iMac Pro X299 - Live the Future now with macOS 10.14 Mojave [Successful Build/Extended Guide]

[pieropontra]

I used the excel-sheet method and put in some values i think will be all right for testing if anything will change with your setup. No overclocks but only lower fan speeds.

The corresponding sheet can be found here (if not already known to you):
https://www.tonymacx86.com/threads/...control-properties.267519/page-6#post-1884585

Although quite overwhelming at first, i think this excel-sheet is a nice alternative to the VGTab method as one can alter more values.

BTW: i'm also using Noctua fans (NF-A12x25) to cool my air cooled Vega 56.

I used a downgraded rpm curve kext with my Vega 64 LC made by VGTab and it won't work so why the card should take care of your kext? Just asking...

Moreover, please note that I'm using a miniPWM -> PWM adapter to a PWM splitter and to 2x non stock fan on my rad. Even in Windows RPMs are higher than they should!

 

[kubusikan]

I used a downgraded rpm curve kext with my Vega 64 LC made by VGTab and it won't work so why the card should take care of your kext? Just asking...

It was just an idea as my experience is that VGTab, as comfortable as it is, doesn't resulted in something i wanted.
In short, VGTab did something but idid not exactly get what happened and how i should adjust it to the result i wanted exactly.
Maybe i'm exaggerating ab bit here.

Moreover, please note that I'm using a miniPWM -> PWM adapter to a PWM splitter and to 2x non stock fan on my rad. Even in Windows RPMs are higher than they should!

I am using the same Adapter to connect my two fans.
But the thing regarding Windows seems strange to me as my system behaves unter Windows even without adjustments in WattMan.

How are your temps in Windows/MacOS without and under load?
My air cooled card sits at say 35-40°C idle and gets to 60°C when under load by heaven or furmark if the two Noctuas are at full blast (~2000rpm). And that's in a very tiny case.
If i set the target temp at the default 72°C the fans will not spin up that high and the whole thing is much quieter.

I would assume that a water cooled card will be way cooler without high spinning fans?

 

[pieropontra]

It was just an idea as my experience is that VGTab, as comfortable as it is, doesn't resulted in something i wanted.
In short, VGTab did something but idid not exactly get what happened and how i should adjust it to the result i wanted exactly.
Maybe i'm exaggerating ab bit here.

I am using the same Adapter to connect my two fans.
But the thing regarding Windows seems strange to me as my system behaves unter Windows even without adjustments in WattMan.

How are your temps in Windows/MacOS without and under load?
My air cooled card sits at say 35-40°C idle and gets to 60°C when under load by heaven or furmark if the two Noctuas are at full blast (~2000rpm). And that's in a very tiny case.
If i set the target temp at the default 72°C the fans will not spin up that high and the whole thing is much quieter.

I would assume that a water cooled card will be way cooler without high spinning fans?

I just noticed system info says my GPU is in slot 6 and it's actually in slot 1! Check the image attached

My card with Noctua at the 120mm rad is around 26° in idle so it's definitely cold and fans should spin slower. Maybe the kext don't work 'cause MacOS thinks it's in slot 6!? I don't know what to do, I'm really close to surrender and buy an air cooled Vega 64 or Frontier (and sell this nuclear engine I got now). How the heck MacOS thinks GPU is in slot 6!??! Can I fix this someway?!

 

[kubusikan]

I just noticed system info says my GPU is in slot 6 and it's actually in slot 1! Check the image attached

Well, as i'm not able to use my System today i can not check about this.

One thing you might try to see if the kext is actually loaded is to put in low core clock values (say 1000MHz and lower) for P5, P6 and P7 in VGTab and monitor if the GPU only reaches these low values when under load.
If it uses the higher stock values, then the kext does nothing indeed which indicates it is not loaded.

But i would advise to not surrender yet.
And keep in mind that an air cooled card will run hotter in any case and the fans will spin higher.

 

[izo1]

I just noticed system info says my GPU is in slot 6 and it's actually in slot 1! Check the image attached

My card with Noctua at the 120mm rad is around 26° in idle so it's definitely cold and fans should spin slower. Maybe the kext don't work 'cause MacOS thinks it's in slot 6!? I don't know what to do, I'm really close to surrender and buy an air cooled Vega 64 or Frontier (and sell this nuclear engine I got now). How the heck MacOS thinks GPU is in slot 6!??! Can I fix this someway?!

That doesn't really matter.

Every thread I've seen people who use LC Vegas they have the same issue with the fans as you do.

I would suggest continue using the radiator fan connected to your motherboard and capping the RPM to maybe 1200 constant (I am not sure at what RPM your current Noctua becomes audible). The GPU will not be damaged even if it overheats, it will just throttle.

Try the above and run a Unigine benchmark for 30 mins and check your temps with HWInfo under Windows. As long as your GPU thermals are below 84c and your HBM thermals below 90c you're good.

I've had similar issues, like previously mentioned on NVIDIA 980Ti EVGA Hybrids and I was forced to use the motherboard fan connectors on the radiator fans. It was fine and it was barely audible even on full load.

Maybe the Radeon VII will be better? It has 3 fans on it, although same hot Vega chip...

Also remember, you most likely have it running on Performance mode, that's why it runs at higher wattage and the fan kicks up faster.

Have you tried standard mode? The little switch close to the ports needs to be switched away from the ports for standard mode.

 

[kgp]

I thought that you might mention something about the 4 ports hub USB 2.0/USB 3.0. I tested every possible scenarios and combinations with the HUB ( Except x4 USB 3.0, as I did not have enough USB 3.0 Drive to connect ). The results are identical. Only HS05/SS05 would post no matter what. Attached is a few screenshots with various configurations. ( Tested upon your guide recommendation, without Kext on Mojave 10.14.0 ). HS09/11 are connected respectively to Airport/Bluetooth Broadband Card & wireless keyboard.

In regards to the ports HS 06, HS07, HS08 and HS13,HS14. Even after connecting the Aura RGB Xtra Christmas lights, no connections ( with or without XHCI usb Kext ) seem to interfere with the ports. AFAIK, my USB kext is fully functional.
Thank you for your input/ feedback. They are well received.

There is a general misunderstanding. You just need one USB3.0 stick for the port discovery. Not need to connect several USB3.0 sticks at the same time.

The 4 XHCI back panel connectors are USB3.0 and not USB2.0/USB3.0. Like all USB3.0 connectors they are of course backwards compatible to USB2.0 as long you do not only assign SS-ports to these connectors but also the respective HS-ports.

It might well be that all 4 connectors share the same HUB and only use HS05/SSP5. But then why you implement SSP6 at all in your kext? If you implement SSP6, you should also implement HS06 for USB2.0 backwards compatibility. But as you say that SSP6 and HS06 are not used at all, you can drop both ports from the kext.

Considering your above findings the fully implemented XHC USB kext for the GA X299 AORUS GAMING 7 only has 14 HS/SS ports and should be implemented as attached. The truncated 15-port kext is identical and has just a different file name for its implementation in the Github repository. I once more also updated the port-layout drawing for consistency with your above findings.

As soon you confirm that all USB2.0 and USB3.0 ports/headers of the GA X299 AORUS GAMING 7 work now with the files attached, I will upload both files to the Github XHC USB kext library.

Thanks for all your efforts and your kind collaboration.

 

[kgp]

That doesn't really matter.

Every thread I've seen people who use LC Vegas they have the same issue with the fans as you do.

I would suggest continue using the radiator fan connected to your motherboard and capping the RPM to maybe 1200 constant (I am not sure at what RPM your current Noctua becomes audible). The GPU will not be damaged even if it overheats, it will just throttle.

Try the above and run a Unigine benchmark for 30 mins and check your temps with HWInfo under Windows. As long as your GPU thermals are below 84c and your HBM thermals below 90c you're good.

I've had similar issues, like previously mentioned on NVIDIA 980Ti EVGA Hybrids and I was forced to use the motherboard fan connectors on the radiator fans. It was fine and it was barely audible even on full load.

Maybe the Radeon VII will be better? It has 3 fans on it, although same hot Vega chip...

Also remember, you most likely have it running on Performance mode, that's why it runs at higher wattage and the fan kicks up faster.

Have you tried standard mode? The little switch close to the ports needs to be switched away from the ports for standard mode.

@pieropontra, @izo1, @kubusikan,

No idea what this discussion is all about.

1.) Nearly all custom Vega implementations have 3 fans. No idea what the Radeon VII should improve in this context.
2.) A silent air flow custom Vega with stock performance is a wishful thinking. If you want a silent custom Vega with stock performance or even with some OC, opt for custom water blocking, which works perfect on my X299 rig and with my respective Vega-SSDT!
3.) Undervolting a custom air-flow Vega as mentioned by @kubusikan is a valid approach to make it more silent. However I strongly recommend to modify the load table of the respective Vega-SSDT and to avoid any additional kext.
4.) It is true that VGTab is complex, does not fully work as described, but with a little bit of patience and wit some willing for fine-tuning, it certainly helps in any case.
5.) That a SSDT with a load table or some kext instead does not help is a misleading statement, which I have to reject completely. Also the statement that all forums report fan issues with all custom Vegas is incorrect and misleading. See, e.g. THIS THREAD. They clearly state that at least for the Sapphire Nitro+ Vega with the respective firmware, all fans behave as expected and I can confirm this statement with an air-flow Sapphire Nitro+ Vega 64 implemented in my X99 rig for supporting all air-flow Vega guys.
6.) There are indeed two implemented stock firmware settings on the Sapphire Nitro+ Vega 64, which can be activated by the little onboard switch. However this solution is rather thought for guys not being able to work with a custom SSDTs or kexts. Nobody skilled is forced to only remain with one out of two stock settings and is free to use a SSDT load table or kext to fine-tune or optimise the Vega and its respective fans up to his personal taste.

I hope this sheds some additional light on all your above discussion.

@pieropontra, it is fine for me that you basically use this thread to make others work for you as you apparently do not have all necessary technical background to get most things done by yourself. However, I would like to ask you to consider when writing your posts that the latter are also read by others. As soon also others start to comment on your sometimes confusing and incorrect statements and contribute with their own not always adequate ideas and statements, the entire discussion starts to leave its rails and ends up leaving an impression to the reader, which is also inadequate in any case.

A great day to all you guys,

KGP

 

[Ellybz]

There is a general misunderstanding. You just need one USB3.0 stick for the port discovery. Not need to connect several USB3.0 sticks at the same time.

The 4 XHCI back panel connectors are USB3.0 and not USB2.0/USB3.0. Like all USB3.0 connectors they are of course backwards compatible to USB2.0 as long you do not only assign SS-ports to these connectors but also the respective HS-ports.

It might well be that all 4 connectors share the same HUB and only use HS05/SSP5. But then why you implement SSP6 at all in your kext? If you implement SSP6, you should also implement HS06 for USB2.0 backwards compatibility. But as you say that SSP6 and HS06 are not used at all, you can drop both ports from the kext.

Considering your above findings the fully implemented XHC USB kext for the GA X299 AORUS GAMING 7 only has 14 HS/SS ports and should be implemented as attached. The truncated 15-port kext is identical and has just a different file name for its implementation in the Github repository. I once more also updated the port-layout drawing for consistency with your above findings.

As soon you confirm that all USB2.0 and USB3.0 ports/headers of the GA X299 AORUS GAMING 7 work now with the files attached, I will upload both files to the Github XHC USB kext library.

Thanks for all your efforts and your kind collaboration.

I guess I misunderstood the kext creation guidelines and thats why I left the kext with 15 ports even if only 14 ports were used. My bad . SSP06 can of course be dropped as it is unused. I tried the 14 ports kext. All is fine. Have a nice week end.

 

[kgp]

I guess I misunderstood the kext creation guidelines and thats why I left the kext with 15 ports even if only 14 ports were used. My bad . SSP06 can of course be dropped as it is unused. I tried the 14 ports kext. All is fine. Have a nice week end.

Both modified kexts are now part of the library.

Thanks for all your efforts and help

A nice weekend to you, too

 

[izo1]

@pieropontra, @izo1, @kubusikan,

No idea what this discussion is all about.

1.) Nearly all custom Vega implementations have 3 fans. No idea what the Radeon VII should improve in this context.
2.) A silent air flow custom Vega with stock performance is a wishful thinking. If you want a silent custom Vega with stock performance or even with some OC, opt for custom water blocking, which works perfect on my X299 rig and with my respective Vega-SSDT!
3.) Undervolting a custom air-flow Vega as mentioned by @kubusikan is a valid approach to make it more silent. However I strongly recommend to modify the load table of the respective Vega-SSDT and to avoid any additional kext.
4.) It is true that VGTab is complex, does not fully work as described, but with a little bit of patience and wit some willing for fine-tuning, it certainly helps in any case.
5.) That a SSDT with a load table or some kext instead does not help is a misleading statement, which I have to reject completely. Also the statement that all forums report fan issues with all custom Vegas is incorrect and misleading. See, e.g. THIS THREAD. They clearly state that at least for the Sapphire Nitro+ Vega with the respective firmware, all fans behave as expected and I can confirm this statement with an air-flow Sapphire Nitro+ Vega 64 implemented in my X99 rig for supporting all air-flow Vega guys.
6.) There are indeed two implemented stock firmware settings on the Sapphire Nitro+ Vega 64, which can be activated by the little onboard switch. However this solution is rather thought for guys not being able to work with a custom SSDTs or kexts. Nobody skilled is forced to only remain with one out of two stock settings and is free to use a SSDT load table or kext to fine-tune or optimise the Vega and its respective fans up to his personal taste.

I hope this sheds some additional light on all your above discussion.

@pieropontra, it is fine for me that you basically use this thread to make others work for you as you apparently do not have all necessary technical background to get most things done by yourself. However, I would like to ask you to consider when writing your posts that the latter are also read by others. As soon also others start to comment on your sometimes confusing and incorrect statements and contribute with their own not always adequate ideas and statements, the entire discussion starts to leave its rails and ends up leaving an impression to the reader, which is also inadequate in any case.

A great day to all you guys,

KGP

Not a lot of people have the LC edition of Vega Frontier LC or Vega Limited Edition LC.

But from other forums and subreddits, which I won't mention here, most that I've read about LC edition Vegas were the fact that the radiator fans run at high speeds even on half loads. I'm too lazy to find the threads, but take my word for it, if you're so inclined.

As a rule of thumb for AIO coolers, anything above 250watts will require something bigger than a 120mm radiator to keep it "silent," or at least "inaudible." Thus, you either have to let it kick up the fan on full load or control the fan through means of using a motherboard connector or software.

But anyway, like you said, best is to have either a triple fan Vega like the Sapphire Nitro+, or control the rad fan on a LC edition of Vega imo maybe with Standard mode on and a little bit of undervolting.



And yes there are Vegas with double fans, example Gigabyte....