Jump to content

Xeon W and Asus WS C422


fabiosun

Recommended Posts

"fabiosun" post_id="12170" time="1534587536" user_id="56">

Sorry @obus

 

 

 

your first MSR patch togheter fakecpuid produces 1 for both terminal command?

 

Yes but with "_xcpm_pkg_scope_msrs_PMhart" and FakeCPUID.

 

Added in 3 minutes 21 seconds:

 

AppleIntelInfo.

 

 

 

Added in 33 minutes 51 seconds:

 

Ok fabiosun.

 

 

 

I have tested a little bit further.

 

With FakeCPUID my rig boots with full powermanagment (xcpm) together with this patches:

 

1._xcpm_pkg_scope_msrs_PMhart 31D2E8AE FCFFFF ->31D29090 909090

 

2.pikeralpha reboot fix554889E5 41574156 41554154 53504189 D64889FB 4189F54C 8D3D7662 -> C34889E5 41574156 41554154 53504189 D64889FB 4189F54C 8D3D7662

 

3._xcpm_smt_scope31D2E87E FCFFFF -> 31D29090 909090

 

 

 

With xcpm_bootstrap © Pike R. Alpha 89D804C4 3C227722 -> 89D804C1 3C227722 my rig is booting without FakeCPUID and with any other patch but always without native powermanagment (xcpm).

 

 

 

Could it be a bug in Clover? Obviously I can achieve xcpm mode if I use a fake skylakecpuid. If I use 0.0506E3 or E4 or even some other skylake cpuid it seems to work but as soon as I use the correct id 050654 for Skylake X everything seems to get fucked up.

AppleInteInfo.txt.zip

Link to comment
Share on other sites

  • Supervisor

Ciao obus

 

It seems we can not avoid using fakecpuid..

 

Probably also others like Broadwell Ep 0x40764 (I don't remember well) do the same job

 

It is very strange also you can use Appleintelinfo kext because in new I9 architecture it is not working well if MSR is checked in option

 

Analyzing it you miss C3 states in output but it is pretty normal

 

Previously when I had an unsupported cpu there was a kernel patches by SammlerG to activate C3 output, in real use however no difference

 

to come back to our "issue"

 

I am searching a kernel patch for adding cpu id in kernel section..but I can't find it

 

Also it would be the same thing of using fakecpuid, so no great catch

 

No I do not thing it is a clover bug?

 

but also it is strange if you use KernelXCPM flag is not working for you..

 

this flag was an automated way to patch unsupported Cpu we had asked the past to clover's coders to not use in automatic way.

 

 

 

My advice is to launch appleintelinfo and use your system loaded/unloade for few time and see if all CPU pSTATES come out

 

then evaluate if for you it is better to use a Broadwell EP fakecpuid or Skilake X in terms of stability and performances

 

Hoping Apple programmers include your cpu in next OSX release

 

 

 

Ha scritto dopo 1 minuto 51 secondi:

 

PS remember also to try unlock MSR bios option to see if you can avoid using MSR patch or any kernel patches

 

Using a fakecpuid you are using a way to use XCPM of another family of CPU

 

But it not should be your case because Xeon W family is A CPU supported in iMac Pro...( ;( )

  • Like 1
Link to comment
Share on other sites

  • Supervisor

C3 states could be visible only programming MSR 0xe2 from this value 0x1E000005 to another with the help of a kernel patch which an user )SammlerG) found for old OSX version

 

 

 

If I was you I would test with software like cinebench, Geekbench, luxmark 3.1 all your system and post the result also grabbing output and temperature with IPG

 

To see performances, stability and temperatures in both condition

 

I think however which your system could reach all States also with Skylake x fakecpuid..maybe it needs more time to reach all states

 

 

 

then you should have a clear situation of optimal condition for your system

 

 

 

Ha scritto dopo 5 minuti 41 secondi:

 

to explain better C3 states stuff (this is an extraction from an old Insanely Mac thread I created time ago)

 

follow patch I was using in Sierra to have C3 States, in bold the one which does what I have said in previous message (MSR 0xE2 programming)

 

 

 

Solution 2 (same results and behavior of solution 1)

 

 

 

Bios Unlocked

 

EIST Enabled

 

​4g enabled (only for Sierra)

 

 

 

​kext Patch

 

 

 

IOPCIFamily

 

find

 

4881FB00000040

 

replace

 

4881FB00000080

 

5960X 10.12 Brumbaer

 

 

 

6 patches

 

find

 

AA010000DC3300000000000000000000000000000000000001

 

replace

 

AA010000000000000000000000000000000000000000000000

 

SammlerG Sierra xcpm MSR Patch 1 / xcpm support Pike R. Alpha

 

 

 

find

 

A0010000DC33000000000000000000000000000040000000010005

 

replace

 

A0010000DC33000000000000000000000000000040000000010004

 

SammlerG Sierra xcpm MSR Patch 2 / xcpm support Pike R. Alpha

 

 

 

find

 

3A060000DC33000000000000000000001F00000000000000000000000000000000000000000000000000000000000000

 

replace

 

3A0600000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000

 

SammlerG Sierra xcpm MSR Patch 3 / xcpm support Pike R. Alpha

 

 

 

find

 

42060000DC33000000000000000000001F0000000000000018

 

replace

 

42060000000000000000000000000000000000000000000000

 

SammlerG Sierra xcpm MSR Patch 4 / xcpm support Pike R. Alpha

 

 

 

find

 

E20000004C00000000000000000000000F040000000000000500001E

 

replace

 

E20000004C00000000000000000000000F040000000000000700007E

 

SammlerG Sierra xcpm MSR Patch 5 / xcpm support Pike R. Alpha

 

[/color]

 

find

 

89D8C1E008B99901

 

replace

 

B800250000B99901 [ I use 25 (3700 Mhz) you can change accordly your OC clock]

 

(wrsmr fix to get CPU max) © okrasit

 

 

 

You can disable a couple of them if needed ;-)

  • Like 1
Link to comment
Share on other sites

  • Supervisor

yep another input

 

https://ark.intel.com/products/130046/Intel-Xeon-W-2175-Processor-19_25M-Cache-2_50-GHz">https://ark.intel.com/products/130046/I ... e-2_50-GHz">https://ark.intel.com/products/130046/Intel-Xeon-W-2175-Processor-19_25M-Cache-2_50-GHz

 

see your CPU specs

 

turbo for few cores is @4300 so CPU P states must be 43 and you have to see all in CPU PSTATES

 

 

 

your cpu has 14 cores / 28 Threads

 

this means you have to see in CPU C6 states all from 0 to 27..

 

when you see this situation without any hole..you are gold (technically speaking) :) in Italian way is simple to figure out this technically speaking..I don't know in English if it is the same :)

  • Like 1
Link to comment
Share on other sites

Thank's again fabiosun I will elaborate a little bit further with this C3 states stuff and report back to you.

 

 

 

Have a nice Saturday evening emoticon_festa

 

 

 

Added in 4 minutes 45 seconds:

 

fabiosun!

 

Could you provide the link to this your old thread about C3 states?

Link to comment
Share on other sites

Ok. What you telling me is that max turbo frequency is 4.30 GHz = 43 P-states

 

and according to my AppleInteInfo I can se 27 CPU C6-Cores = 28 cores (0 - 27)

 

 

 

Added in 5 minutes 18 seconds:

"fabiosun" post_id="12188" time="1534603295" user_id="56">
check your PM

 

Thank's empticons_thu

Link to comment
Share on other sites

  • Supervisor

ideal condition

 

 

 

CPU C6-Cores [ 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 ]

 

 

 

CPU P-States [ 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 (33) 34 35 36 37 38 39 40 41 42 43 ]

  • Like 1
Link to comment
Share on other sites

  • Supervisor

run 4 times Cinebenchz

 

for cpu and graphic with appleintelinfo terminal executed

 

also during test run Intel power gadgets

 

and post here image of these datas for both fakecpuid togheter complete appleintelinfo output (you have to run command in termal many times (sorry if you know this)

 

 

 

you have to test for about 15/20 minutes...

  • Like 1
Link to comment
Share on other sites

"fabiosun" post_id="12208" time="1534609107" user_id="56">

run 4 times Cinebenchz

 

for cpu and graphic with appleintelinfo terminal executed

 

also during test run Intel power gadgets

 

and post here image of these datas for both fakecpuid togheter complete appleintelinfo output (you have to run command in termal many times (sorry if you know this)

 

 

 

you have to test for about 15/20 minutes...

 

Here you go.

 

In 0x0506E4 CPU P-States 36 38 and 42 is missing.

 

In 0x040670 no CPU P-States is missing.

Archive.zip

Link to comment
Share on other sites

  • Supervisor

Ciao @obus

 

IPG graphics are pretty similar

You should do also 3 test if you can

Geekbench (CPU and GPU) in free version only OpenCL is available and it is ok

then in cinebench do some test also for gpu (as I was asking before) and for cpu single core like you can see in my attachment

 

I am asking because I can't see in both IPG graphics higher freq than 3.3 or 3.4 I think

In some condition you should see some peak around 4.0 and more

however outpinfo show all states so your cpu scales well in both case (few missing states are irrelevant..they could come far in time during testing

 

 

 

Ha scritto dopo 7 secondi:

LINK_TEXT>

 

 

 

Ha scritto dopo 1 minuto 9 secondi:

 

then important other things..stability is the same?

 

temperature seems pretty identical in both cases

 

Ha scritto dopo 9 minuti 58 secondi:

 

LINK_TEXT>

 

 

Ha scritto dopo 4 minuti :

 

LINK_TEXT>

 

during single CPU and Graphic cinebench run

  • Like 1
Link to comment
Share on other sites

"fabiosun" post_id="12262" time="1534687092" user_id="56">

Ciao @obus

 

IPG graphics are pretty similar

 

You should do also 3 test if you can

 

Geekbench (CPU and GPU) in free version only OpenCL is available and it is ok

 

then

 

in cinebench do some test also for gpu (as I was asking before) and for cpu single core like you can see in my attachment

 

I am asking because I can't see in both IPG graphics higher freq than 3.3 or 3.4 I think

 

In some condition you should see some peak around 4.0 and more

 

however outpinfo show all states so your cpu scales well in both case (few missing states are irrelevant..they could come far in time during testing

 

 

 

Max turbo frequency is 4.3 GHz but that never happened.

 

Could it have to be something with my FakeCPUID?

Link to comment
Share on other sites

  • 4 weeks later...

Hi again @fabiosun[/mention]

 

There is another guy on the InsanelyMac forum how bought

 

the same mobo as I and I invited him to this forum. I will wait for the official candidate release of Mojave before I take any further steps.

 

 

 

Everything is working stable on my rig now but I'm not capable to get turboboost over 3.5 or something like that.

 

If you have any news please inform me. empticons_thu empticons_thu

Link to comment
Share on other sites

  • Supervisor

ciao Obus

 

you and yapan4 have to understand if your MSR0xe2 is really unlocked for all cores of your cpu

to do this task you can use @vit9696 tools

 

first of all

 

VirtualSmc.efi in driver64uefi

then delete your fakesmc and all kext related to it (sensors, acpi an so on)

put in kext/others

VirtualSMC.kext

SMCProcessor.kext

then put

https://github.com/acidanthera/AptioFixPkg/releases

Aptiofix..../tools/VerifyMsrE2.efi

 

inside clover/drivers64uefi

you should see some lines before clover bootmenu like this one

 

LINK_TEXT>

 

if you see all your cores and similar text

you have a really unlocked bios 🙂

  • Like 1
Link to comment
Share on other sites

Perfetto!

 

 

 

I will check that during the evening. I will report back as soon as possible.

 

I'm already using virtualSMC and it's working flawlessly on my computer (I don't need any information about fans) even if I know that yapan4 has some problems with the fan sensors.

 

 

 

What about SMCHelper-64.efi? should I leave that in drivers64efi? It's not compatible with vitualSMC.efi if I understand it right?

 

Salve!

Link to comment
Share on other sites

  • Supervisor

this result is with no other kext in my efi/clover/others

only new bit Virtualsmc and smcprocessor

many things are missing but I think app developer have to adopt some modification to be fully compatible with new VirtualSMC era 🙂

 

LINK_TEXT>

 

Ha scritto dopo 1 minuto 39 secondi:

I am using only these ones in my system

 

LINK_TEXT>

  • Like 1
Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
  • Recently Browsing   0 members

    • There are no registered users currently online
×
×
  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue.