GA-EX58 and GA-X58A DSDT native power management modifications

jhrfc · January 17, 2018

No issues here. 1.6GHz (x12) on all cores at idle.

Thanks! actually, mine is now idling properly, i think it may have been spotlight indexing or something, happening after my upgrade.

TheBloke · February 12, 2018

Just wanted to add another success to this thread, and say thanks so much to everyone who contributed so much amazing info. Especially mcbeth303 for posting a correctly edited DSDT for the GA-X58A-UD3R, and d00d for, well, everything!

I'm running on the X58A-UD3R Rev 2.0 with a Xeon X5670 OC'd to 4.4ghz (22 x 200, no turbo), in macOS 10.13.3. I've had a Hackintosh install on and off on this hardware for years, but today is the first day that absolutely everything worked right. When I first installed, originally on Lion or Mountain Lion, most things worked, so beyond [url="http://www.insanelymac.com/forum/topic/279450-why-insanelymac-does-not-support-tonymacx86/"]#####[/url] I didn't bother doing much more research or work. That continued up to El Capitan, until in the middle of last year I updated Clover to prepare for Sierra. Then everything went haywire - KP's on boot and the BIOS trashed every time I restarted. Due to pressing life stuff I just abandoned macOS at that point. Until this week.

Now I have a system updated to High Sierra that I can boot, reboot, sleep and wake to my heart's content, with hardware/temperature/fan monitoring and working power saving/speedstep. The full works!

In case it helps any other newbies, the stuff I needed to do for my X58A-UD3R was:

[url="http://www.insanelymac.com/forum/topic/279450-why-insanelymac-does-not-support-tonymacx86/"]#####[/url] run with Legacy quick settings
Installed RehabMan's latest FakeSMSC set into EFI/CLOVER/kexts/Other - including the extra files for HWMonitor. Downloaded from here.
Then in Clover Configurator I edited config.plist as follows:
- under ACPI: enable FixUSB and FixDarwin (without which I got a USB-related KP on boot since I updated Clover in mid-2017)
- also under ACPI, I selected "Generate P States"
- and under System Parameters, enabled NVidiaWeb to enable the drivers for my 980Ti (after installing them, of course)
Then I installed mcbeth303's DSDT.aml in EFI/CLOVER/ACPI/patched

I initially allso tried mcbeth303's SSDT file, but this did not give me full performance: the CPU was locked at a 12x multiplier. The same thing happened with no SSDT at all, and it still happened after I tried to edit his SSDT.aml to add some more states for my slightly faster CPU.

But Clover's "Generate P States" worked immediately, and I guess is the standard way now? Is there any reason one might need an SSDT as well on these motherboards?

The steps I ran were on a system that originally ran El Capitan and was then upgraded first to a later Clover and finally to 10.13 using the direct App Store upgrader. A fresh install of 10.13 might not need all the steps - eg maybe [url="http://www.insanelymac.com/forum/topic/279450-why-insanelymac-does-not-support-tonymacx86/"]#####[/url] will tick FixUSB and FixDarwin if it's run now, where I had to add them manually after upgrading Clover.

My Geekbench 4 score is 3777 for single processor, 18439 for all six cores. This compares to 3724 and 17787 respectively under Windows 10; not a bad increase on the multi-threaded score!

Thanks again to everyone who contributed to this amazing thread.

TheBloke · February 12, 2018

Hardware: X58A-UD3R, rev.: 2.0, Bios FH; Xeon W3680; 12GB 1600 DDR3; GTX770 on native Drivers; USB 3.0 on native drivers and an ethernet card that runs native aswell. HDA DSDT edits and Toleda script. Chameleon, Yosemite 10.10.3

This is the last thing I am needing to get working - USB 3. I've searched the thread and see mention of PCIe cards from eg Orico. However when I look at those cards on Amazon, they all say they're only compatible with OSX 10.9.5 or earlier. They specifically state they won't work with 10.10 or later.

Maybe this doesn't apply on a Hackintosh though? Or what card are you guys using to get working USB3 on 10.10 or later? Specifically I need support for 10.13 but it sounds like 10.10 was when something changed regarding USB 3 compatibility.

Right now I have my onboard USB3, which does nothing at all, and an PCIe USB3 card from Anker, which uses a VIA chipset I believe. This is a bit odd: it mostly doesn't work. Plugging in a mouse or bluetooth adapter gives nothing. But if I plug in a USB stick to one of these USB3 slots, it mounts and functions OK! So there is some functionality on those ports, which I find very odd.

I'd be most grateful for any advice on getting USB 3 on my X58A-UD3R. I don't mind buying a PCIe card if that's what I need to do, but given the cards I see mentioned years ago in this thread seem to state they won't work in latest macOS, I'd be glad for some confirmation before I buy.

purpleorangebeach · February 12, 2018

This is the last thing I am needing to get working - USB 3. I've searched the thread and see mention of PCIe cards from eg Orico. However when I look at those cards on Amazon, they all say they're only compatible with OSX 10.9.5 or earlier. They specifically state they won't work with 10.10 or later.

I’m using an Orico card in my system. Works OOB in High Sierra. Details in signature.

Sent from my iPhone using Tapatalk

TheBloke · February 12, 2018

I’m using an Orico card in my system. Works OOB in High Sierra. Details in signature.

Sent from my iPhone using Tapatalk

Sweet, thanks very much! They're cheap, too. I'm going to order one right now. Thanks for replying so fast, really appreciated.

TheBloke · February 12, 2018

One more thing: regarding speedstep/P-states, I'm noticing that I only appear to have two available multipliers. Watching HWMonitor, it seems that all cores either run at a 12x multiplier or 22x. Nothing in-between. And either they're all at x12 or they're all at x22, I never get some at one speed and some at another (which I definitely did see sometimes in Windows.)

I'm on X58A-UD3R with X5670, block clock 200 with 22 multiplier and turbo disabled. Voltage set to Normal with a dynamic voltage added. Level 2 Load-Line calibration. Under CPU Core features, I have enabled "C3/C6/C7", "C1E", and "CPU EIST". Only one disabled is "CPU Thermal Monitor".

I'm not using an SSDT, instead I have "Generate P States" enabled in the latest Clover. I also tried adding "Generate C States" as well. I did initially try using an SSDT, but that caused me to only ever run at 12x. Specifically I used the SSDT provided by macbeth303, and later I also tried editing and compiling it myself to increase the max speed as my processor is running a little faster than his. But I stopped using it and went to Clover config instead, which at least got me the choice of x12 and x22.

I've read through the OP quite thoroughly and I think I'm right in saying that Clover's Generate C/P states are meant to be the working solution now, which is what I have. That certainly did work to get me speedstep, I just can't get a range of different speeds. The OP's later edits talks about various P State mods no longer being needed as of a particular Clover release.

I've checked the DSDT I'm using (provided by mcbeth303) and it doesn't have any P-State editing. In my active DSDT the CPUs are listed like this:




  Scope (_PR)

    {

        Processor (CPU0, 0x00, 0x00000410, 0x06) {}

        Processor (CPU1, 0x01, 0x00000410, 0x06) {}

        Processor (CPU2, 0x02, 0x00000410, 0x06) {}

...

So as far as my level of knowledge goes, I think I'm set up OK to use Clover's Generate C/P states. But I don't know why this gives me only two multipliers. It's certainly not the end of the world only having 12x and 22x, but it would be nice to have more granularity if it's possible.

Thanks in advance.

pkdesign · February 12, 2018

Ordered mine at Aliexpress.

https://www.aliexpress.com/item/External-4-USB3-0-ports-PCI-e-x1-Card-PCI-express-to-USB-3-0-SATA/32391196624.html

Been working great and only $15.

The key is that this card is based on Fresco Logic FL1100 chip.

Emilo · February 12, 2018

So as far as my level of knowledge goes, I think I'm set up OK to use Clover's Generate C/P states. But I don't know why this gives me only two multipliers. It's certainly not the end of the world only having 12x and 22x, but it would be nice to have more granularity if it's possible.

To get all multipliers and power states you must enable "Intel® Turbo Boost" in BIOS.

You will find more info in the following post: http://www.insanelymac.com/forum/topic/319980-overclocking-6-core-cpu-on-ga-ex58-ud5/?p=2345859

Here is some info on peripherals that work natively OOTB: http://www.insanelymac.com/forum/topic/303804-usb-3-sata-3-wifibluetooth-4-on-older-motherboards/

TheBloke · February 12, 2018

To get all multipliers and power states you must enable "Intel® Turbo Boost" in BIOS.

You will find more info in the following posts:

http://www.insanelymac.com/forum/topic/319980-overclocking-6-core-cpu-on-ga-ex58-ud5/?p=2345859

or

http://www.insanelymac.com/forum/topic/196771-ga-ex58-and-ga-x58a-dsdt-native-power-management-modifications/page-77?do=findComment&comment=2336714

Here is some info on peripherals that work natively OOTB: http://www.insanelymac.com/forum/topic/303804-usb-3-sata-3-wifibluetooth-4-on-older-motherboards/

Thanks so much for the reply, and for those awesome guides you wrote. Very helpful.

Wow you have 20874 Geekbench 3 score from 4.5Ghz with Turbo. I'm at 18439 on GB4, but I have 4.4Ghz without Turbo, so all cores can run at 4.4ghz. I thought that meant I would get a higher multi-core speed but lower single-core. In fact my multi-core is lower than yours, but my single-core is higher (3777 for me, 3644 for you). Although mine is GB4 and yours is GB3, so that might account for a slight increase in my single-core speed.

But it's interesting you still get higher multi-core than me despite having a lower non-turbo speed - your GB3 says you are at 4.1ghz without Turbo. Unless your UD5 mobo is somehow faster than my UD3R or something. We both have X5670 and 24GB DDR3-1600. Do you know what you get in latest GB4 with that X5670 system?

OK so I need Turbo for all multipliers. Hmm, I will have to think about that. I spent many hours/days re-OCing a few months ago and I did try Turbo for a while, but in the end I found I could get higher overall speeds with Turbo disabled. At least I thought so - now I see your GB3 score, I'm not so sure!

Also your BIOS settings are way different to mine - like you have most of your non-CPU voltages to Auto! Wow, I had to tweak mine way up:

 
Load Line Calibration ............... Level 2
CPU Vcore ................... Normal
Dynamic Vcore (DVID) ............ +0.28125
QPI/VTT Voltage 1.175v .............. [Auto]  1.355
CPU PLL 1.800v ......................... [Auto]  1.880
 
MCH/ICH
PCIE 1.500v ......................................... [Auto]  Normal
QPI PLL 1.100v .................................... [Auto]  1.200
IOH Core 1.100v .................................. [Auto]  1.300
ICH I/O 1.500v ..................................... [Auto]  Normal
ICH Core 1.1v ...................................... [Auto]  1.300

DRAM
DRAM Voltage 1.500V .................. [Auto]  1.68
DRAM Termination 0.750v ........... [Auto]  0.840
Ch-A Data VRef. 0.750v ............... [Auto] 0.840
Ch-B Data VRef. 0.750v ............... [Auto] 0.840
Ch-C Data VRef. 0.750v ............... [Auto] 0.840
Ch-A Address VRef. 0.750v .......... [Auto] 0.840
Ch-B Address VRef. 0.750v .......... [Auto] 0.840
Ch-C Address VRef. 0.750v .......... [Auto] 0.840

Obviously there are very different ways to do this. When I first started OCing this board, years ago, I was told that Auto was bad and to use Normal for everything and then tweak up voltages from there. Clearly you are doing the opposite and it's working fine, maybe even better than it is for me.

One thing I think I do have wrong - I have my Uncore set to x18, as I thought it was meant to be half QPI which I have at x36. You list it as double SPD +1, which is 17x. Interesting you also have QPI to Auto, so I am running my QPI faster than you (7.2 vs 6.4), but still getting lower benchmark results!

OK so maybe I need to go right back to step 1 with this OC and see if I can replicate what you have I thought non Turbo was going to be faster, but your GB scores are telling me otherwise, and if that also gets me full P-states it sounds like I am much better trying to replicate your OC than the one I have now.

Thanks again for all the info.

TheBloke · February 13, 2018

I just upgraded beyond 24GB, which until recently I didn't realise was possible:

Though I do now see it has been discussed before

Updated my machine from 6x2GB Corsair XMS3 to 6x8GB Corsair XMS3, all 48GB detected in BIOS, Windows7, Clover 3763 and macOS 10.12.2.

In order to test it I yanked 3 x crappy Kingston 4GB DDR3-1066 chips out of my secondary desktop and added them to the 3 x 8GB DDR3-1600 Corsairs I had already. And after a couple of false-negative tests that turned out to be me knocking the existing 8GB chips a little loose when I added the new 4GB.. success.

I'm not going to stick using these particular chips, both because that leaves my spare desktop at only 4GB RAM, and more importantly as my overclock is running RAM at 1600mhz and 1.68v, and it might well kill these cheap 1066mhz chips. Or at least contribute to crashes.

Now I know it works I am going to buy 3 x 4GB Corsair DDR3-1600s to take my permanent RAM to 36GB. Of course 48GB would be nicer, but RAM prices seem to be insane again at the moment and I can't afford £150-£200 to buy another 3 x 8GB chips (and those are used prices :w00t: )

But 36GB already feels like a very chunky upgrade, and I am once again amazed at how this old desktop of mine keeps on staying relevant and even getting better, now nearly eight years after I bought it

Anyway, if anyone else out there didn't realise the 24GB 'limit' is not an actual limit, I hope this helps

Emilo · February 13, 2018

Here are my scores with Geekbench 4: https://browser.geekbench.com/v4/cpu/7012455

It is 3847 for single core and 18547 for multi-core.

At the beginning I used the same overclocking approach as yours with "Intel® Turbo Boost" disabled and everything worked fine.

Later I desired for my overclock to be "by the book" and within the following guidelines:

- CPU VCore should not exceed 1.4175V at maximum

- CPU temperature should not exceed 69 C at maximum load during regular use

- All possible C/P States should be used

If your overclock gives you what you want there is no reason to change it.

TheBloke · February 13, 2018

Here are my scores with Geekbench 4: https://browser.geekbench.com/v4/cpu/7012455

It is 3847 for single core and 18547 for multi-core.

At the beginning I used the same overclocking approach as yours and everything worked fine.

Later I desired for my overclock to be "by the book" and within the following guidelines:

- CPU VCore should not exceed 1.4175V at maximum

- CPU temperature should not exceed 69 C at maximum load during regular use

- All possible C/P States should be used

If your overclock gives you what you want there is no reason to change it.

OK great, thanks. So actually our scores are very similar in GB4. Your single-core is slightly higher - no surprise with Turbo - but interestingly you're still higher on multi-core even without Turbo. I suppose that could be due to some other difference, eg in motherboard, but I would have assumed I'd have quite a big edge in multi-core due to all cores being at 4.4Ghz, and your score strongly suggests that's not the case.

And yeah, I was totally happy until I saw the lack of P-States. Not that it's a big deal at all, but it would be nice if I had more granularity in the speed control, and I'd think that would help with temperatures as well.

Plus your method of having lots of Autos just seems better in principle - why lock to a single, specific voltage if the motherboard is able to automatically select the best one which works fine? I changed them because I was told that the opposite would happen - the mobo would choose bad values and it would never be stable - but clearly that's wrong. I suppose it's possible that it works on your UD5 but not on my UD3R, but hopefully not.

Also I like that you are running your RAM at 1.5v. EDIT: Actually you have it at Auto, so maybe it goes higher? But perhaps not as high as my 1.68v. I never quite understood why I had to have my RAM at 1.68v when it's 1600mhz RAM which I am running at 1600mhz. But lower voltages than that would blue screen immediately. If I could get lower voltage it would be good for RAM longevity and temperatures, which will be even more important when I upgrade to six DIMMs.

So overall I think it will be worth my time having a go at re-tuning the OC. I can always go back to my known-good if it proves too much trouble.

One final question: I assume that you do your OC testing in Windows? How is macOS at giving useful info when an OC fails? In Windows I used "Blue Screen View" to give me the exact reason for each blue screen, and I found a forum post that told me what to change for each type of blue screen; code XX means more VID, code YY means more IOH voltage, etc. Not that the latter should apply with Auto, but still it's useful. Is it possible to diagnose OC problems in macOS in the same way, or should I stick to Windows? I'll definitely need Windows for IntelBurnTest at least.

PS. Out of interest, in your sig you list your CPU as "X5670 & i7-980x" - do you have two systems? Or you swap out the processors for different purposes?

Emilo · February 13, 2018

One final question: I assume that you do your OC testing in Windows? How is macOS at giving useful info when an OC fails? In Windows I used "Blue Screen View" to give me the exact reason for each blue screen, and I found a forum post that told me what to change for each type of blue screen; code XX means more VID, code YY means more IOH voltage, etc. Not that the latter should apply with Auto, but still it's useful. Is it possible to diagnose OC problems in macOS in the same way, or should I stick to Windows? I'll definitely need Windows for IntelBurnTest at least.

PS. Out of interest, in your sig you list your CPU as "X5670 & i7-980x" - do you have two systems? Or you swap out the processors for different purposes?

I have two identical computers with the exception of CPUs. One has Xeon X5670 and the other i7-980X. The both overclocked to 4.54GHz as explained in my guide.

When it comes to stress testing on Mac side I used Geekbench 3 Stress Test feature. On Windows I used Intel Burn Test.

My overclocking method is very simple:

At the beginning I set DVID to +0.24375 as this value would give me desired maximum CPU VCore of 1.408V.

Than I left all other voltages I could at Auto and just played with BCLK value until I reached stability.

Then using Easytune6 and Hardware Monitor on Windows I reviewed actual values of BIOS parameters that were set to Auto.

Since all values looked fine to me I decided to leave them at Auto.

RAM Notes:

In both of my computers all 6 RAM slots were always occupied.

Initially I used Corsair XMS3 RAM running at 1.65V and with "Intel® Turbo Boost" enabled in BIOS I could not obtain stability beyond 174 BCLK/4.17GHz.

Corsair XMS3 RAM is falsely advertised by many internet sellers as 1.5V type. This RAM will not work properly below 1.65V.

When I switched to Low Power CORSAIR Vengeance LP RAM and left it at default 1.5V and 9-9-9-24 timings, higher stable CPU overclock became possible.

In general I do not overclock my RAM, trying to run it at the lowest voltage as increasing RAM voltage is an obstacle in getting the best CPU overclock possible.

TheBloke · February 14, 2018

Thanks a lot Emilo, that's great info.

I used to have XM3 RAM but for the last couple of years I have been using Corsair Vengeance, but unfortunately not LP. My current 3 x 8GB DIMMS are Vengeance CL9, and I just bought 2 x used 8GB Vengeance Pro CL9 - at a lower price than 3 x 4GB would have been.

From the specs it appears Vengeance 'Pro' is completely identical to Vengeance except for a different heatsink design! Stupid 'enthusiast' gear

So I'm going with 40GB total in five DIMMs. Probably not ideal for performance - one channel will have only one DIMM - but I doubt I will notice any day-to-day difference, and I'd rather have 40GB with the option to upgrade to 48GB any time, versus 36GB and not able to upgrade further without removing/re-selling three DIMMs.

Anyway, I guess the Vengeance RAM has the same issue of not working below 1.65V, because that's exactly what I found when I tried lowering voltage. I think I didn't even manage to drop to 1.64V without blue screens, which matches your findings.

So I suppose I won't be able to do as you did and have Auto or low voltage on the RAM and will need to stick to the 1.6 range. Maybe that will limit my OC and I won't be able to replicate your settings. I will have to try and see. But if I have to stick with what I have now, 4.4ghz no Turbo and therefore only two P-States, then I can live with that. Although I don't mind going over 1.408V if necessary, so maybe that will help.

In other news my Orico card turned up and as expected I now have working USB3 - thanks again for the recommendations, guys. Oh, and I've spent my evening getting one of my spare LSI SAS/SATA cards working so I can now have lots of 6Gb/s SATA3 connections as well Not that I will need many as I have a separate NAS, but I have more SATA3 SSDs than my mobo has SATA3 ports, so this will help. Although I'm not yet sure if I will be able to boot Clover from the LSI.

TheBloke · February 14, 2018

Hey guys, have any of you ever seen this screen before? I think it's an EFI type boot screen.. except my motherboard doesn't have EFI!?

EDIT: It's not the motherboard, it's Clover generating this. Found references to some of the strings in Clover files. So I suppose Clover can't find any disks, or more likely it can't even load itself properly.

This is what I got when I tried to boot Clover from my LSI card. Specifically, I attached my OSX/Clover boot SSD to the LSI card; I moved it top of the boot order in the X58A-UD3R BIOS, then rebooted. When it booted, I saw the usual cursor that appears at top left of the screen before the Clover UI appears. Then this weird screen appeared.

What on earth is this? Is it some kind of half-EFI mode in the BIOS? It certainly knows the name of the motherboard, the BIOS revision, and CPU/RAM details. And it talks about "directing the system to continue the boot". So surely it's part of the BIOS.. but how, when I don't have EFI?

The only mention of EFI in my BIOS is an EFI CD/DVD booting option, meant for Windows 7 for example, which was set to Auto. I tried setting that to Disabled to see if it was the cause of this screen, but after rebooting I still got it instead of Clover.

So I am getting into Clover, but apparently it can't then continue the boot, hence it shows this screen. I know of other users booting from LSI controllers (RAID ones specifically), but I think it's quite likely that they're using EFI boot which I can't use.

Anyway I know this is a bit OT for this thread - I posted here because at first I thought it might be a mobo related screen.

konondoyl · February 14, 2018

This is what I got when I tried to boot Clover from my LSI card. Specifically, I attached my OSX/Clover boot SSD to the LSI card; I moved it top of the boot order in the X58A-UD3R BIOS, then rebooted. When it booted, I saw the usual cursor that appears at top left of the screen before the Clover UI appears. Then this weird screen appeared.

at first establish Clover then transfer the necessary folders, has taken out in a text editor

TheBloke · February 14, 2018

at first establish Clover then transfer the necessary folders, has taken out in a text editor

Sorry konondoyl, I don't follow? What are the necessary folders and what do I do in a text editor?

Do you mean I need to re-install Clover when the SSD is on the LSI controller?

purpleorangebeach · February 14, 2018

Sorry konondoyl, I don't follow? What are the necessary folders and what do I do in a text editor?

Do you mean I need to re-install Clover when the SSD is on the LSI drive?

That's what I would do. I've seen that screen before when, in my excitement to get up and running, I've forgotten to install Clover

TheBloke · February 14, 2018

That's what I would do. I've seen that screen before when, in my excitement to get up and running, I've forgotten to install Clover

OK yeah makes sense. I'll give it a go tonight.

Previously when I've installed Clover/OSX I used the tools that are apparently so evil and dastardly we're not supposed to talk about them here But this would be a good opportunity to learn the underlying/direct methods.

Thanks.

TheBloke · February 15, 2018

Here are my scores with Geekbench 4: https://browser.geekbench.com/v4/cpu/7012455

It is 3847 for single core and 18547 for multi-core.

At the beginning I used the same overclocking approach as yours with "Intel® Turbo Boost" disabled and everything worked fine.

Later I desired for my overclock to be "by the book" and within the following guidelines:

- CPU VCore should not exceed 1.4175V at maximum

- CPU temperature should not exceed 69 C at maximum load during regular use

- All possible C/P States should be used

If your overclock gives you what you want there is no reason to change it.

Thanks again for all the info and help. I've followed your OC guide and I'm pretty happy with the results:

Using your OC settings exactly - except for RAM voltage which I left at 1.68V as I'm on Vengeance RAM - gave me the following:
- macOS GB4: 3824 / 18425. Within less than 1% of your scores;
  - Windows is still lower, and the difference is now even greater: 3719 / 17457
  - I wonder if this is partly down to differences in GB4 between platforms rather than macOS actually running faster. But maybe it is.
- Better P-States: as you said, P-States have been improved. I now see in macOS different cores at different multipliers, where previously all were the same. I also see a few more multipliers: specifically, I have seen x12, x21, x22, x23 and x24. Which I suppose is just one extra compared to before (x21), as the latter two are Turbo states.
  - So in macOS I'm still not seeing the granularity I expected - I thought I'd see eight or ten possible states throughout the range of x12 to x24. But it's still an improvement.
  - In Windows I do see more states - I've definitely seen 3610Mhz for example, which is x19. But it's hard to monitor in Windows as HWMonitor's P-state monitoring is now partly broken for some reason (see below)

I then made one further tweak: turned off Auto SPD, and configured timings manually so I could select Command Rate 1 instead of 2. I was using CR1 on my previous OC with success, and I'd read that CR1 vs CR2 can be equivalent to lower RAM timings, eg equivalent to setting 8-8-8 instead of 9-9-9.
- This boosted my macOS GB4 multi-core, giving scores of 3821 / 18728 (an increase of about 1.5% on the multi-core figure.)
- However when I then upgraded to 40GB RAM, it dropped quite a bit, to 3803 / 17814.

I did briefly test pushing things further - eg 195 x 24 with a couple of attempts at higher +VIDs. It did boot into Windows and GB benchmark ran OK, but I got some hard freezes and invalid results in IntelBurnTest, and the temperatures were way too high; I stopped my last test when one core hit 99C. I don't have the cooling to even attempt this I think, nor is it really worth the risk and hassle to get another 5%.

Beside the different RAM voltage and now the manual SPD, I also have one other setting different to you: Virtualisation enabled. I had this enabled on my previous OC and it was fine, and I plan to use VMWare Fusion, so this will potentially be useful.

What I find interesting is how this new OC has some quite different behaviour to my previous one, in several ways:

I now see the true, expected VID values, eg maxing at 1.440V. Before, my VID never went above the high 1.3x's, and I could never understand why - several times when tweaking I added more to Dynamic VID and saw exactly the same maximum values, which seemed weird but I never investigate further.
Windows HWMonitor now shows that my max power is 108W, where before it never exceeded 95W. I didn't realise until now that that was unusual - I thought the chip simply couldn't go over 95W. But with the new settings, it can and does. This is obviously related to the first point, the VID not going as high as expected. Something in my old settings was capping it.
For some reason P-State monitoring is now broken in Windows HWMonitor: after launching it, the processor multipliers don't appear at all for at least a minute, and after that they don't update properly. They work fine in macOS HWMonitor, and Windows CPU-Z also seemed to show it OK, but it only displays one figure for the whole CPU.
My Northbridge temperatures are much lower with the new OC. With the old one, I was going above 75C quite often, and I actually put a Noctua 40mm fan on the heatsink to make sure this didn't go too much higher. That took me to an average of 68C. With the new OC, and that fan still in place, I'm seeing 55C max. So I can probably now remove that fan.

That VID capping also means that I was wrong when I previously said I was at +0.28V Dynamic. That's what I set, but clearly I never reached it. I was never quite sure on that, because I wasn't certain what the base voltage was meant to be - it seemed like every time I went into the MIB settings, it showed a different base voltage for "Normal", and whatever it showed never seemed to match the figures I saw quoted So I couldn't be sure whether the VID I saw in HWMonitor was applying the full +Dynamic or not. I just knew that I sometimes had to boost Dynamic two or three steps before I saw my actual VID go up. Now I see that actually something was capping it to the max 95W TDP of the chip.

One possible explanation for the capping is that I was running Load-Line Calibration level 2, which I suppose does more smoothing and could create a max power/VID cap? If not that I suppose it was due to my never using Auto, manually setting every voltage; one or more of those settings may have needed Auto in order to go beyond 95W.

I still find it weird that we're getting higher Multi-Core scores in GB4 with a lower non-Turbo speed. If it truly runs all cores at once, surely 6 x 4.4Ghz is better than 6 x 4.18Ghz? Perhaps it doesn't truly stress all CPUs at once, such that the Turbo sometimes has a chance to kick in. I suppose I could test that further with some non-synthetic benchmarks - some long running task, like compiling GCC or Firefox with -j12 - that could be used to verify on my old and new OC to compare against GB4's synthetic benchmarking. Then again that sounds time consuming and I am happy with what I have so it'd probably only be for academic interest. I think I will definitely stick with the new OC even if some benchmark did prove the old one to be 2% or whatever faster in some specific case.

Anyway, all in all a good evening. Thanks again for documenting all your research and experiences

Emilo · February 16, 2018

Thanks again for all the info and help. I've followed your OC guide and I'm pretty happy with the results.

I am glad that my OC guide worked for you.

Thank you for sharing your thoughts in such a detailed report.

Hopefully it will encourage others to try this method as well.

P.S. Would you please copy/paste your last post to my OC guide page.

Atomic_Sheep · February 28, 2018

TheBloke has redirected me to this thread, so I'll ask my questions here.

I have a legacy BIOS but to create my bootable flash drive, I mounted the EFI partition with Clover Configurator and put my EFI folder in there. When I boot up from my flash drive, it boots up Clover. How is this possible?

purpleorangebeach · February 28, 2018

TheBloke has redirected me to this thread, so I'll ask my questions here.

I have a legacy BIOS but to create my bootable flash drive, I mounted the EFI partition with Clover Configurator and put my EFI folder in there. When I boot up from my flash drive, it boots up Clover. How is this possible?

Clover probably isn’t running off the flash drive. Are you sure you’re booting from the flash drive? Do you have a Clover install on another drive in your machine?

Sent from my iPhone using Tapatalk

Atomic_Sheep · February 28, 2018

Yep makes no sense to me either. Clover is definitely not installed anywhere other than my USB.

Another question. When you compile DSDT, various errors/warnings pop up. Should I be aware of something in this regard? Should they all be rectified or just ignore these?

TheBloke · February 28, 2018

Guys, could someone let me know what SMBIOS' would be valid for my system? (UD3R with X5670)

I'm running MacPro 5.1 at the moment. I believe I've read that at least one iMac one should work too?

The reason I ask is that I'm trying to use a USB 3.0-to-DVI display adapter to add a fifth screen to my setup. The adapter I have uses a DisplayLink DL3000 chip and is stated to work with macOS, with 10.13 confirmed working in some Amazon reviews. DisplayLink's drivers also state support for 10.13, and were recently updated.

The device works perfectly in Windows 10, but is unusable in macOS - I get a picture on an attached monitor, but it is constantly corrupted when anything changes on the screen. I've tested many different screens, cables, resolutions, display modes and USB ports. I've described my experiences fully in this thread.

I don't have much hope to fix it, but one thing I want to try is setting a later SMBIOS. The DisplayLink macOS troubleshooting page lists a whole bunch of potential issues, many of which could describe what I see (though mine is so bad as to be unusable), but says all the major ones should be fixed in High Sierra 10.13 on devices that support Metal. It then links to Apple's Metal support page, which describes which Macs support Metal, like Mac Pros from late 2013 and iMacs from 2012.

I definitely have working Metal. But I thought maybe using an SMBIOS for a system that shouldn't support Metal is in some way stopping the DisplayLink driver - or some part of macOS itself - from using Metal, and thus causing the DisplayLink to break horribly. It's a longshot, but maybe the DisplayLink drivers specifically check for Mac model and then do or don't use Metal based on that. So my last thing to try is changing to a later SMBIOS.

TLDR: on my system can I use an SMBIOS from a 2012 iMac, or late-2013 Mac Pro? And are there any implications or downsides to doing so? (I think I read I need a special Kext to fix a black screen on MacPro 6.1, if I used that?)

TIA.

GA-EX58 and GA-X58A DSDT native power management modifications

Recommended Posts

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites