OS X. Mac OS X is written for a max core speed of 4.2 GHz and there is very little to be gained past that speed. That said some application can use a higher core clock. I take no responsibility for any damage that may happen due to the Overclock. For LGA 2011 Four Core users you may need more Vcore voltage then the setting listed below. More information will be coming later with these settings but I have been very busy the past 2 weeks.

The settings changed are the following:

XMP Over-clock

CPU Core Ratio: Sync All Cores

CPU Core Ratio: 42 to 45

CPU VCORE Manual Voltage: 1.35

VTT CPU Voltage: 1.2

2nd VTTCPU Voltage: 1.20302

Load Line Calibration: High

CPU Current Capability: 130% to 140%

This entry was posted in Uncategorized. Bookmark the permalink.

62 Responses to LGA 2011 Overclocking On A Hackintosh

  1. artyom says:

    Hi!
    I have several issues.
    I’ve set my bios to this OC settings and update SSDT-1 to 4,5 G When i try to test CPU with terminal command “yes” but it doesn’t work whiles with SSDT for stock speed it loads all cores.
    May be I’ve missed some BIOS settings to change.

    After several reboots mobo shows me screen with F1 options and information about unsuccessful overclock.

    DVD tray always opens when my system start and finder asks me what i have to do with blank CD that just inserted :-)) Do you know how to fix it?

    My rig: Asus Rampage IV Black Edition, Intel 4930k, 32 GB Kingstone 2400 Mhz, GTX Titan -B. 10.9.2, Clover bootloader

    Thanks alot

  2. Joshua M says:

    Are we not using the BCLK or the CPU strap?

  3. flmmkr says:

    What settings would you recommend for 4.5GHz on the 4930K?

  4. speck says:

    I have a 3930k (Sandy-E) on a P9X79-WS – and a 3820 on a GA-X79-UP4 – does this apply to those of us still on Sandy-E?

      • speck says:

        Sorry — that was poorly worded – I am on 10.9.2 (using your DMG SSDT – I need to donate, by the way!) – I meant to ask if the posted OC settings were for 6-core Sandy-E, Ivy-E or both?

        1 – I know we heard about limited X79 power management native including boost bins, but if I read right, this was only on the Ivy-E chips? Or has Sandy-E gotten power management working? (3930k on asus P9X79, 3820 on GA-X79-UP4)… If only ivy-E and no foreseeable future for PM on Sandy-E parts, I may consider a Trade up to an Ivy-E part as that can be a lot of watts, if I am understanding correctly that with NullCPUPM these are running at full bore all the time?

        2 – Will we / are we able to get rid of the VoodooTSCSync, or is that something else?

        3 – And finally, are we still optimal to use MP5,1 SMBIOS profile on X79?

      • rampagedev says:

        1. Yes 2. No 3. No

  5. Very odd, I had a perfectly working overclock on my system, but as soon as I updated to the most up to dat Rampage IV Extreme BIOS, for whatever reason my system will no longer budge higher than 3.2GHz in Mac OS X, but it works perfectly fine in Windows.

    I’ve tried updating my Chameleon, SSDT, FakeSMC and even cleared my CMOS, however nothing is getting my overclocking to work whatsoever.

    CPU: Core i7 3930K
    Mobo: Rampage IV Extreme
    BIOS Rev: 4804

    Mac OS X 10.9.2

  6. Hack Mac says:

    I’m running this command “sudo /Users/macpro/Downloads/ssdtPRGen.sh -turbo 4400 -t 130 -c 1 -w 3” – with a RIVE and 3930K (desired OC of 4.4GHz) – then I rename the output file (SSDT-1.aml) and put into /Extra. After rebooting several times, the system always boots up with my OSD MONITOR displaying “CPU RATIO 32.0 x 100.0” – OC not working…

  7. enesemini says:

    Hey Andrew,

    My overclock (4.4 GHz) on the i7 4930k works perfectly on windows, but not on OS X 10.9.2.
    I have installed the AICPM kext using the Kext Helper b7 and placed the generated (command: “sudo /Users/user/Downloads/ssdtPRGen.sh-master/ssdtPRGen.sh -turbo 4400”) SSDT-1.aml in the /extra directory. In your Read Me.txt for the Power Management it says that I have to update the SMBIOS.plist and the FakeSMC, however It doesn’t say what to change and how.

    I probably have missed something, but I can’t get my head around it.

  8. flmmkr says:

    Finally had time to do this overclock, and cannot get it to work. I have a P9X79E WS and 4930K. I used the SSDT and SSDT-1 from the DMG, as well as the SMBIOS, and made sure my boot plist contained the same settings as the boot plist from the DMG. Per the instructions, I did not use the AICPUPM kext from the DMG (“4930K does not need the patched kext”). iStat Menus shows 3.37 GHz and 1.35V. Intel Power Gadget shows 3.40GHz. Temps go up when benchmarking, but clock doesn’t and speeds are slow. Geekbench is under 18K. Latest version of Chameleon is installed.

    Things I’ve tried to get it working:
    1. Make my own SSDT-1 using ssdtprgen.sh. No change.
    2. My AICPUPM kext did have a different date than my Mac (same OS version), so I installed that kext instead thinking it may have been updated/patched at some point, so now that is the OS X native version. No change.

    Thoughts?

  9. frankiee says:

    Hi Andrew,

    two questions about OC – might be stupid ones, but I am quite new to OC:

    1) Esp. for x42 your Voltages seem a tad high? I just ramped up my multiplier to x42, and I used the following (lower) values: CPU VCORE: 1.2V / VTT CPU: 1.175 Volts / 2nd VTT: 1.05 (unchanged!). I let Prime95 run for about 1 hour without problems, Temps reached: max 81 Degrees. So the temps might be a tad too high … but would be the rest of my settings OK?

    2) XMP profile: my RAM is rated at 2133 (64 Gigs!). So while I would like to go up from the 1333MHz, I wonder if the RAM still would be stable at 2133. So would it make sense to use the XMP profile, but actually clock down a bit to 1866 MHz? And how is the best way to do this?

    Thanks for any insights!

    • rampagedev says:

      Each CPU is different. I only use voltages that are under Intel max spec. Need to keep more generic with the settings.

      Your Ram will be fine at the full spec.

      • frankiee says:

        OK, I see your point. But seems that many users might go away with lower settings. A problem I see with your recommendation is that this translates to rather high voltages (up to and slightly above 1.4V) and with LLC enabled, you could have some even higher – and undetected – spikes in voltage when load changes, which in turn might degrade your CPU faster than necessary.

        So I choose to set LLC to “auto” and also CPU Current capability back to 100%. I had a maximum of 127W during prime95, so it seems that I am just a tad below the TDP even with CPU-CC set to 100%. In conjunction with more aggressive fan settings, highest temps reached (during a rather short burn test though) were 71 degrees, so that is just a tad over TCase. Since prime95 is rather extreme, would that be OK for 24/7?

        As for RAM: so I also enabled XMP profile #1 and I immediatly saw some voltages go up: VCCSA from 0.93V up to 1.18V, VTT from 1.1V up to 1.225V. On the other hand, OCing my RAM did not nearly have as much effect as OCing the CPU: while CPU OC brought me nearly 3K in GB, OCing the RAM only contributed to 800 pts. Total OCed score (bothe RAM and CPU) leads to 22.7K GB, think I am fine with that.

        So for me, three questions arise:

        1) Since these Voltages seem quite higher, will it shorten the life of my RAM / CPU?
        2) Still, would it make sense to downclock to say 1866 MHz and lower these 2 Voltages a bit?
        3) With the rather lowperf increase, does it make sense at all to OC my RAM?

        Again, thank you vor any insights!

      • rampagedev says:

        These CPUs love current. In 10 years perhaps you could see a issue but very doubtful.

        Test with 1866 and work your way up. I run my 1866 at 2133 with timings: 13 13 12 28 @1.65v.

      • frankiee says:

        Update: so it seems that RAM is not stable anymore when using XMP, bc Intel Burn Test consistently fails with these settings 😦 I already upped Vcore to 1.3, and also re-enabled LLC and upped CPU capability to 130%, also played with VCCSA LLC – still fails! Without XMP, all seems to be fine.

        So how can I combat this? Upping 2nd VTT? Or actually downclock from the XMP settings? Any help appreciated!

      • rampagedev says:

        Odd. What are the RAM timings when not in XMP?

      • frankiee says:

        Timings before (@1333): 9-9-9-24-2T
        Timings with XMP (@2133): 11-11-11-32-2T

        Volates went up like stated above. “Rampage Mode” set to “auto” – maybe setting to “3” (Ivy-E optimized) would help?

        What is also a bit odd is that both XMP profiles seem to be the same, but one reading “2133MHz” while the other is reading “2134MHz”.

      • rampagedev says:

        Up the timings in XMP mode to: 12-12-11-32-2T

      • frankiee says:

        Yes, loosen this timings did indeed seem to help, at least at a first glance! Next – being curious – I then tried to use the 11-11.. timings, but instead downclock to 1866 to achieve stability. The result was almost the same benchmark scores, maybe a tiny bit slower (speak of 60 pts in GB for example).

        But on the other hand reported VCCSA and VTT Voltages were way down, actually down again to the original level with XMP disabled! (All these controls left to “auto” in the BIOS so it is the BIOS who decided to set these). And I ran stability tests even longer, all seem to be fine. Therefore, this seem to result in the same performance but using way less energy and therefore strain. Temps were down, too – about 3 degrees while running IBT. I think I might re-add a tiny bit to VCCSA and VTT, just to make sure.

        Since I also read elsewhere that in many cases: tighter timings > more MHz it seems that this might be also a viable option, what would you think?

  10. frankiee says:

    OK, so I played a bit more, this time with voltage offset instead of manual voltage. And the results are quite pleasing so far! Since I actually _under_volted by setting offset to -0.01 I get even lower voltages, and the big big difference is that now I have only about 0.8V while idling instead of a steady 1.2V. This translates in a CPU which is about 7-10 degrees cooler than with manual voltage! And still, I have LLCs set to “regular” and Current Capabiltiy to 100%, also VTT and VCCSA are still at their original low levels before RAM OC. Of course, your YMMV but it seems that esp for a more “mild” OC, you can get away with _much_ lower values. Of course maybe I have a “golden chip” or something, but it may be worth it to play with lower settings or using voltage offset. But carefully watch your voltages under load when trying this. Just FYI!

    • rampagedev says:

      You have won the silicon lottery.

      • frankiee says:

        Yeah, ist it so? Currently, I am now at a reported Vcore of 1.18V with x42 and while running prime95. Temps never exceed 70 degrees (am on Air). So you think that’s fine? Maybe I should try higher 😉 But even with a “lesser” chip, trying offset voltage may be well worth it, think degragation should be much lower then.

  11. kjforeman says:

    I’m unable to get power management working on an Asus Deluxe X79 + 4930k. I’ve installed AICPM and the 4.5Ghz SSDT-1.aml from your X79 DSDT DMG, used the ASUS FakeSMC, as well as removed NullCPUPowerManagement.

    Is there some other step I’m missing? Wasn’t quite sure what the SMBIOS update in the power management readme was referring to – I’m using the one from the Chameleon folder of your DMG right now. Should I change that?

    Thanks!

    • rampagedev says:

      First you did not read the read me… the patched kext is not needed for your CPU. Reinstall the latest update and make sure you are on the latest BIOS.

Leave a reply to artyom Cancel reply