Z490 & Z590 - Will Z590 ever have macOS Support ?

[dehjomz]

I sent your comments to Gigabyte.
But they haven't responded to my previous submission yet.
I also wrote that this whole design does not make sense without a Thunderbolt hot-plug.
If you have the opportunity, also report it.

Thanks ! I have also been in communication with their tech support team, which to be honest is very slow, they took over 17 days to initially respond. But as of now, I'm losing hope.

Gigabyte does not appear to realize that their motherboard's thunderbolt implementation has a problem... They responded to me with the classic "we don't see what you see" when it comes to thunderbolt hotplug. I'm not confident that the person responding to tech support even used the z590 Vision D, they don't appear to have first hand knowledge. They even told me that thunderbolt "should be" Plug and Play. I am not confident that they have actually tested whether hotplug is working with thunderbolt3 devices...especially those of the Alpine Ridge variety.

I even told them that their BIOS's GUI has a bug in it, in that when you disable the thunderbolt controller in the bios, then the thunderbolt settings disappear completely from the BIOS, and the only way to get it back is to clear CMOS. They actually with a straight face tried feeding me the line that "this is normal" behavior to which I pushed back heavily and told them simply that their z490 Vision D doesn't do that, so if it wasn't 'normal' for the z490, then how can you believe it 'normal' for the z590? A user should never have to clear CMOS to regain access to a BIOS setting that was once there.

They even asked me to send them my thunderbolt 3 dock to them...it appears the number of thunderbolt devices they use to "test" is limited. Uhh no way, you're not getting my dock.

Intel and these board partners market thunderbolt as a super usb, but when I first started using USB many moons ago, once I plugged in a device, it came on, no problem. So I don't understand how the so-called super usb thunderbolt4 of 2021 doesn't do hotplug properly (with thunderbolt 3 devices). As of now, thunderbolt3 seems to have been largely debugged with Titan Ridge and it is working fine (as far as we can tell without drilling down on a machine-code level). So what's the point of Maple Ridge if it still needs to cook? I honestly don't understand the benefits.

We're speaking to Gigabyte, but I wonder if we need to speak to Intel as maybe Intel's firmware is the root of the issues. It's hard to know if it's a BIOS ACPI implementation error or a firmware error.

Finally, some makers of thunderbolt4 docks claim that to use the hubbing feature of their docks (i.e., the additional thunderbolt 4 ports), you need thunderbolt4 host controller, but this is not true in all instances, as my thunderbolt4 dock works beautifully on the z490 Vision D and of course my macBook pro.

The only issue I have on the z490 Vision D is an issue of my own making which is that I spilled thermal paste into the cpu socket as I was removing comet lake, and some of it also got into the graphics card socket. I cleaned it, but unfortunately now it doesn't recognize 2 of the 4 dimm slots. Which is why I was hoping for z590 to work properly...

I think I'm just going to give up on z590 and hopefully get a used Z490 Vision D as the new ones seem to be sold out world wide...

If the performance regressions we're seeing with the 11900K's overclocking ability and multicore performance (compared to the 10900K), and now Maple Ridge's spotty hotplugging ability with some thunderbolt3 devices (seems to me the problem is Alpine Ridge support as compared to Titan Ridge support), then is it any wonder why Apple broke up with Intel? Apple had to have seen Intel's roadmap and thought it was a joke, and probably also saw tons of bugs on the firmware level with thunderbolt... I mean how interested would customers be in a new iMac if it had an 11900K inside after it was largely panned in the press and was even called worst than a waste of sand? (which I think is a bit harsh by the way, Rocket Lake is okay to me) Apple's prospective customers are going to be an order of magnitude more interested in and excited by a new Apple Silicon iMac, so Apple made the right decision to move on.

Hopefully Gelsinger can turn this ship around. Intel needs to understand the power of inspiring excitement in their customers, and move on from this "I'm the dominant supplier so you'll buy from me even if I my product is buggy and slower than the competition" attitude. Users do not get excited by firmware bugs or performance regressions.

 

[mattystonnie]

I have a new Asus Z590M Prime-Plus paired with a i5 10400 and didn't manage to enable the UHD 630. The same CPU paired with a Z490 works perfectly fine.

 

[UtterDisbelief]

I have a new Asus Z590M Prime-Plus paired with a i5 10400 and didn't manage to enable the UHD 630. The same CPU paired with a Z490 works perfectly fine.

Hi there.

This is what I have found too. Same for me. I enabled it okay but no acceleration. So far no-one that I have seen has managed to get the iGPU working with acceleration using a 500-series motherboard and 10th gen CPU.

 

[dehjomz]

So the Z590 Asus HERO XIII comes with Maple Ridge Thunderbolt NVM firmware version 24.0, and this firmware has working hot-plug with Thunderbolt 3 devices. It appears that NVM firmware 26.x, like v26.5 on the Z590 Vision D, does not properly hot=plug Thunderbolt 3 devices...at least on some motherboards.

 

[c-o-pr]

Thanks ! I have also been in communication with their tech support team, which to be honest is very slow, they took over 17 days to initially respond. But as of now, I'm losing hope.

There is nothing new about this situation WRT evolving standards and interfaces.

Apple couldn't figure out USB on Intel Macs for a long long time even though Apple leas the launch USB.

Your points about Apple wanting off Intel's ride are well taken and I agree. I saw Intel Architecture R&D from inside for 20 years and they have never had a coherent architectural vision much less less a philosophy. Even as they dominate the market they are endlessly scrappy, and it shows in the complexity and ham-fisted product lineups and evolution. But my point is not to bash Intel. They also deliver! It's just very rarely beautiful in execution.

Thunderbolt has no champion except Apple. The origins of TB in serial interconnects for high-performance computing (HPC) are close to Intel's heart, but Intel never took interest in advancing it to commodities; well they did in sense of letting Apple do it. Microsoft doesn't care except that Apple cares. HPC I/O is a nightmare of goofy tradeoffs.

Apple supplied the coherent consumer vision for TB, and it works, but it has been far from compelling due to ridiculous mismatch of interface capabilities with expectations and Apple's significant but niche market share: TB has been constantly changing and not simplifying things as promised; it's too complex and too simple at once, and makes ambiguous crossover into USB in vastness of PC-land.

... So far.

USB 4 brings 25 years of clunky fits-and-starts of Intel I/O work together into a coherent and performant interface, its not quite done yet re deployment.

Too bad that at moment it's about to be baked, Apple can develop own systems top-to-bottom and bails.

So now PC Thunderbolt > USB 4 completion is left PC market which never much cared. It's just one more way to things that are already done.

The flip-side is it's a fairly complete and deep I/O subsystem now with interesting and desirable traits. There's so much inertia, it will get done.

So it's natural that in first few months of rollout, it's gonna be buggy because PC (everything) is always like this, and because of change in Apple direction. Rocket Lake seems anti-climactic but it's a decent step ahead: PCI-4, USB 4 and strong single thread while holding onto multi performance are fine. It just that it's not at all amazing, while M1 is fantastic! The Mac Mini M1 is a total marvel of price/performance.

I dig your discontentment, and thanks for helping to cross check with Windows and engage board vendors. But it's in no way a crisis, just one more clumsy step in evolving tech and things will get smoother.

I think 500 series platforms might come to be loved, in same sense as 2012 Macbook Pro, because they are right at crossing of architectural era between fully baked old ways and an impending shift to more performant new ways that will also cause a lot of forgetting, whether we like it or not. Just look at all those stupid type-A ports on rear-IO. That shiite's gotta go soon.

I can stand to give this ASUS a few months to straighten out and look forward to M2 and Alder Lake.

 

[mattystonnie]

Hi there.

This is what I have found too. Same for me. I enabled it okay but no acceleration. So far no-one that I have seen has managed to get the iGPU working with acceleration using a 500-series motherboard and 10th gen CPU.

Hello there.

Connected remotely I can see the UHD 630 is identified correctly but no video output on any port: DVI, HDML or DP.

I tried for one week and gave up, the best I could was to see a video output with 14Mb VRAM and no acceleration.

Let's hope they will release an iMac based on Intel 11th processor, maybe on 20 April.

 

[ori69]

Hello there.

Connected remotely I can see the UHD 630 is identified correctly but no video output on any port, dvi, hdmi or dp.

I tried for one week and gave up, the best I could was to see a video output with 14mb vram and no acceleration.

Let's hope they will relase an iMac based on intel 11th processor, maybe on 20 April.

What did you enter in Devices?

 

[UtterDisbelief]

Hello there.

Connected remotely I can see the UHD 630 is identified correctly but no video output on any port: DVI, HDML or DP.

I tried for one week and gave up, the best I could was to see a video output with 14Mb VRAM and no acceleration.

Let's hope they will release an iMac based on Intel 11th processor, maybe on 20 April.

Yes, I noticed that macOS identified the real Device ID of the new iGPU correctly - no matter what I was spoofing - and actually gave me output so I knew the connectors were roughly okay, but in my case only 7MB VRAM and no acceleration.

 

[mattystonnie]

What did you enter in Devices?

AAPL,ig-platform-id: 07009B3E, 0000A53E and device-id: 9B3E0000, A53E0000.
I tried connectors patching and framebuffer-stolenmem but no luck.

Only works with device id 3E9B0000 without hardware acceleration.

 

[mattystonnie]

Yes, I noticed that macOS identified the real Device ID of the new iGPU correctly - no matter what I was spoofing - and actually gave me output so I knew the connectors were roughly okay, but in my case only 7MB VRAM and no acceleration.

14MB VRAM if you use the DisplayPort, 7MB with the HDMI.

So you're saying it is identifying the real id of the iGPU 0x9BC8 / 0x9BC5 ??

If yes, I know a way of spoofing the id, it might work.