How to build your own iMac Pro [Successful Build/Extended Guide]

[kgp]

So, wouldn't it be logical to try and duplicate the results you got with rember with another memory test just to rule out a complication with rember? You should be able to duplicate the issue with another memory test, shouldn't you?

As for Apple, It's my opinion that the main reason they prefer Xeons in their "professional" machines is because their Xeons systems support ECC memory. The X299/core i9 platform does not. Interestingly, AMD's new Threadrippers do support ECC memory.

But, this issue may not be a memory issue, it could easily involve a number of facilities, even in High Sierra itself, although it looks like a memory corruption issue based on your tests. Guess I'm kind of glad I stayed with X99.

@mm2margeret,

a.) I did not say that remember is faulty, I said that memory testing programs are on general standalone programs and should not be used in line with other programs or memory extensive programs like heavy USB bombing, especially when assigning the entire RAM to the memory testing program! Thus, why would one do the same misleading approach by means of any other memory testing program? It should definitely not be the aim to replicate supposed issues by means of an ill-posed approach ( like e.g. running standalone memtest programs in line with e.g. heavy USB bombing), which could be the source all problems and issues!

b.) Why just the supposed USB issue, which I cannot replicate or detect on my system in any case (without running a memtest in line with e.g. heavy USB bombing), should cause a RAM memory allocation error? But else, if you run remember, or any other memtest tool in standalone mode as generally recommended, the latter program does not detect any memory problem at all? If there would be a severe RAM memory allocation problem on X299 systems due to non-ECC RAM, wouldn't remember always detect errors, even when running the latter memetest just standalone?

c.) Can you tell me, which X99 board is using ECC RAM? Moreover, X99 really does have a memory allocation problem, that's why on X99 system you need to use OsxAptioFix2Drv-free2000.efi to bypass the memory allocation problem, which by the way has nothing to do with ECC or non-ECC Ram in any case. Just read the explication of Nick Woodhams, why to use OsxAptioFix2Drv-free2000.efi to avoid random reboots on X99 systems! Without OsxAptioFix2Drv-free2000.efi, one would indeed permanently witness random reboots an X99 systems due to memory allocation errors!

Thus, now.. do you witness any random reboots on X299 when just using OsxAptioFix2Drv.efi instead of OsxAptioFix2Drv-free2000.efi? I guess, the answer is clearly "NO"!

Thus, I really don't understand what you are talking about, sorry! Read and inform yourself primer to arguing with statements based on wrong assumptions, misinformation or lack of knowledge. Up to my knowledge, you built your X99 systems by means of my X99 guide ( a guide that grew over nearly 2 years and reached more and more sophistication and now is the base of your fully functional X99 system), but you question my statements and conclusions with respect to X299, without having any personal experience with X299 systems? You know that I have both kind of systems (X99 and X299) up and running, and you can be sure that I tested what I propose in my guides a hundred times more than any other did. So please, what are you telling me here, estimated lady?

My proposal: Just try to reproduce the supposed USB issues on your X99 system by using remember in line with heavy USB bombing and then continue arguing here in this thread about X299 with a little bit more background and knowledge. Also try to use your X99 system with OsxAptioFixDrv.efi or OsxAptioFix2Drv.efi instead of OsxAptioFix2Drv-free2000.efi and you will finally understand what it means to have RAM memory allocation errors .. It appears just as lack of knowledge to spread the misleading opinion that X99 systems are free of errors and likely the much better choice when compared with X299 solutions.

I am really tired of all these absurd discussions...

Cheers,

KGP

 

[mthenders]

@mthenders

First, don't worry about the scrolling console messages. This is normal if you boot with the -v kernel option (verbose).

Is your KP before or after a row of +++++++?

And here are some more suggestions for you to check:

1. Double check the BIOS.
2. Have you tried different versions of OsxAptio?
3. Do you have FakeSMC?
4. Do you have VoodooTSCSync?
5. Are all inject kext in Clover's Other (not 10.13)?
6. What are your kernel boot options besides -v?

These are all things that can create such a KP.

Good luck.

KP is after a long row of +, which is preceded by the message:

"OsxAptioFix2Drv: Starting overrides for \.IABootFiles\boot.efi
Using reloc block: no, hibernate wake: no"

1. For every attempt, I'm reinstalling BIOS 0802 and configuring it according to section B1 of the guide.
2. Not yet.
3. Yes, it's including in KGP's "EFI-X299-10.13.1-Release-061117.zip" attached to the OP.
4. Yes, I add it to "/EFI/CLOVER/kexts/Other". I've set the IOCPUNumber key of its plist to 27, as my CPU has 14 cores.
5. Not sure what this means, exactly, but in KPG's archive, the "/EFI/CLOVER/kexts/Other" folder contains "FakeSMC.kext" and "IntelMausiEthernet.kext", and per the guide, I've added "VoodooTSCSync.kext". There is no "10.13" folder in "/EFI/CLOVER/kexts/".
6. npci=0x2000, slide=0, dart=0, debug=0x100, keepsysm=1

Thanks!

 

[kgp]

KP is after a long row of +, which is preceded by the message:

"OsxAptioFix2Drv: Starting overrides for \.IABootFiles\boot.efi
Using reloc block: no, hibernate wake: no"

1. For every attempt, I'm reinstalling BIOS 0802 and configuring it according to section B1 of the guide.
2. Not yet.
3. Yes, it's including in KGP's "EFI-X299-10.13.1-Release-061117.zip" attached to the OP.
4. Yes, I add it to "/EFI/CLOVER/kexts/Other". I've set the IOCPUNumber key of its plist to 27, as my CPU has 14 cores.
5. Not sure what this means, exactly, but in KPG's archive, the "/EFI/CLOVER/kexts/Other" folder contains "FakeSMC.kext" and "IntelMausiEthernet.kext", and per the guide, I've added "VoodooTSCSync.kext". There is no "10.13" folder in "/EFI/CLOVER/kexts/".
6. npci=0x2000, slide=0, dart=0, debug=0x100, keepsysm=1

Thanks!

Always the same kind of problems and issues, always the same kind of questions and comments, basically due to some tiny user errors, which finally could be identified at at the end.. pages of questions and answers due to the fact that people seem unable to carefully read an apply the proposed!

Just once more revise your BIOS settings, once more revise your EFI configuration and try to loose your particular knot which prevents you from boot. With the correct BIOS settings and with the distributed EFI-Folder, your system should boot OoB after some likely few minor adaptations and verifications! Carefully read my guide and exactly and precisely apply the proposed!

I cannot tell you more.. please understand that I am simply tired of always answering the same questions and always commenting on the same issues, ad infinitum like within a perpetum mobile...

Good luck!

 

[braindeadmac]

KP is after a long row of +, which is preceded by the message:

"OsxAptioFix2Drv: Starting overrides for \.IABootFiles\boot.efi
Using reloc block: no, hibernate wake: no"

1. For every attempt, I'm reinstalling BIOS 0802 and configuring it according to section B1 of the guide.
2. Not yet.
3. Yes, it's including in KGP's "EFI-X299-10.13.1-Release-061117.zip" attached to the OP.
4. Yes, I add it to "/EFI/CLOVER/kexts/Other". I've set the IOCPUNumber key of its plist to 27, as my CPU has 14 cores.
5. Not sure what this means, exactly, but in KPG's archive, the "/EFI/CLOVER/kexts/Other" folder contains "FakeSMC.kext" and "IntelMausiEthernet.kext", and per the guide, I've added "VoodooTSCSync.kext". There is no "10.13" folder in "/EFI/CLOVER/kexts/".
6. npci=0x2000, slide=0, dart=0, debug=0x100, keepsysm=1

Thanks!

Probably overkill to reinstall BIOS each time, I've only done that once or twice, and that was after some particularly flaky behavior.

What happens if you just try booting with OsxAptioFixDrv instead of OsxAPtioFix2Drv? You should be seeing "Using reloc block: yes, hibernate wake: no"

 

[flymovies]

Hey @flymovies, so sorry I have a long a$$ day at work, couldn't get back to you any time sooner..

1) For M2 performance, could be the board, could be the M2 itself
when you get a chance, try to use another slot for M2, (I would assume right now you using the ones under the heatsink, so try another slot which you have to mount the little stand-off metal piece) and check if the BlackMagic disk app improves the speed.. Also, if you don't mind to format that M2 card with Win 10, I found the BlackMagic Disk Test app for Windows,(attached below) see if the speed showing correctly on Window.. so in that case we know if your hardware are performance good or not
2) OC performance, Geekbench score
Try to load-up my OC profile below, see if its improve your score.
Pls let me know how it goes, thanks

To @ALL

Lately I built another ones with Core i7-7800X, sometime I turn Off/On the system and it stuck at BIOS error code 00, freezing over there and have to force shutdown the system. Power it back on and the Asus OC showing instead of 37%, now showing 14% and giving the message Safe mode for BIOS was failed the OC. Went in the BIOS and I saw the DRAM Timing got f$%^ck up. Reload the OC profile and everything back to normal.. (It never happens to my i9-7900X at all).. Same idea as thinking my board might be defective, call ASUS support (we all do, I know , blame on the hardware, first thing first lol), still in progress troubleshoot the system. If anyone having the same issue, pls keep me posted if you solve the problem, thanks in advance..

Ha, I should have mentioned this ;-) ... I have 2x times NVME installed in both onboard M2 slots. They are the same type, size and have the same firmware. I have been testing under windows with some SA Speed Tool. Let me try BlackMagic under Windows a bit later.

Last night I realised I still had to do the 10.13.1 update, because I used the other USB stick installer with 10.13. on it. The update fried my system (sigh).

But I noted something - when I upgraded from 10.13. to 10.13.1 I got a lot of afps errors during verbose boot before the system finally got stuck at boot. It reads something like "correcting size x to size y".

I am so sick of afps ... I am now trying to setup the system (again) with hfs+ only ... and will stay there. Eventually I also will install on NVME, but only later - I have ordered the Asus HYPER x16 M2 card. Maybe this will fix my NVME speed issue. I will report back, stay tuned ;-)

 

[kgp]

Probably overkill to reinstall BIOS each time, I've only done that once or twice, and that was after some particularly flaky behavior.

What happens if you just try booting with OsxAptioFixDrv instead of OsxAPtioFix2Drv? You should be seeing "Using reloc block: yes, hibernate wake: no"

Up to my knowledge, OsxAptioFixDrv does not work... random reboots..

Use it at own risk..

Why one would like to that anyway? What is the reson for?

Do you have problems with system stability using OsxAptioFix2Drv??

I DON’T!!!!

 

[paulotex]

2. Not yet.
5. Not sure what this means, exactly, but in KPG's archive, the "/EFI/CLOVER/kexts/Other" folder contains "FakeSMC.kext" and "IntelMausiEthernet.kext", and per the guide, I've added "VoodooTSCSync.kext". There is no "10.13" folder in "/EFI/CLOVER/kexts/".
Thanks!

2. So I suggest you cycle through all the possibilities: OsxAptioFixDrv, OsxAptioFix2Drv, OsxAptioFix2Drv-free2000.efi, and also the Test2.efi (together with either FixDrv or Fix2Drv). If none of them works, then go back to the recommended
OsxAptioFix2Drv-free2000.efi.

Also, between each try remember to fully shutdown the computer, not simply restart.

5. Your answer is exactly what I wanted to know: you've placed the kexts in "Other", which is where they should be. Once I decided to create the 10.13 folder and place the kexts there, and the system stopped booting properly. Hence my question.

6. Your options look good to me.

Finally, you can try to boot clover in debug mode: in the Clover screen, press "o" for options, then "binaries patching", then select "debug". Try to film this boot process. It might have some more important info to understand what might be wrong.

 

[paulotex]

But I noted something - when I upgraded from 10.13. to 10.13.1 I got a lot of afps errors during verbose boot before the system finally got stuck at boot. It reads something like "correcting size x to size y".

I got issues with apfs when I forgot to update the apfs.efi in Clover. I was using a version from 10.13 beta, and after sometime using apfs it started giving me strange errors at boot. Sometimes failing to load the kernel or the kernel cache, sometimes simply crazy apfs errors during boot. Updating the Clover apfs.efi to the one in my system fixed these problems, without need to reinstall.

I find APFS quicker to check when doing a force reboot, and also more resilient to errors. But I agree that it is a bit strange to have this extra virtual disk always mounted and to have several partitions sharing the same space on disk. I am actually considering moving my data disk (I boot from a small NVME, and most of my data is in a SSD) to APFS too.

 

[mm2margaret]

@mm2margeret,

a.) I did not say that remember is faulty, I said that memory testing programs are on general standalone programs and should not be used in line with other programs or memory extensive programs like heavy USB bombing, especially when assigning the entire RAM to the memory testing program! Thus, why would one do the same misleading approach by means of any other memory testing program? It should definitely not be the aim to replicate supposed issues by means of an ill-posed approach ( like e.g. running standalone memtest programs in line with e.g. heavy USB bombing), which could be the source all problems and issues!

b.) Why just the supposed USB issue, which I cannot replicate or detect on my system in any case (without running a memtest in line with e.g. heavy USB bombing), should cause a RAM memory allocation error? But else, if you run remember, or any other memtest tool in standalone mode as generally recommended, the latter program does not detect any memory problem at all? If there would be a severe RAM memory allocation problem on X299 systems due to non-ECC RAM, wouldn't remember always detect errors, even when running the latter memetest just standalone?

c.) Can you tell me, which X99 board is using ECC RAM? Moreover, X99 really does have a memory allocation problem, that's why on X99 system you need to use OsxAptioFix2Drv-free2000.efi to bypass the memory allocation problem, which by the way has nothing to do with ECC or non-ECC Ram in any case. Just read the explication of Nick Woodhams, why to use OsxAptioFix2Drv-free2000.efi to avoid random reboots on X99 systems! Without OsxAptioFix2Drv-free2000.efi, one would indeed permanently witness random reboots an X99 systems due to memory allocation errors!

Thus, now.. do you witness any random reboots on X299 when just using OsxAptioFix2Drv.efi instead of OsxAptioFix2Drv-free2000.efi? I guess, the answer is clearly "NO"!

Thus, I really don't understand what you are talking about, sorry! Read and inform yourself primer to arguing with statements based on wrong assumptions, misinformation or lack of knowledge. Up to my knowledge, you built your X99 systems by means of my X99 guide ( a guide that grew over nearly 2 years and reached more and more sophistication and now is the base of your fully functional X99 system), but you question my statements and conclusions with respect to X299, without having any personal experience with X299 systems? You know that I have both kind of systems (X99 and X299) up and running, and you can be sure that I tested what I propose in my guides a hundred times more than any other did. So please, what are you telling me here, estimated lady?

My proposal: Just try to reproduce the supposed USB issues on your X99 system by using remember in line with heavy USB bombing and then continue arguing here in this thread about X299 with a little bit more background and knowledge. Also try to use your X99 system with OsxAptioFixDrv.efi or OsxAptioFix2Drv.efi instead of OsxAptioFix2Drv-free2000.efi and you will finally understand what it means to have RAM memory allocation errors .. It appears just as lack of knowledge to spread the misleading opinion that X99 systems are free of errors and likely the much better choice when compared with X299 solutions.

I am really tired of all these absurd discussions...

Cheers,

KGP

KGP:

I'm not arguing (or at least I'm not trying to be argumentative), I'm trying to stimulate thoughtful discussion and understanding.

Sorry if you see my attempts to understand as a debate - it's not. In one of your other posts on another one of your threads, I highly praised you and your contributions to this community! And, I will continue to do so!

I am considering a X299 system and want to understand the issues. Also, I doubt X99 systems are completely free of errors. I will consider performing the test that's been under discussion here in this thread on my X99 system (given some time), and will report back.

It appears that there may be a small inconsistency/error in the way X299 systems approach memory allocation. This could easily be an issue in High Sierra itself (after all, it's quite new). It could be mainly, if not exclusively related to USB issues. It could be just pushing the system to the very edge of its capabilities, and that causes a failure. It could be a lot of things.

I just want to try and understand these issues before I buy.

 

[braindeadmac]

Up to my knowledge, OsxAptioFixDrv does not work... random reboots..

Use it at own risk..

Why one would like to that anyway? What is the reson for?

Do you have problems with system stability using OsxAptioFix2Drv??

I DON’T!!!!

It's not just a matter of system stability, my machine will not boot with OsxAptioFix2Drv, as I've previously mentioned. I get roughly the same crash that mthenders reports, to my recollection. I rather stubbornly tried to make OsxAptioFix2Drv boot and it simply doesn't. I have not tried again since before 10.13 SU and 10.13.1 releases. I'm working today, but I'll try again tonight with 10.13.1 and post results here. Recall please that I have a different video card, as well as a different processor than yourself, @kgp. I have NO random reboots at this time. Best uptime so far is just shy of 120 hours, and I just interrupted that to play some windows games with my kids.