MSI X299 SLI Plus and Intel 7820X: installation and post installation steps

[bbuster]

I think so for i have 1 more problem. I am able to install MacOS, but as soon as the pc needs to reboot, i get into Clover but only see the USB device. No SSD with MacOS anymore.

 

I tried to install it as GPT (guid), but also as apfs. In both cases the hdd is invisible in clover. Did i forget something?

Installation seems to work so far

 

 

BTW: I want to install Windows 10 on another SSD too. Should i install Windows 10 first?

[slurpi]

In a PM someone asked:

 

I have a fully working and stable system but I can't for the life of me get my RAM to work at it's intended speed of 3000mhz. It shows up in both Geekbench and OS X at 2133mhz after trying numerous different combos in clover (XMP =YES, XMP =1, XMP =NO ) 

 

I think I didn't mention this in the guide, but unfortunately I have the same issue. My XMP in the UEFI sets it to 2800MHz, but "About this Mac" and Geekbench both report 2133MHz. I didn't try to fix this, since I've read that memory frequency is not so important. But it is a bit frustrating. Fast memory is more expensive than slow memory, so paying extra and not getting the expected performance is a bummer.

 

If someone knows how to fix this, please let me know.

[slurpi]

 

I think so for i have 1 more problem. I am able to install MacOS, but as soon as the pc needs to reboot, i get into Clover but only see the USB device. No SSD with MacOS anymore.

 

I tried to install it as GPT (guid), but also as apfs. In both cases the hdd is invisible in clover. Did i forget something?

Installation seems to work so far

 

 

I haven't installed in a  while, so here's a detailed guide for the installation (I didn't do the disable APFS conversion part):

https://hackintosher.com/guides/macos-high-sierra-hackintosh-install-clover-walkthrough/

 

Just make sure you have an extra boot option when you reboot the first time, that says something like "Boot macOS Install from xxxx", where xxx is the name of the partition you installed to before. Perhaps you thought it was the installer again?

 

Here are a few more things to check:

 

One possibility is that you don't have apfs.efi installed in drivers64UEFI in CLOVER, and you are trying to install on an APFS partition. Make sure you have apfs.efi there. On a running system, you can copy it from /usr/standalone/i386/.

 

Another possibility is that your your ESP isn't correctly formatted. I suggest you use the Disk Utility that comes with the macOS installer to partition (make sure it is GPT) and format (HFS or APFS) the disk again and redo the installation. In one of the installation steps there is a menu with Utilities and there you can launch Disk Utility.

 

Finally, make sure you have disabled CSM in the UEFI. This makes sure that you never boot in legacy mode. Clover will disguise this issue by pretending you are booting in UEFI, but I have found that this can trigger strange errors, both in Windows and in macOS, that are very difficult to debug.

 

And note that GPT and APFS are two unrelated things. GPT is how the disk is partitioned, and APFS is how a partition is formatted.

 

BTW: I want to install Windows 10 on another SSD too. Should i install Windows 10 first?

 

 

 

Ideally yes, since Windows tends to mess up the ESP, even if it is on a different disk. However, you can simply remove the macOS disk while installing Windows, to make sure it does not see its ESP and mess with it. Later you can install Clover on both SSDs, so you can boot from either one.

 

In my case, I have my NVME with macOS, Windows and Linux. In addition I have a SSD for data, but it also has Clover there. I make sure I can boot Clover from both storage devices, and use one Clover installation for testing Clover. But I have to be really careful so that Windows upgrades (Fall Creator and big things like that) don't mess up the ESP. I keep a bootable USB available with a Linux rescue installation (gparted), so that I can boot with it and manually fix the ESP if necessary.

 

In your case, with a disk for macOS and another for Windows, things should be much smoother. 

[bbuster]

Thanks, i rememberd i used the apfs.efi, but i took it from my /usr/standalone/i386/ from my macbook pro, but now i think i can't use this because my macbook has Sierra, not High Sierra... can this be the problem?

 

Exactly, i still booted from the usb stick. First i installed OSX, the drive was formated as GPT, then i run the installer and formatted the partition as Journaled, the installer run and i rebooted into the clover usb stick. I saw my installer and also the mac ssd. I selected the mac ssd and the installation continued. Then it needed to reboot again and my ssd was gone.

 

I tried to reinstall mac osx again but now i wasn't able to use journaled, i only could format it as apfs. I did this too, but now the disk isn't showing anymore after the first reboot.

 

I have a 960 evo 250gb which i will use as Mac OSX.

I also have a 960 evo 1TB which i will use as windows, i was just thinking of maybe buy another 850 evo 250gb for windows so i can use the 1tb as storage without windows.

 

Is it difficult to install clover on 2 disks? I would like to always start in Windows except if i want to use mac (hold a key or something.. i don't know yet what is possible).

 

I already had windows installed but i messed up this by trying and doing something wrong.

Thanks again for your help! 

[Matthew82]

I haven't installed in a  while, so here's a detailed guide for the installation (I didn't do the disable APFS conversion part):

https://hackintosher.com/guides/macos-high-sierra-hackintosh-install-clover-walkthrough/

 

Just make sure you have an extra boot option when you reboot the first time, that says something like "Boot macOS Install from xxxx", where xxx is the name of the partition you installed to before. Perhaps you thought it was the installer again?

 

Here are a few more things to check:

 

One possibility is that you don't have apfs.efi installed in drivers64UEFI in CLOVER, and you are trying to install on an APFS partition. Make sure you have apfs.efi there. On a running system, you can copy it from /usr/standalone/i386/.

 

Another possibility is that your your ESP isn't correctly formatted. I suggest you use the Disk Utility that comes with the macOS installer to partition (make sure it is GPT) and format (HFS or APFS) the disk again and redo the installation. In one of the installation steps there is a menu with Utilities and there you can launch Disk Utility.

 

Finally, make sure you have disabled CSM in the UEFI. This makes sure that you never boot in legacy mode. Clover will disguise this issue by pretending you are booting in UEFI, but I have found that this can trigger strange errors, both in Windows and in macOS, that are very difficult to debug.

 

And note that GPT and APFS are two unrelated things. GPT is how the disk is partitioned, and APFS is how a partition is formatted.

 

 

 

Ideally yes, since Windows tends to mess up the ESP, even if it is on a different disk. However, you can simply remove the macOS disk while installing Windows, to make sure it does not see its ESP and mess with it. Later you can install Clover on both SSDs, so you can boot from either one.

 

In my case, I have my NVME with macOS, Windows and Linux. In addition I have a SSD for data, but it also has Clover there. I make sure I can boot Clover from both storage devices, and use one Clover installation for testing Clover. But I have to be really careful so that Windows upgrades (Fall Creator and big things like that) don't mess up the ESP. I keep a bootable USB available with a Linux rescue installation (gparted), so that I can boot with it and manually fix the ESP if necessary.

 

In your case, with a disk for macOS and another for Windows, things should be much smoother. 

Hello. A have the same cpu and Asus x299. I trying boot to OS X from yesterday but no lack. Can you share your EFI? You made same special changes in bios?

[slurpi]

I think the apfs.efi from Sierra will not work in High Sierra. During installation, go to the Utilities menu, open a terminal, and see if you get the apfs.efi from the BaseSystem image. Otherwise just google it, you'll probably find it.

[zambostraut]

In a PM someone asked:

 

 

I think I didn't mention this in the guide, but unfortunately I have the same issue. My XMP in the UEFI sets it to 2800MHz, but "About this Mac" and Geekbench both report 2133MHz. I didn't try to fix this, since I've read that memory frequency is not so important. But it is a bit frustrating. Fast memory is more expensive than slow memory, so paying extra and not getting the expected performance is a bummer.

 

If someone knows how to fix this, please let me know.

 

Thanks for posting my question. I realised I had to pass the quiz to respond in here or as my Gf put it "That's a hard CAPTCHA"

 

Anyway, I think it may be cosmetic but have seen numerous other OS X installs with ram reading the correct XMP profile (over the standard 2133mhz) so I'm assuming it's possible. I've also tried playing with the iMac Pro SMBIOS but I get mixed results when booting and have to change SMBIOS back from clover in the boot. 

 

If anyone manages to get there RAM to run at full speed I'm sure Slurpi and I would be very grateful to hear from you!

Hello. A have the same cpu and Asus x299. I trying boot to OS X from yesterday but no lack. Can you share your EFI? You made same special changes in bios?

I made three seperate USB installers with different versions of clover before attacking my High sierra Hackintosh. I found that two of them gave me the same issue you are having but I managed to get a working installation by following this guide. 

 

APFS was initally causing my problems although I managed to get it installed on a further new USB install after attaching this in my clover apfs.efi

[slurpi]

Hello. A have the same cpu and Asus x299. I trying boot to OS X from yesterday but no lack. Can you share your EFI? You made same special changes in bios?

The only change to the BIOS is reported in the OP: disable CFG. I also disabled CSM. However, the CFG thing might not be relevant for you BIOS.

 

What kind of error do you get? Saying "no luck" does not give much information. Just black screen? Do you get to the +++? Do you get a KP?

 

In another place there is an amazing thread about an Asus X299 motherboard. You should google it, you'll get much more info there.

 

Here's my config.plist, with smbios serials removed:

 

config.zip

[Matthew82]

OK. I found solution. Just add one patch for kernel "xcpm_idle patch by Pike R. Alpha" I have installed system and start testing everything today. 

[bbuster]

Hmm, i thought i had it working.

I installed MacOs on another new SSD. The installation worked and i got into Mac OSX. I installed some kext (internet worked), i was trying to install uni_beast for High Sierra. It looked liked everything was okay.

 

I tried to reboot to see if it would start without an usb stick. I removed the usb stick, rebooted the machine and got into Clover.

I selected my mac os ssd, got an apple sign for 2 seconds and then an error sign.

 

I rebooted again in the verbose mode, and now i see this message:

OsxAptioFixDrv: AllocateRelocBlock(): can not Locate relocation block (0x1AB45 pages below 0x10000000000): Not Found.

 

I then tried to reboot with my usb stick, but there is my "Does printf work??" error.

What did you do after the installation? Do you always have to use the USB stick to boot?

 

Thank you

[Matthew82]

Try with OsxAptioFix2Drv-64.efi and slide =39. This set up work for me Asus x299

[slurpi]

 

 

I rebooted again in the verbose mode, and now i see this message:

OsxAptioFixDrv: AllocateRelocBlock(): can not Locate relocation block (0x1AB45 pages below 0x10000000000): Not Found.

 

I then tried to reboot with my usb stick, but there is my "Does printf work??" error.

What did you do after the installation? Do you always have to use the USB stick to boot?

 

Thank you

 

 

Hi:

 

Try with OsxAptioFixDrv+Test2 (see the first post). It never failed me.

 

I boot from the internal storage, never from the USB. It works well for me.

 

Each time you get these OsxAptioFixDrv errors (can't allocate block or printf), remember to fully shutdown the machine and boot from cold. Don't do a warm reboot, they might trigger the error. But note that  my system with OsxAptioFixDrv+Test2 also work with reboots.

 

The slide suggestion from Matthew82 might work. However, my experience is that slide values are system dependent (MoBO, CPU, BIOS version, installed hardware), so they might not be valid on different machines. And my efforts a few days ago to calculate a slide for my system failed.

 

 

Finally, this thread seems to point to a new fix:

http://www.insanelymac.com/forum/topic/331381-aptiomemoryfix/

It has a repository here:

https://github.com/vit9696/AptioFix

Since it is all in a preliminary state, if the the OsxAptioFixDrv+Test2  works for you, I suggest you stick with it.

 

Good luck

[slurpi]

Updated to 10.13.2 (Supplemental Update) 17C205.

 

Nvidia drivers updated to  378.10.10.10.25.104.

 

Nothing to report, system seems smooth. It was possible to boot into GUI and from there install the Nvidia WebDrivers.

 

Note that there is also another macOS version 17C2205, which is only downloaded in systems with the iMacPro1,1 SMBIOS. This would need a different Nvidia WebDriver (I think the 106 version).

 

Like I said before, my system is working well with the iMac18,3 SMBIOS. I will wait for the release of 10.13.3 to change to iMacPro1,1. And I am also planning to create another installation guide for that version of the OS.

 

 

[zambostraut]

I also updated straight from the App store (Supplemental update) an system booted straight back up as if nothing had happened.

 

I was using Nvidia WebDrivers 378.10.10.10.25.102. and apparently I have no further updates available. 

 

Still haven't worked out the RAM XMP issue although I read on another thread that XMP Profile 1 may not work and instead it should be set to "Off" in bios and manually configured to run at the speed your RAM can handle. I don't have the knowledge about RAM timings as of yet and it seems there are a lot of different options in BIOS to do this so not sure if I'll try it at the moment or not. 

 

If anyone else has a shot or gets anywhere do please update your findings in here.

[zambostraut]

Here are my Benchmark scores in the attached images with the settings posted below. If anyone else on here want's to throw theirs in to the thread and Slurpi doesn't mind this content in his thread I think that would be useful to work out if the RAM speed issue is purely cosmetic, BIOS or OS X based. 

 

GeekBench SINGLE 5501 : MULTI = 35195

CineBench 1895 Best so far

Power Consumption MAX 183Watts

Temp MAX 79c

 

CPU Clock Ratio 4.4 (All Cores)

CPU Mesh Ratio 3.2

Energy Efficient Turbo Disabled

Enhanced Multi Core Enabled

Hyper-Threading Technology: Enabled

CPU EIST Function: Auto

 

 

Extreme Memory Profile (X.M.P): Profile1 (OS X Reports 2133mhz even though my RAM is set to 3000mhz and BIOS correctly reports this speed. Not sure if this is a cosmetic error of if the RAM is actually running at XMP Profile 1 true speed.)

 

 

[slurpi]

Not a problem if you want to investigate your XMP issues here. You might not have a large audience ;-)

 

My scores @ 4.4GHz are:

 

Cinebench: 1923. That's the best with 17C205. I had a slightly better one before: 1928.

​Temps at 99C, but no throttling. 

 

Geekbench: 5490 / 36053.

I also used to have a better score (a bit above 5500 / 36300).

[zambostraut]

Hey, 

 

Yes I imagine it's not going to be the most talked about subject ever but you never know. 

 

Your scores are pretty good and interesting to see the difference even though we are using the same hardware basically. Those temps are pretty crazy mind you, is that on water or just air? 

 

So we're both at 4.4ghz and it seems your scores are about 3% faster or so. Negligble I guess but the only difference could be the RAM / Motherboard. I get the feeling if our RAM was running at the correct XMP speed we'd see another 3-4% again. 

 

Think the new Spectre / Meltdown patch has taken a little off of our scores but I'll have another test today with different bios settings. Manually setting the timings has done some good things for other users but it seems like a lot of assing around for a small gain. It's addictive getting every last point out mind you.

 

Will report back in due course. 

[slurpi]

My cooling is air, the Noctua D15. Tests show that it is equivalent to an 240mm AIO (like the Corsair 115i) so I went for that, due to the simple setup. It forced me to have a big case though, because of it's height.

 

The temps are so high because the core voltage for 4.4GHz is at 1.21V. Lower voltage than (or auto) and I was getting random reboots when doing the performance tests. Cheap motherboard, I guess, even with the latest UEFI.

 

Ultimately, the thermal compound Intel used on the heat shield is really to blame, plenty of people that delid get 10C-20C lower.

 

My usual temps under load (compilation of programs, parallel processing of data using MPI) is usually 70-80C, so I'm ok with this.

 

And I also suspect it's the Meltown patch that cause the performance to be a bit lower. 

[zambostraut]

Did you change your CPU Mesh Ratio or Frequency at all? 

 

I left my CPU votalge at auto and haven't noticed any reboots or stability issues. Warmest I get is 81c at full whack 4.4Ghz on all cores. Sure I can push it a bit higher but not sure the extra 500 points on GB is worth the heat / power usage. I'd far rather get this RAM speed thing working for sure. Although it's meant to be minimal in performance gainz we both have 25 -30% extra RAM speed to get out of our systems as it stands. Other X299 builds seem to be picking up the correct speeds so no idea why we aren't. 

 

Temps below 90c are ok with me I guess and I'm really quite happy with the stability of X299 on High Sierra. I'd have to find a really good reason to bump up to the SMBIOS Mac Pro 1.1. Not sure what benefit it will bring but if you go for it and notice anything do notify. 

Cheers!
 

[slurpi]

Did you change your CPU Mesh Ratio or Frequency at all? 

I just changed the multiplier to 44 (to get 4.4GHz) and the core voltage as low as I could get to get stability. At 1.20V, the machine would reboot during the Prime95 (in the first 10 min). This is the first time I am OC (it's the first time in 10 years that I built a computer), so everything is a bit overwhelming. Changing just the multiplier and adjusting the voltage was as far as I wanted to go.

 

That's why I am also using XMP to OC the memory, and I don't do it manually. On paper my memory claims to be 3200MHz, but XMP only gives it 2800MHz, and I don't want to push it further.

 

A finally issue is the dubious quality of some X299 motherboards. It has been shown that some of them can't really support the high power draw of OC, and they can overheat. My MB is on the lower end, but it has behaved well so far (knock on wood). I think 4.4GHz is as far as I dare to go, even if I had a liquid cooling system.

 

About the iMacPro1,1, I want to give it a try. I noticed that some of the USB 3.0 ports are not recognized, and I hope with the new SMBIOS it will be fixed ;-) Mostly, I don't like to use all the DSDT + Clover patches to get a system running. The more native the better. I tried yesterday the 10.13.2 installer from the Apple store, but it still needs some fixing to install on a iMacPro SMBIOS. There are some instructions on the net on how to do that, but it is too much unnecessary work for me. Like you, I enjoy the stability and performance as it is.

 

In addition, I am hope that 10.13.3 will be easy to install on an iMacPro hack, without any special changes, so I'll wait for that.

 

Good luck with your issues. BTW, I read in another forum that the MSR 0xE2 Register on your BIOS can be patched so that it is writeable by macOS. The removes a Clover patch and allows for better performance and better sleep/wake functionality. Have you tried that yet?

[bbuster]

I'm back to zero again... 

I recreated my usb stick the way i did before, but now i alway get a "does printf work??" error.

Bios is good, usb stick is the same as before, i use the GT 710 2gb

 

Why am i getting this problem now? I tried with clover 4369, with the OsxAptioFix2Drv-64.efi, the OsxAptioFixDrv-64.efi, the Test-...efi but all give me the same error.

 

I turn off my pc completely, but it still occures

[zambostraut]

I just changed the multiplier to 44 (to get 4.4GHz) and the core voltage as low as I could get to get stability. At 1.20V, the machine would reboot during the Prime95 (in the first 10 min). This is the first time I am OC (it's the first time in 10 years that I built a computer), so everything is a bit overwhelming. Changing just the multiplier and adjusting the voltage was as far as I wanted to go.

 

That's why I am also using XMP to OC the memory, and I don't do it manually. On paper my memory claims to be 3200MHz, but XMP only gives it 2800MHz, and I don't want to push it further.

 

A finally issue is the dubious quality of some X299 motherboards. It has been shown that some of them can't really support the high power draw of OC, and they can overheat. My MB is on the lower end, but it has behaved well so far (knock on wood). I think 4.4GHz is as far as I dare to go, even if I had a liquid cooling system.

 

Hey, 

 

Yes that seems all very reasonable and I'm in the same boat in terms of thinking. I don't want to push higher than 4.4OC although from what I've read this Gigabyte board has a pretty decent VRM and doesn't get to hot. I'm really happy with how it's working right now anyway so I don't mind an extra 5-7% gain for nothing. Only that RAM issue is pissing me off but I can live with it for now. 

 

In terms of the MSR 0xE2: I'm not sure the Gigabyte boards have that issue, maybe so but even with the patched clover I'll leave it as is for now. Power consumption, temps and performance are all good and I've yet to have any sort of crash or reboot with the CPU voltage left at stock. 

 

I agree about waiting for the next version of High Sierra too. I think the iMac Pro SMBIOS might just be a little too immature to safely move to. If it doesn't show any performance enhancements I'll stick with this one. USB 3.0 / 3.1 have all worked OOTB for me at full speed too. 

 

Keep me posted if you find anything else out and I will likewise. 

I'm back to zero again... 

I recreated my usb stick the way i did before, but now i alway get a "does printf work??" error.

Bios is good, usb stick is the same as before, i use the GT 710 2gb

 

Why am i getting this problem now? I tried with clover 4369, with the OsxAptioFix2Drv-64.efi, the OsxAptioFixDrv-64.efi, the Test-...efi but all give me the same error.

 

I turn off my pc completely, but it still occures

I had that error randomly also but changed up USB slots and or USB stick. In the end I finally got a working install using a USB 3.0 stick with a standard copy of High Sierra copied using Apple's own method of dd via terminal. Then installed clover (most up-to-date) and it made it to the end. 

 

You said "I'm back to zero again" but where exactly did you get up to before going back to the start?

 

[MaLd0n]

@slurpi

 

u try a new driver from here?

http://www.insanelymac.com/forum/topic/331381-aptiomemoryfix/

AptioMemoryFix.zip

 

 

[surfinchina]

I just changed the multiplier to 44 (to get 4.4GHz) and the core voltage as low as I could get to get stability. At 1.20V, the machine would reboot during the Prime95 (in the first 10 min). This is the first time I am OC (it's the first time in 10 years that I built a computer), so everything is a bit overwhelming. Changing just the multiplier and adjusting the voltage was as far as I wanted to go.

 

That's why I am also using XMP to OC the memory, and I don't do it manually. On paper my memory claims to be 3200MHz, but XMP only gives it 2800MHz, and I don't want to push it further.

 

A finally issue is the dubious quality of some X299 motherboards. It has been shown that some of them can't really support the high power draw of OC, and they can overheat. My MB is on the lower end, but it has behaved well so far (knock on wood). I think 4.4GHz is as far as I dare to go, even if I had a liquid cooling system.

 

About the iMacPro1,1, I want to give it a try. I noticed that some of the USB 3.0 ports are not recognized, and I hope with the new SMBIOS it will be fixed ;-) Mostly, I don't like to use all the DSDT + Clover patches to get a system running. The more native the better. I tried yesterday the 10.13.2 installer from the Apple store, but it still needs some fixing to install on a iMacPro SMBIOS. There are some instructions on the net on how to do that, but it is too much unnecessary work for me. Like you, I enjoy the stability and performance as it is.

 

In addition, I am hope that 10.13.3 will be easy to install on an iMacPro hack, without any special changes, so I'll wait for that.

 

Good luck with your issues. BTW, I read in another forum that the MSR 0xE2 Register on your BIOS can be patched so that it is writeable by macOS. The removes a Clover patch and allows for better performance and better sleep/wake functionality. Have you tried that yet?

 

Overclocking the X299 motherboard is great compared to the older ones.

For 4.4 you can just change the multiplier and nothing else. For a higher overclock you can change volts but I only bother over 4.6.

Here is my overclock guide for Gigabyte users.

Leave everything else, including mesh, alone. Apart from XMP of course.

Prime 95 was designed for less cores so in my opinion it's not good to run. Too stressful for the modern CPUs. AIDA is better, but not CPU and AVX together.

If you must run Prime 95 offset your AVX by 2, 512 by 3.

If you only have an 8 pin CPU power cable then don't overclock. 12 pins is ok, 16 is best.

 

I've had nothing but trouble with imac pro 1,1. It slowed my computer down by heaps and won't let me update OS x by app store. I'd wait until it's solid before jumping in. There are no advantages and a lot of issues right now. I've spent a LOT of time with this and so far no good.

 

Only Asus motherboards have MSR locked and require bios patching so far as I know. All other motherboards are fine, so either keep away from Asus or get comfortable with bios patching

 

Hope this helps

Pete

[bbuster]

 

I had that error randomly also but changed up USB slots and or USB stick. In the end I finally got a working install using a USB 3.0 stick with a standard copy of High Sierra copied using Apple's own method of dd via terminal. Then installed clover (most up-to-date) and it made it to the end. 

 

You said "I'm back to zero again" but where exactly did you get up to before going back to the start?

 

 

Last week i managed to complete the installation, but i didn't know exactly how and what bootloader i had to install (now i think i need to install clover again but now not on the usb stick but on the hackintosh ssd. After that my pc wouldn't boot anymore and i got this error: OsxAptioFixDrv: AllocateRelocBlock(): can not Locate relocation block (0x1AB45 pages below 0x10000000000): Not Found.

 

 

I tried to recreate the setup usb stick because with the same usb stick i still got the same error trying to install. Since then i can't get it to install anymore. The only difference is that i now bought a samsung 850 evo instead of installing it to the 960 evo (which i want to keep using on windows).

But i don't think it does anything with that already