Contribute
Register

Z690 Chipset Motherboards and Alder Lake CPU

EFI folder v. 0.5

I have amended the standard OpenCore 0.7.5 EFI folder for Alder Lake with the feedback of the last days.
  • There are now two different EFIs: Debug-TXT and Release-GUI.
  • 'Debug' is a minimal folder, which is primarily intended to dump a SysReport for reference and adjustments. It has quirk XhciPortLimit enabled, for USB mapping under Catalina or Big Sur 11.2.3 and lower, and features the free version of memtest to validate a build. When it's done, switch to 'Release'.
  • 'Release' is the full folder, with GUI by default. It features UsbInjectAll for USB mapping with Big Sur 11.3. (If you have made a map with XhciPortLimit under an older OS, or mapped from Windows or Linux, just use this map and disable the USB quirk/kext.)
  • 'Release' has two sample.plist: One for iMacPro1,1 (simplest setup) and one for MacPro7,1. The latter uses RestrictEvents, CPUFriend and the frequency vectors from iMP1,1 (provided by @darthsian ) to match the performance of the former (as shown in benchmarks by @CaseySJ ).
  • To save some space, memtest is only present in 'Debug' and AppleALC is only present in 'Release'. Copy them around and activate the corresponding entries as you want.
  • Networking assumes Realtek 2,5 GbE by default. Amend as needed. There is a stub device-i225.plist with the Device Properties values as used by @CaseySJ and @Stork ; if your board has an i225V use either—but not both at the same time.
As with previous versions, pick the sample.plist which most suits you, add serial numbers and rename to config.plist. This should work for most Z690 boards. In BIOS, either disable E-cores (P-cores +HT) or disable Hyper-Threading (P+E cores, no HT). Happy hackintoshing!

Edit. Added missing attachment, and caveat:
I compile these folders without having an actual Alder Lake build to validate, so there could be errors. Please carefully check the settings and use ocvalidate on the final config.plist.
Unfortunately, my MSI Z690-A Pro can't load with this two EFIs: Monterey, Install Monterey and Recovery.
I tried with HT off.
 

Attachments

  • IMG_4561.jpg
    IMG_4561.jpg
    1.3 MB · Views: 50
Last shipment. Assembly time!
 

Attachments

  • tempImageuEoK4a.png
    tempImageuEoK4a.png
    11.9 MB · Views: 78
In my case, as CaseySJ knows, my adventure with Asus motherboards. I started my latest hackintoshes with Asus z390-A Prime and Maximus Hero XI, and I experienced many many freezes for months w/o been able to solve any of this no matter what I tried. Then Casey suggested the "Designare z390" and since then my 2 hackintoshes run w/o any issues! Gigabyte doesn't have the best BIOS or OC possibilities like other brands like Asus and MSI. But in terms of stability I'm sure it exists on the top!
GIGABYTE boards were the most compatible dating back to the P55 SoC and OS X Snow Leopard days.
 
35E6F4CC-0319-48EF-A3ED-9AB16C72E22F.jpeg


So I swapped over the WD M.2 from the B550 Vision D, and this is what came up for the AJA disk test.
Much improved for sure!
 
Last edited:
In my case, as CaseySJ knows, my adventure with Asus motherboards. I started my latest hackintoshes with Asus z390-A Prime and Maximus Hero XI, and I experienced many many freezes for months w/o been able to solve any of this no matter what I tried. Then Casey suggested the "Designare z390" and since then my 2 hackintoshes run w/o any issues! Gigabyte doesn't have the best BIOS or OC possibilities like other brands like Asus and MSI. But in terms of stability I'm sure it exists on the top!
I tried an Asus board years ago and had a bad experience where the board died, so I RMA'd it and they claimed to have fixed it, then when it came back two weeks later and I tried to power it on, it was still dead. I was so disappointed because all the tech websites at the time were recommending Asus over everyone else.

Ever since then I never used an Asus board again, but I've heard from others good and bad things about their Asus experience. I built a Skylake system several years after that, and used an MSI z170 Gaming M7 on the Skylake 6700K.. it is still working like a champion all these years later. Very stable. But then I wanted thunderbolt so I tried the Gigabyte z490 Vision D, and it has been very stable, working like a champion, very fast POST times. No issues whatsoever. Then somehow I had the bright idea to also try the z590 Vision D but it has been quirky particularly with thunderbolt 4... but overall still very stable. But it took a couple weeks of debugging and then getting Gigabyte to send me a test BIOS.

I'm sure Asus has made changes over the years but I will never forget my experience with their customer support. I'm looking forward to Alder Lake, Raptor Lake, Meteor Lake and the future lakes and seeing how Intel improves its platform. Z690 is a very good platform in terms of PCIe lanes which enables features on some boards like 10 gbe + thunderbolt + 4 m.2 pcie4.0 slots + wifi6e. Alder Lake is very impressive and I will probably build a system using it... but I may also hold out for Raptor Lake. We'll see. I'm looking forward to Zen4/5 as well. All in all, I'm excited about the competition that has come to the PC market. Intel, AMD, Apple, and now Qualcomm will drive innovation for many more years. I am getting the feels of how I used to feel in the 1990s when I first started to immerse myself into technology. Very exciting.
 
So I did some further tests today.
Disk speed read tests are certainly good. These are results from the Samsung 980 on Big Sur.

View attachment 534967

View attachment 534970


However performance is sporadic, with reads going from 1247-1501 and reads from 1747-2166

However it seems that there's one more setting we can turn on in BIOS to improve performance.
Under the Asus BIOS with AI Tweaker there is one setting called Asus Multicore Enhancement.
You have to set it to Enabled 'Remove All Limits'. With this turned on I/O was much higher (read was 2200MB/s).

View attachment 534968

View attachment 534969
You would be more pleased if that 980 was delivering the 3GB/s throughputs it is designed for.

Re sporadic results, note that the AJA test file size is 1GB, which I would expect to overstate throughout because the test never gets the drive fully pumping. These newer drives will have a bursty response up to some 10s to a 100s of GB written, then roll off on due to page pipeline swamping (so to speak, not a technical term).

On read throughput, media reliability may be a bugaboo as ECC works to deal with inevitable imperfections of flash. Thermal throttling can also slow things down on bigger chores, read or write.

Samsung claims to have a "Full Power" mode which must be enabled in Magician (under Windows). I use a 980 Pro 2T but never bothered to look into Full Power mode it because my drive gave expected performance of ~ 3GB/s on PCIe3 via AJA 10GB test file size without having to futz with anything. Then I ran into media errors and data loss, which changed my focus somewhat <g>

Later, running ddrescue to zero the whole drive, I saw expected behavior of writes slowing down by 50% after about 400GB and not recovering for the remainder of the job. After the zeroing, I also saw a full-drive-read stalls with throughput decreasing to 100K/s for a few seconds without generating errors, which I presume was the drive dealing with bad flash. Note that this specific read performance problem this is purely anecdotal with a sample size of 1/me, whereas there's lots of discussion on the web about general NVMe write performance slowing for large sequential workloads.

I say all this to help others consider how there are various factors potentially at play.

However, in your case, better performance than you report can be expected given your new board and drive.

But where to look for bottleneck?

Well, ASUS multicore enhancer is not the best place to look. It has nothing to do with IO limits which are all about DMA by drive controller. Multicore enhance is about removing power / clock frequency limits for compute loading going beyond Intel recs for OC full load. For OC champions with exotic liquid cooling this opens door for leaderboard claims. For regular joes, it opens door for 1-7% gains based on highest power draw depending on model and bin. Note the degree of diminishing returns at the top end of any overclocking. For example 10900K goes from 4.9Ghz max all-core to 5.1 for luck of draw best-bin parts at cost of 50-100W power increase and attendant thermals. ASUS makes this an optional tweak to remind users that they are at hairy edge.

Rather than looking for a compute-related BIOS adjustment to increased performance, look to ensure HW config is right re bus slot / lane assignments, BIOS NVMe config and drive power.

For benchmarking, you can try resetting the drive with Magician Secure Erase to get it to best-case state for testing. Keep in mind this is just to aid system performance eval. Real work is always another fish kettle. Also you might check SMART for stats indicating either media or bus errors, just to size up situation. And consider that drive usage age / wear can take its toll. You can do this in Magician, or with smartmontools from homebrew or macports or by booting Linux.

For z690 kit users will prefer PCIe4 980 Pro which is widely known to deliver over 6GB/s on PCIe4 slots. I haven't heard anything about PCIe 5 but I am not really looking yet.

Finally, per expectations: sequential rate bragging rights tend follow the biggest drive form-factors due to nature of flash.
 
For anyone who needs a new pcie 4.0 m.2 : XPG 2TB GAMMIX S70 Blade - Works with Playstation 5, PCIe Gen4 M.2 2280 Internal Gaming SSD Up to 7,400 MB/s (AGAMMIXS70B-2T-CS) https://www.amazon.com/dp/B093DNV47J/?tag=tonymacx86com-20

This seems like a good drive although I am just using it for windows gaming at the moment . Still using my xpg8200 for macOS.

**UPDATE**
I do not recommend this SSD I got a kernel panic today related to it:

panic(cpu 0 caller 0xffffff800a942836): nvme: "3rd party NVMe controller. Command timeout. Write. fBuiltIn=1 MODEL=XPG GAMMIX S70 BLADE FW=3.2.F.P7 CSTS=0x1 US[1]=0x0 US[0]=0x1 VID=0x1dbe DID=0x5236 CRITICAL_WARNING=0x0.\n" @IONVMeController.cpp:6053
Panicked task 0xffffff859b2e36a0: 227 threads: pid 0: kernel_task
Backtrace (CPU 0), panicked thread: 0xffffff859ea10b70, Frame : Return Address
0xfffffff0fd1e3a20 : 0xffffff800809c00d mach_kernel : _handle_debugger_trap + 0x41d
0xfffffff0fd1e3a70 : 0xffffff80081f5d85 mach_kernel : _kdp_i386_trap + 0x145
0xfffffff0fd1e3ab0 : 0xffffff80081e5763 mach_kernel : _kernel_trap + 0x533
0xfffffff0fd1e3b00 : 0xffffff800c0178f4 as.vit9696.VirtualSMC : __ZN18VirtualSMCProvider10kernelTrapI22x86_saved_state_1010_tEEvPT_Pm + 0x454
0xfffffff0fd1e3b80 : 0xffffff800803ba60 mach_kernel : _return_from_trap + 0xe0
0xfffffff0fd1e3ba0 : 0xffffff800809c3dd mach_kernel : _DebuggerTrapWithState + 0xad
0xfffffff0fd1e3cc0 : 0xffffff800809bb96 mach_kernel : _panic_trap_to_debugger + 0x2b6
0xfffffff0fd1e3d20 : 0xffffff8008918649 mach_kernel : _panic + 0x54
0xfffffff0fd1e3d90 : 0xffffff800a942836 com.apple.iokit.IONVMeFamily : __ZN16IONVMeController14CommandTimeoutEP16AppleNVMeRequest.cold.1
0xfffffff0fd1e3da0 : 0xffffff800a925ac7 com.apple.iokit.IONVMeFamily : __ZN16IONVMeController13FatalHandlingEv + 0x141
0xfffffff0fd1e3dd0 : 0xffffff8008860c75 mach_kernel : __ZN18IOTimerEventSource15timeoutSignaledEPvS0_ + 0xa5
0xfffffff0fd1e3e40 : 0xffffff8008860b78 mach_kernel : __ZN18IOTimerEventSource17timeoutAndReleaseEPvS0_ + 0xc8
0xfffffff0fd1e3e70 : 0xffffff80080ecc85 mach_kernel : _thread_call_delayed_timer + 0x4f5
0xfffffff0fd1e3ee0 : 0xffffff80080edcd2 mach_kernel : _thread_call_delayed_timer + 0x1542
0xfffffff0fd1e3fa0 : 0xffffff800803b18e mach_kernel : _call_continuation + 0x2e

I might return this and get the SN850
 

Attachments

  • 76EABF24-34E1-4F41-9EC0-8ACC93A86947.png
    76EABF24-34E1-4F41-9EC0-8ACC93A86947.png
    68.1 KB · Views: 39
Last edited:
Unfortunately, my MSI Z690-A Pro can't load with this two EFIs: Monterey, Install Monterey and Recovery.
I tried with HT off.
You may try with all SSDTs disabled, except SSDT-AWAC and SSDT-CPUR-Z690.
Can you post the SysReport dumped by the debug EFI to see if there's something in there which requires adaptation?
 
Back
Top