Contribute
Register

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

Status
Not open for further replies.
I guess you intended to refer to the implemented Gigabyte Designare EX BIOS description rather than to a non-existing MSI BIOS description. ;)

Congrats to your build anyway :thumbup:

you are to much right :thumbup:
 
  • Like
Reactions: kgp
  • Like
Reactions: kgp
Great post! Thank you for clarification!

If you are interested about more detailed instructions how to find out the offset for your motherboard BIOS CFG Lock, which are varying, please check these detailed instructions with screenshots:

https://www.tonymacx86.com/threads/...-cfg-lock-msr-ex02.253466/page-2#post-1847437

In fact there was a dedicated section for patching ASUS X299 BIOS firmware in the originating post of this thread to formerly unlock the MSR register on all ASUS X299 mainboards. As ASUS upon my initial request implemented a feature to manually lock/unlock the MSR register within the BIOS settings of all ASUS X299 mainboards, our former detailed approach to unlock the MSR register by BIOS firmware patching became obsolete.

Interesting information that you prove in the other thread anyway.. congrats! :thumbup:
 
My Sapphire Nitro + Vega 64 (not limited edition) reaches when I play games >90°C.
If I remove the PP-PHmSoftPowerPlayTable from my SSDT it keeps apr. 10°C colder.
I'd say that these temps are not healthy.

At first i used @kgp kext with the embedded SoftPowerPlayTable just out of curiosity with my air cooled Vega 56.
As expected the temps were extreme and graphic artifacts were immediately visible.

I would suggest to either remove the SoftPowerPlayTable as you did or generate one with VGTab more suited for your air cooled Vega 64 and use the resulting kext.
I did so and even undervolted the graphics card to some extend. I love the fact that we have got the possibility to tweak the GPU to our needs under MacOS. But this project is ongoing for me. ;)


One thing to note - in my case iStat Menus reports the temperature of my Sapphire Pulse Vega 56 using the hotspot value for GPU Die which is not the GPU temp as reported on Windows. This is quite some 10+°C higher as the actual GPU Die temp.
So this Value may not be reliable to really judge the temperature.
 
I'd say that these temps are not healthy.

At first i used @kgp kext with the embedded SoftPowerPlayTable just out of curiosity with my air cooled Vega 56.
As expected the temps were extreme and graphic artifacts were immediately visible.

I would suggest to either remove the SoftPowerPlayTable as you did or generate one with VGTab more suited for your air cooled Vega 64 and use the resulting kext.
I did so and even undervolted the graphics card to some extend. I love the fact that we have got the possibility to tweak the GPU to our needs under MacOS. But this project is ongoing for me. ;)


One thing to note - in my case iStat Menus reports the temperature of my Sapphire Pulse Vega 56 using the hotspot value for GPU Die which is not the GPU temp as reported on Windows. This is quite some 10+°C higher as the actual GPU Die temp.
So this Value may not be reliable to really judge the temperature.

1.) Vega 64 temps are the same under iStatMenus, Heaven or Valley..
2.) How do you compare OC temps between Windows and macOS?
3.) I am not using any kext but a SSDT for Vega 64
4.) The implemented SoftPowerPlayTable is for the Sapphire Nitro+ Vega 64
5.) Yet I did not release any SSDT for a Vega 56

If you have a SSDT-X299-Vega 56 SSDT including an appropriate SoftPowerPlayTable let me know. BTW there is one by @Loloflatsix, see post #445.

If you want a SSDT without any SoftPowerPlayTable use my SSDT-X299-RX580.aml
 
MacOS Mojave 10.14.2 Beta 2 (18C38b)

Screenshot 2018-11-08 at 00.49.33.png


Enjoy and have fun,

kgp.png
 
Last edited:
here I share my SSDT-X299-Vega56.aml adopted from kgp SSDT-X299-Vega64.aml for my Msi VEGA56 Air Boost OC with
PP_PhmSoftPowerPlayTable with these main parameters:
GPU max frequency 1663 Mhz
Memory HBM2 max frequency 900 Mhz
Fan Idle 1540 Rpm

This SSDT-X299-Vega56.aml is for my Gigabyte GA X299 UD4 PRO : SB_.PC02.BR2A but can be modified according to IOReg with other mobos.

Fantastic! SSDT-X299-Vega56.aml now official part of the X299 System SSDT repository.
 
1.) Vega 64 temps are the same under iStatMenus, Heaven or Valley..
2.) How do you compare OC temps between Windows and macOS?
3.) I am not using any kext but a SSDT for Vega 64
4.) The implemented SoftPowerPlayTable is for the Sapphire Nitro+ Vega 64
5.) Yet I did not release any SSDT for a Vega 56

If you have a SSDT-X299-Vega 56 SSDT including an appropriate SoftPowerPlayTable let me know. BTW there is one by @Loloflatsix, see post #445.

If you want a SSDT without any SoftPowerPlayTable use my SSDT-X299-RX580.aml
2. in both cases i used a run of Heaven Benchmark.
In Windows i use HWMonitor to monitor Temps
On MacOS i used the output of: ioreg -l |grep \"PerformanceStatistics\" | cut -d '{' -f 2 | tr '|' ',' | tr -d '}' | tr ',' '\n'|grep 'Temp\|Fan\|Core\|GPU Activity\|Memory ' and compared it to the output of iStat.
iStat reports at this time while running Heaven 95°C - the ioreg output reports 85°C which seems reasonable given that i run the Vega 56 at stock speeds for now. Heaven reports the same value as iStat but ithink that the output of ioreg is the correct one.
3. I know but wanted to hint at the possibility to use a kext instead of the implementation of the SoftPowerPlayTable in the SSDT. For testing this might be beneficial as it might be simpler to just replace the kext than to edit the SSDT every time one changes some values.
4. Sure! And also for water cooled in the case of the provided SSDT by you? Or am i wrong with my assumption?
5. No problem with that! Maybe i will contribute with one if i settled with values i think are safe for mainstream Vega 56.

At this time i am using (for testing) your SSDT for Vega 64 without PP_SoftPowerPlayTable, PP_DisablePowerContainment and PP_FuzzyFanControl. The kext which was at first generated by VGTab only contains PP_DisableFanControl at this time.
By this i'm testing out the Vega 56 with stock settings without the problem of the fans at 100% all the time.
Works! :thumbup:

Next i will reinstate my PP_SoftPowerPlayTableValues for my use case (tiny case, air cooled, custom 12cm noctua fans as exhaust from the Vega 56).
But i think that the solution with only PP_DisableFanControl should give reasonable results with stock Vega 56/64 cards without the 100% fan issues. Stock means the values as provided by the GPU BIOS.
 
Last edited:
Fantastic! SSDT-X299-Vega56.aml now official part of the X299 System SSDT repository.
Please be careful as the MSI card uses slightly higher core clocks than any other Vega 56. 1622 vs 1590. The higher HBM clock of 900 vs 800 is in line of what i observed as safe on my Sapphire Pulse 56 too.
 
2. in both cases i used a run of Heaven Benchmark.
In Windows i use HWMonitor to monitor Temps
On MacOS i used the output of: ioreg -l |grep \"PerformanceStatistics\" | cut -d '{' -f 2 | tr '|' ',' | tr -d '}' | tr ',' '\n'|grep 'Temp\|Fan\|Core\|GPU Activity\|Memory ' and compared it to the output of iStat.
iStat reports at this time while running Heaven 95°C - the ioreg output reports 85°C which seems reasonable given that i run the Vega 56 at stock speeds for now. Heaven reports the same value as iStat but ithink that the output of ioreg is the correct one.
3. I know but wanted to hint at the possibility to use a kext instead of the implementation of the SoftPowerPlayTable in the SSDT. For testing this might be beneficial as it might be simpler to just replace the kext than to edit the SSDT every time one changes some values.
4. Sure! And also for water cooled in the case of the provided SSDT by you? Or am i wrong with my assumption?
5. No problem with that! Maybe i will contribute with one if i settled with values i think are safe for mainstream Vega 56.

At this time i am using (for testing) your SSDT for Vega 64 without PP_SoftPowerPlayTable, PP_DisablePowerContainment and PP_FuzzyFanControl. The kext which was at first generated by VGTab only contains PP_DisableFanControl at this time.
By this i'm testing out the Vega 56 with stock settings without the problem of the fans at 100% all the time.
Works! :thumbup:

Next i will reinstate my PP_SoftPowerPlayTableValues for my use case (tiny case, air cooled, custom 12cm noctua fans as exhaust from the Vega 56).
But i think that the solution with only PP_DisableFanControl should give reasonable results with stock Vega 56/64 cards without the 100% fan issues. Stock means the values as provided by the GPU BIOS.

But is the clocking of the GPU the same under Windows and macOS?

Sure, you can also use the kext in case you exactly know what else the kext is doing. ;) At some point you will need the SSDT and it can't work in line with the kext anyway. VGTab also provides you the PP_SoftPowerPlayTable directly to be implemented in the SSDT by copy & paste.. I would not know why this should be more difficult than using the kext anyway.

The SSDT-X299-Vega.aml implemented in the repository is for the airflow Nitro+ Vega 64
 
Status
Not open for further replies.
Back
Top