SalPro Dual Xeon WorkStation: ASUS Z9PE W8 DS - 2x XEON E5 2660 - GTX 590 3GB - 64GB

[TpwUK]

If you see a message like "should have X threads but only found Y" above the AppleACPIPlatform panic, it must be caused by Processor scope issue.
I couldn't avoid the panic without Processor scope edit (or not using 8-core Xeon).

I find it strange that the Sandy Bridge Xeon E5-26?? EP v1 chips do not generate this error, my first experience of this issue was with the Ivy Bridge v2 Xeons. Without your kid bit regarding Processor Scopes i would have been well lost at that point.

Now I'm planning to write small script to patch DSDT Processor scope.

This would be good for those who can not work this out for themselves. Once i had waded through all the SCK? declarations and saw where it needed to go it was a doddle, time consuming aside that is. So if you can do the script that would be cool.

I made some mistake about how to fix DSDT. So I fixed the post above. Sorry.

It was easy enough to get what you were referring to ... It was needed and I am grateful

I also think that is the problem of Maxwell GPU.
I couldn't boot UniBeast installer with GTX970 even if I use nv_disable=1, so I had to use GTX760 while installing.
After installing, GTX970 probably works with Web driver and nvda_drv=1, but I haven't tested yet.
I used default smbios setting of UniBeast, so it was probably MacPro3,1. After install, I introduced MacPro6,1 smbios.

I don't think Maxwell is included with default Mac Nvidia drivers yet, I might be wrong though, nv_disable=1 is used to turn off Apples drivers and nvda_drv=1 is to tell your Mac to use the Web drivers provided by Nvidia, but of coarse you can't do that on a clean install since you have not got them installed yet. That's the circle i got trapped in with an clover always adding the nvda_drv=1 so i had to add nvda_drv=0 to the end of boot args and to remove or 0 off nv_disable.

Newer clover does not have this issue

Martin

 

[rolandino]

I find it strange that the Sandy Bridge Xeon E5-26?? EP v1 chips do not generate this error, my first experience of this issue was with the Ivy Bridge v2 Xeons. Without your kid bit regarding Processor Scopes i would have been well lost at that point.



This would be good for those who can not work this out for themselves. Once i had waded through all the SCK? declarations and saw where it needed to go it was a doddle, time consuming aside that is. So if you can do the script that would be cool.



It was easy enough to get what you were referring to ... It was needed and I am grateful



I don't think Maxwell is included with default Mac Nvidia drivers yet, I might be wrong though, nv_disable=1 is used to turn off Apples drivers and nvda_drv=1 is to tell your Mac to use the Web drivers provided by Nvidia, but of coarse you can't do that on a clean install since you have not got them installed yet. That's the circle i got trapped in with an clover always adding the nvda_drv=1 so i had to add nvda_drv=0 to the end of boot args and to remove or 0 off nv_disable.

Newer clover does not have this issue

Martin

I'm a bit busy at the moment, so I will continue testing on Monday.
Maxwell isn't included in the Apple drivers, but nv_disable=1 also does not work.
I will try to boot the installer with a non-Maxwell card.

Meanwhile here is my modified DSDT.aml.
So we can compare the files, if you want.

 

[ccx4700]

I attached a small ruby script to patch DSDT.aml
Here is usage.

$ mv proc_patch.rb.txt proc_patch.rb
$ chmod +x proc_patch.rb
$ ./proc_patch.rb DSDT.dsl 16 > DSDT_patched.dsl

Specify first argument for DSDT.dsl path, second for threads per CPU.
If you have 8-core CPU with HT, you should specify 16.
(Because of limitation of this board, I have to add .txt extension.)

I find it strange that the Sandy Bridge Xeon E5-26?? EP v1 chips do not generate this error, my first experience of this issue was with the Ivy Bridge v2 Xeons. Without your kid bit regarding Processor Scopes i would have been well lost at that point.



This would be good for those who can not work this out for themselves. Once i had waded through all the SCK? declarations and saw where it needed to go it was a doddle, time consuming aside that is. So if you can do the script that would be cool.



It was easy enough to get what you were referring to ... It was needed and I am grateful



I don't think Maxwell is included with default Mac Nvidia drivers yet, I might be wrong though, nv_disable=1 is used to turn off Apples drivers and nvda_drv=1 is to tell your Mac to use the Web drivers provided by Nvidia, but of coarse you can't do that on a clean install since you have not got them installed yet. That's the circle i got trapped in with an clover always adding the nvda_drv=1 so i had to add nvda_drv=0 to the end of boot args and to remove or 0 off nv_disable.

Newer clover does not have this issue

Martin

As far as I know, Processor scope problem is caused by 8-core v2 only.
People who use 8-core v1 (like E5-2680) had no problem.
And I had owned E5-2660 v2 that has 10 cores, I had no trouble. At the time Yosemite couldn't support over 32 threads, so I limited E5-2660 v2 to 8 cores by UEFI, but it properly worked.
I don't know why 8-cores v2 Xeons have these problems.

 

[TpwUK]

Great !! ccx4700 Many thanks for sharing that, I think we must be about ready to start a new Z9PE-D8 WS how to with ElCap thread, what do you think ?

 

[ccx4700]

I think so too. There are no problems to solve now. We have fully working systems.

 

[TpwUK]

I guess the only way to know for sure is to do the suck it and see approach, expensive experiment if it don't pan out though !!

Martin

 

[rolandino]

Finally I got my system up and running (10.11.5).

Thanks for your DSDT and SSDT-1, power management is working.
Also I was able to inject HDEF into the DSDT, so audio is working without VoodooHDA.
For USB3 I ordered one of the Inateck controllers.

Last issue is: only one ethernet interface is recognized. Do you have both interfaces running?


Cheers
Roland

 

[TpwUK]

10.11.5 Knocked out my PowerManagement but Clover ASUS fix cures that
Not tried injecting HDEF for a while on this board as i normally fall foul of it somewhere :/
USB3 should be working with GenericUSBXHCI but not showing in your system profile, so you should be able to use them but not see them. Both ethernet ports are working here, is the one that's missing the one with the USB3 hub, if so it might be board failure, try your front USB3 Hubs if you have them connected - Test under windows or linux to see if it works there.

Feel free to post your DSDT on here for others to get HDEF working

Martin

 

[rolandino]

The USB3 Ports are working but only USB2 devices show up. Had this before and 20 Euros for the Inateck controller is worth a try.
Some people say, it is running smoother and faster than the Asmedia crap. I will report back.

The network interface working is the right one above the USB3 ports next to the audio connectors.
On Windows7 both interfaces worked, on Windows10 there was also a problem with the left port. Strange...
Which version of AppleIntelE1000 are you using?

I will post my DSDT and the Clover on-the-fly AppleHDA kext patch tomorrow.

 

[TpwUK]

The USB3 Ports are working but only USB2 devices show up. Had this before and 20 Euros for the Inateck controller is worth a try.
Some people say, it is running smoother and faster than the Asmedia crap. I will report back.

The network interface working is the right one above the USB3 ports next to the audio connectors.
On Windows7 both interfaces worked, on Windows10 there was also a problem with the left port. Strange...
Which version of AppleIntelE1000 are you using?

I will post my DSDT and the Clover on-the-fly AppleHDA kext patch tomorrow.

My USB3 ports have been known to fall asleep and not wake up until after a full power down, but they have always shown whatever has been connected to them. Windows 10 steadfastly refused to install on my old board, not tried it with the new one yet, but the old board always reported ACPI_BOOT_ERROR with a glum face and no 0x0 code to hunt the problem down with, just a very generous let us restart your machine message :\

I have added my AppleIntelE1000 info.plist but i am pretty certain it's the one from MB for ElCap but i could be wrong

Martin