Search results

Honestly, X670(E) on mini-ITX should be a non-starter. It is not a reasonable design to force two PCH chips on a tiny motherboard which cannot possibly expose all of the I/O that X670 provides. B650 is more than enough in this size—save for bragging rights.

Digital Rights Management This is just cosmetic (and RestrictEvents takes care of it), but note that DDR5 drops speed when running with 2 DIMMs per channel so you'd be better with 2*32 GB than with 4*16 GB.

That was fast! But did I see a screenshot of a code contribution? :eek: Is that the new hardcore trend in the Silicon Valley? (No. NO! Don't send it to THAT boss…) It' sounds almost like you're setting a trap for someone. Almost. :cool:

At least Zen2/3 APUs have been shown to benefit from the patch. The compute dies are the same, whether there is an iGPU or not. It's not clear why the sync issue in compute dies would be worse, or would only occur, when an iGPU is present in the package/within the IO die. But it's also not clear...

Thanks! That pseudo-code makes the change easier to understand. For reference, regular OpenCore (from Library/OcAppleKernelLib/CpuidPatches.c, line 1420) // Perform TSC and FSB calculations. This is traditionally done in tsc.c in XNU. // busFreqValue = CpuInfo->FSBFrequency...

The new calculation is inspired by Clover, but not identical to it. We also see in the table posted by CaseySJ that some TSC-related values, which I suppose derive from secondary calculations, have values which match neither those derived from Clover nor those derived from regular OpenCore...

Somehow this seems a logical next step after having first set up serial debugging on a motherboard without serial ports and then set up the software environment to compile the OS kernel and customised versions of its extensions in order to investigate PCIe issues on AM5… But let me express a...

And how's audio with the "NOT loved bootloader" compared with the "loved" one?

Wouldn't the DEXT require working AppleVTD?

@CaseySJ posted an EFI on page 28. This should work with Big Sur, plus the patches (or a combined version with masking) for Monterey and Ventura if you're running bare metal. If you're installing as a VM things may be different as actual PCIe initialisation would have been done by the hypervisor.

Outstanding work! On behalf of the mere mortals among us who cannot read compiled binary code, may I ask what this decompiles to and what this code is doing?

PCIe4.0 x8 has the same bandwidth as PCIe 3.0 x16 and few workloads manage to make use of all that. Virtualisation may well solve issues with macOS here. Which hypervisor?

This suggest that recent versions of macOS are increasingly reliant on IOMMU (VT-d on Intel platforms), and that solving the issue may require to attach AppleVTD to AMD-V (AMD's implementation of IOMMU). Which is easier said than done…

This was off-topic but it was the first report I've seen of Z790. Thanks!

If it is a TSC Sync issue it may require that the developers of TSC Sync or AMD Sync have a look at it—which would require they have the hardware. But other observations suggest this is an issue with power states, explaining why SpeedKeeper helps. We all hope it might work, but the issues are...

So this is a power state issue… What happens if you disable C6 state in BIOS? (I'm suggesting this because TrueNAS Core, which is based on FreeBSD, requires disabling C6 to run properly on Ryzen. Maybe Darwin has the same underlying issue?)

Including the slight discrepancy between hw.cpufrequency: 4500000000 and machdep.tsc.frequency: 4499995964 ?

Nothing is dropped; this is indeed a patch… and as shown it's disabled anyway. :unsure: General purpose method, to be used by (future) Thunderbolt SSDTs among others. I wonder why USB ports are defined both by SSDTs and by kexts but @CaseySJ must know what he's doing. This is going to be very...

Alder Lake CPU have "only" 20 PCIe lanes, compared with 24 for Ryzen 7000 (16+2*4). So you're out of luck, unless a manufacturer decides on a weird splitting of these lanes instead of the intended 16+4.

This is evolving fast. Thanks for sharing—though you should probably not share your serial numbers. To be clear: This is with the same EFI(2) you shared earlier, without PCIe drives or devices other than the GPU, the OS drive is a SATA SSD and i225 does not work (nothing attached to it in...