Asus X299 - Support

[izo1]

try attached (that's the last option, if this won't work I don't know hot to help you)

Thank you for that. It didn't work, however I finally figured it out after trying a ton of drivers and spending many hours out of frustration. Used a Broadcom 2070 driver and BT was recognized instantly along with all the previous devices linked before the 1903 update, including the Apple space gray keyboard.

Attaching full driver (both working Bluetooth and Wifi) in case anyone comes across this problem with Windows 10 build 1903.

These are for the BCM943602CS module.

 

[astroguy]

Dude you're missing the TSCadjustreset in your kext. Attached is the proper one for your 9980XE. Also, your config plist is a bit messy. the ACPI patches are missing infos. You do not need all those kernel patches. .Why are you using USBInjectall?? You have not filled the Cpu type 0x0f01 in your config and I'm pretty sure I'm missing buch of other infos Someone with Asus Prime X299-Deluxe II can post his EFI with the proper adapted kext for your board. Or you can read the guide ...Sorry I do not have more time to re write your config.
https://www.tonymacx86.com/threads/...ojave-successful-build-extended-guide.255082/ Good luck mate.

Okay, I was hoping I could get away with using what worked for my other computer on this one, with small modifications, despite it being a different system. I started from scratch with the above post and have successfully installed 10.14.4. Seems to be working.

However, there is one critical exception: Disk I/O seems to be off. In my Hackintosh2, I have a a WD SSD as the system disk, and a Samsung 970 Evo as a data drive for my work that requires fast I/O. On Hackintosh3 (the one that's the subject of this post, the new system), I have a Samsung 860 Evo as the SSD system disk, and the exact same 970 Evo for a data drive. Both formatted as HFS+ (though I have tried to fix this problem with formatting as APFS and it didn't change anything). Using Blackmagic Disk Speed Test, I get around 2500/2700 MB/s on the Hackintosh2, and 2300/2700 MB/s on Hackintosh3. Slightly disappointing, but I can live with it.

The issue comes when I do my benchmark for work-work. This involves reading/writing many small files (~MB) and some large files (GB) and doing stuff in RAM. On Hackintosh2, there are no issues, where the test runs as expected, and iStat Menus shows all 12 cores running and roughly 95% user 5% system. However, on Hackintosh3, I get stuck and instead of 36 cores going full-blast, I end up running at 55% capacity, with 30% taken up by the system doing ... something? (Activity Monitor doesn't say) and only 15% seemingly running my code during the most parallelized parts. The disk I/O for these tests is MINIMAL and doesn't come anywhere close to the limit (like, by a factor of 100x).

I thought it might be because it's 36 cores versus 12, but that's not the case: I then moved my files to the SSD instead of the NVMe. The same test ran in 2/3 the time on the SSD as on the NVMe on the same system. It STILL took about 15% longer than on my system at work, and some parts still went down to 30% system 15% user, but only in some areas of my code (despite the exact same steps using 100% about 20 minutes earlier). I'm also not hitting a thermal limit, for I took off the OC and it's running at stock speed and capping out at 65°C. (Note: The same system at work (same motherboard, CPU, and SSD/NVMe) runs these work benchmarks more quickly with Unix as the OS. It's a mixture of Python and code written by the US Geologic Survey.)

I also checked the quoted X299 install guide and made sure the SSD/NVMe TRIM Support kext to patch is enabled, and it is, and I've tried restarting.

Any ideas? I built this system primarily to run this code at this point in time, so if I have to, I'll reformat everything and install Unix, but I'd greatly prefer not to.

 

[Ellybz]

Okay, I was hoping I could get away with using what worked for my other computer on this one, with small modifications, despite it being a different system. I started from scratch with the above post and have successfully installed 10.14.4. Seems to be working.

However, there is one critical exception: Disk I/O seems to be off. In my Hackintosh2, I have a a WD SSD as the system disk, and a Samsung 970 Evo as a data drive for my work that requires fast I/O. On Hackintosh3 (the one that's the subject of this post, the new system), I have a Samsung 860 Evo as the SSD system disk, and the exact same 970 Evo for a data drive. Both formatted as HFS+ (though I have tried to fix this problem with formatting as APFS and it didn't change anything). Using Blackmagic Disk Speed Test, I get around 2500/2700 MB/s on the Hackintosh2, and 2300/2700 MB/s on Hackintosh3. Slightly disappointing, but I can live with it.

The issue comes when I do my benchmark for work-work. This involves reading/writing many small files (~MB) and some large files (GB) and doing stuff in RAM. On Hackintosh2, there are no issues, where the test runs as expected, and iStat Menus shows all 12 cores running and roughly 95% user 5% system. However, on Hackintosh3, I get stuck and instead of 36 cores going full-blast, I end up running at 55% capacity, with 30% taken up by the system doing ... something? (Activity Monitor doesn't say) and only 15% seemingly running my code during the most parallelized parts. The disk I/O for these tests is MINIMAL and doesn't come anywhere close to the limit (like, by a factor of 100x).

I thought it might be because it's 36 cores versus 12, but that's not the case: I then moved my files to the SSD instead of the NVMe. The same test ran in 2/3 the time on the SSD as on the NVMe on the same system. It STILL took about 15% longer than on my system at work, and some parts still went down to 30% system 15% user, but only in some areas of my code (despite the exact same steps using 100% about 20 minutes earlier). I'm also not hitting a thermal limit, for I took off the OC and it's running at stock speed and capping out at 65°C. (Note: The same system at work (same motherboard, CPU, and SSD/NVMe) runs these work benchmarks more quickly with Unix as the OS. It's a mixture of Python and code written by the US Geologic Survey.)

I also checked the quoted X299 install guide and made sure the SSD/NVMe TRIM Support kext to patch is enabled, and it is, and I've tried restarting.

Any ideas? I built this system primarily to run this code at this point in time, so if I have to, I'll reformat everything and install Unix, but I'd greatly prefer not to.

Just follow the guide friend. From the length of it It looks more complex than what it really is. You won't have any issue. You can skip the SSDTs part if you do not wish to install them ( it's only a cosmetic option inside: about this mac /System report/PCI ) Everything else is inevitable. I could post my EFI but it wont help you as I use specific SSDTs & usb kext for my Gigabyte board. thats why I suggested user with same config as you to post theirs ( If you do not want to read the guide ). again, best of luck

 

[astroguy]

Just follow the guide friend. From the length of it It looks more complex than what it really is. You won't have any issue. You can skip the SSDTs part if you do not wish to install them ( it's only a cosmetic option inside: about this mac /System report/PCI ) Everything else is inevitable. I could post my EFI but it wont help you as I use specific SSDTs & usb kext for my Gigabyte board. thats why I suggested user with same config as you to post theirs ( If you do not want to read the guide ). again, best of luck

I've gone through the guide. I followed C.4 (TSCAdjustReset customization), D.1 to make the EFI folder, D.2-.3 I did not follow because I already have the installer, and did a version of D.4-.5 to install (I installed from the USB and made the EFI folder on the SSD after install). I have the xcpm_core_scope_msrs enabled and PluginType checked, verified via the test in E.1.a, and I followed the rest of E.1 and got the responses that he indicated. I skipped E.2 because the GPU is working. I skipped E.3 because I do not have any audio-out on this computer (it's not for that). E.4 is redundant because the kexts are handled in D.1, and the USB ports are generally working (at least the ones I tested) so I have not gone through to make my custom map, but I will later if needed. That takes me to E.5-.6 on the NVMe, and E.5 says it should work, and I verified the patch is implemented in E.6 and nothing changed. I did not do E.7 because I don't have that GPU. I then followed some of E.8 but the one ethernet port is working fine which is enough for me. I just now went all through E.9.1 (you said I didn't need to do E.9.2) and I still have the same issue.

HOWEVER, I just did a bit more testing. It looks like some of this might be an issue with the software I'm calling within my Python scripts, which is very weird because this would be the only computer that has issues, and it's running the same install versions of Python. I tried running some of the individual programs (not written by me) that my Python script calls. Most of them give expected behavior on the SSD and NVMe and use all CPUs with single-digits % system. However, one of the programs that computes statistics on ~MB size files drives CPU usage down to ~40%, with only 10% user and 30% system. On both the NVMe and SSD. BUT, it runs 2x more slowly on the NVMe than the SSD, despite the same CPU speed reported. There is also very little disk I/O during this step because, as I said, it's ~MB-sized files.

While I'm tempted to say, then, that the Hackintosh is working as a Hackintosh, I don't understand why the program (and just that program, apparently) would run more slowly on the NVMe versus SSD if it's not perhaps a Hackintosh issue (and ran quickly on the Unix install on the same build at work). But this would account for why my overall benchmark ran more slowly than at work off the SSD (that stats program is called A LOT) and more slowly on the NVMe (because that stats program is called a lot and takes 2x longer). It's also making much less sense at 2am.

 

[JimmyS]

if your USB kext is configured to make your internal usb 2.0 header to "255" position than it says to macOS to totally shut down that USB port (so the BT attached). You can change that number to "0" but it could prevent your sleep wake functionalities.

Thanks, I changed the port to "0", but unfortunately it wakes instantly after putting to sleep. Oh well...

AHAHAH 2 ANSWERS IN 1 POST!!! I was a total Hackintosh noob 10 months ago, now I'm goin pretty great!!!

 

[izo1]

Thanks, I changed the port to "0", but unfortunately it wakes instantly after putting to sleep. Oh well...

Sleep has always been a hit and a miss for me and for most people in general that Hackintosh. Especially when you start using USB devices.

This is a non-issue on a real Mac. For example, I have an iMac and a Mac Pro 2013 close by and we always have hard drives connected and it's set to auto sleep on idle, it works perfect. The real Macs wake up and wake the external drives back up.

On a Hackintosh it's 50/50 and as much as tweaking you can do it is never guaranteed to work.

For example if I connect all my external drives to the Hacky, and set system to sleep, the drives will eject themselves or at some random time the Hacky will wake up automatically. And sometimes the fan curves I set in the BIOS don't come back as they're supposed to.

So I just gave up on sleep over the years on Hackys and decided just to use more premium fans/coolers (for longevity) and leave the computer on and just let the monitors sleep on idle. Works perfect and I can come wake it up with a simple keyboard or mouse action.

I also set the left bottom Hot Corner to "Display Sleep" so I just swipe my mouse to the bottom left and get off my workstation and it auto locks and sets the displays to sleep and they go on idle. I come back and sit down it auto unlocks via Apple Watch or via Password. I also set a 10 minute timer for the displays to auto sleep and the system to lock for safety reasons.

And also since I'm using big radiators with lots of fans via push/pull and a custom watercooling solution, I can run the fans at a very low speed so even on idle it doesn't collect a ton of dust or make any audible noise.

Anyway, long story short, I think it's a waste of time trying to figure out sleep on a Hacky. You might get it to work if you never use any BT/WIFI cards (BT uses USB and Wifi PCIe) or even any USB device....but that means some type of virgin system which doesn't have a real world scenario setup....if that makes sense.

 

[djlild7hina]

Thanks, I changed the port to "0", but unfortunately it wakes instantly after putting to sleep. Oh well...

What cpu cooler are you using? Could be a usb device preventing sleep.

 

[kubusikan]

Has anyone been able to add an eject icon next to an external USB drive? I can right click and eject disk no problem, but the icon is not there.

Do you mean the litte eject icon in the left finder bar where all the drives are listed?
For me the icon shows up for external drives.

 

[JimmyS]

Sleep has always been a hit and a miss for me and for most people in general that Hackintosh. Especially when you start using USB devices.

This is a non-issue on a real Mac. For example, I have an iMac and a Mac Pro 2013 close by and we always have hard drives connected and it's set to auto sleep on idle, it works perfect. The real Macs wake up and wake the external drives back up.

On a Hackintosh it's 50/50 and as much as tweaking you can do it is never guaranteed to work.

For example if I connect all my external drives to the Hacky, and set system to sleep, the drives will eject themselves or at some random time the Hacky will wake up automatically. And sometimes the fan curves I set in the BIOS don't come back as they're supposed to.

So I just gave up on sleep over the years on Hackys and decided just to use more premium fans/coolers (for longevity) and leave the computer on and just let the monitors sleep on idle. Works perfect and I can come wake it up with a simple keyboard or mouse action.

I also set the left bottom Hot Corner to "Display Sleep" so I just swipe my mouse to the bottom left and get off my workstation and it auto locks and sets the displays to sleep and they go on idle. I come back and sit down it auto unlocks via Apple Watch or via Password. I also set a 10 minute timer for the displays to auto sleep and the system to lock for safety reasons.

And also since I'm using big radiators with lots of fans via push/pull and a custom watercooling solution, I can run the fans at a very low speed so even on idle it doesn't collect a ton of dust or make any audible noise.

Anyway, long story short, I think it's a waste of time trying to figure out sleep on a Hacky. You might get it to work if you never use any BT/WIFI cards (BT uses USB and Wifi PCIe) or even any USB device....but that means some type of virgin system which doesn't have a real world scenario setup....if that makes sense.

I hear you. So far the best working BT during my Hacky journey has been a Belkin USB BT dongle from like 10 years ago. But even my old 24" iMac from 2007 had some problems with sleep: it would wake up immediately if put to sleep using the Alt-CMD-Eject shortcut on either the white plastic or the aluminium Apple USB keyboard.
This is my second Hackintosh build since 2012 and sleep and wake has mostly worked just fine, but now I just can't wake from BT. No wake from BT is not a big problem though, more like a would be nice to have feature since wake from USB mouse works just like one would expect. This is just the way it goes naturally when you make progress - now that the machine is so close to perfect, the smallest things may start to give you trouble

 

[JimmyS]

What cpu cooler are you using? Could be a usb device preventing sleep.

Noctua NH-D15S, so no USB connection. Thanks for suggestion though.

Setting the internal USB header back as "internal" allows the machine to sleep again.