[Guide] Z820 - El Capitan The great guide (Sucess)

[theroadw]

Hi, Just wanted to share, as someone may be in my situation, recently I upgraded my Z820 with Dual E5-2667 V2 (Ivy Bridge) CPU's and I had 2 days of kernel panics... "should have 16 threads but only found 15 for die 1" or similar, and it didn't matter what I changed in boot flags (CPUS=1, Disable HT, -f -v -x, different kernels), the panic was there. But Windows had no trouble, so I knew my CPU's were ok.

 

At one point I found the problem is in the SSDT, the bios reported the CPU# in ranges that OSX didn't like, so after lots of trial and error by comparing against bilbo's SSDT, I managed to get it working, and then further cleaned up the SSDT and DSDT.

 

The CPU information has to be at the bottom of the DSDT, otherwise the SSDT-1 get's loaded and you get a kernel panic.

In my case...

    Scope (_SB)
    {
        Name (\PETE, Zero)
        Name (\PSTE, Zero)
        Name (\TSTE, Zero)
        Name (\NPB0, One)
        Name (\NPB1, One)
        Name (\NPB2, Zero)
        Name (\NPB3, Zero)
        Name (P0ID, Zero)
        Name (P1ID, 0x20)
        Name (P2ID, 0x40)
        Name (P3ID, 0x60)
        Name (P0BM, 0x04)
        Name (P1BM, 0x04)
        Name (P2BM, Zero)
        Name (P3BM, Zero)
        Name (CUU0, Zero)
        Name (CUU1, Zero)
        Name (CUU2, Zero)
        Name (CUU3, Zero)
        Name (\TST0, Zero)
        Name (\TST1, Zero)
        Name (\TST2, Zero)
        Name (\TST3, Zero)
        Method (PSTA, 1, NotSerialized)
        {
            If (LEqual (Arg0, Zero))
            {
                If (NPB0)
                {
                    Return (0x0F)
                }
                Else
                {
                    Return (Zero)
                }
            }

            If (LEqual (Arg0, One))
            {
                If (NPB1)
                {
                    Return (0x0F)
                }
                Else
                {
                    Return (Zero)
                }
            }

            If (LEqual (Arg0, 0x02))
            {
                If (NPB2)
                {
                    Return (0x0F)
                }
                Else
                {
                    Return (Zero)
                }
            }

            If (LEqual (Arg0, 0x03))
            {
                If (NPB3)
                {
                    Return (0x0F)
                }
                Else
                {
                    Return (Zero)
                }
            }

            Return (Zero)
        }

        Method (CSTA, 2, NotSerialized)
        {
            Store (Zero, Local0)
            If (LEqual (Arg0, Zero))
            {
                ShiftRight (P0BM, Arg1, Local0)
            }

            If (LEqual (Arg0, One))
            {
                ShiftRight (P1BM, Arg1, Local0)
            }

            If (LEqual (Arg0, 0x02))
            {
                ShiftRight (P2BM, Arg1, Local0)
            }

            If (LEqual (Arg0, 0x03))
            {
                ShiftRight (P3BM, Arg1, Local0)
            }

            And (Local0, One, Local0)
            If (LEqual (Local0, Zero))
            {
                Return (Zero)
            }
            Else
            {
                Return (One)
            }
        }

        Device (SCK0)
        {
            Name (_HID, "ACPI0004")
            Name (_UID, "CPUSCK0")
            Name (SCKN, Zero)
            Name (LSTA, 0xFF)
            Method (_STA, 0, NotSerialized)
            {
                Store ("CPUSCK0", CUU0)
                Store (PSTA (Zero), Local0)
                And (Local0, 0x03, Local1)
                Store (Local1, LSTA)
                Return (Local0)
            }

            Processor (C000, 0x00, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP000")
            }

            Processor (C001, 0x01, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP001")
            }

            Processor (C002, 0x02, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP002")
            }

            Processor (C003, 0x03, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP003")
            }

            Processor (C004, 0x04, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP004")
            }

            Processor (C005, 0x05, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP005")
            }

            Processor (C006, 0x06, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP006")
            }

            Processor (C007, 0x07, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP007")
            }

            Processor (C008, 0x08, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP008")
            }

            Processor (C009, 0x09, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP009")
            }

            Processor (C00A, 0x0A, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP00A")
            }

            Processor (C00B, 0x0B, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP00B")
            }

            Processor (C00C, 0x0C, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP00C")
            }

            Processor (C00D, 0x0D, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP00D")
            }

            Processor (C00E, 0x0E, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP00E")
            }

            Processor (C00F, 0x0F, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP00F")
            }
        }

        Device (SCK1)
        {
            Name (_HID, "ACPI0004")
            Name (_UID, "CPUSCK1")
            Name (SCKN, One)
            Name (LSTA, 0xFF)
            Method (_STA, 0, NotSerialized)
            {
                Store ("CPUSCK1", CUU1)
                Store (PSTA (One), Local0)
                And (Local0, 0x03, Local1)
                Store (Local1, LSTA)
                Return (Local0)
            }

            Processor (C100, 0x10, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP100")
            }

            Processor (C101, 0x11, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP101")
            }

            Processor (C102, 0x12, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP102")
            }

            Processor (C103, 0x13, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP103")
            }

            Processor (C104, 0x14, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP104")
            }

            Processor (C105, 0x15, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP105")
            }

            Processor (C106, 0x16, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP106")
            }

            Processor (C107, 0x17, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP107")
            }

            Processor (C108, 0x18, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP108")
            }

            Processor (C109, 0x19, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP109")
            }

            Processor (C10A, 0x1A, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP10A")
            }

            Processor (C10B, 0x1B, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP10B")
            }

            Processor (C10C, 0x1C, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP10C")
            }

            Processor (C10D, 0x1D, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP10D")
            }

            Processor (C10E, 0x1E, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP10E")
            }

            Processor (C10F, 0x1F, 0x00000410, 0x06)
            {
                Name (_HID, "ACPI0007")
                Name (_UID, "PCI0-CP10F")
            }
        }
    }

I think the SSDT-1 options are overkill and they also add more chances for trouble, I guess Z820 were made for a 112 Core CPU config? Anyways, just re-arranged the ID's into my configuration and deleted the rest.

 

***(edit)***

Unfortunately after spending a week or so trying all the tricks I know for getting PM to work, it seems to me that the CPU information I had to remove in the DSDT for the kernel to recognize the CPU properly and boot is also needed to be able to use PM, so... there... it works!! with these CPU's (I've read that E5-2687 V2 may have the same issue) but there is no PM.

Using Chameleon Enoch I can get PM to "partially" work under Mavericks (no P states, but all C states) followed by a freeze about 30 seconds after boot which requires a hard reboot. El Capitan will not even bother with Chameleon and will only get me 3 C states on Clover, no P states.

It looks like the Z820 with dual E5-2667 V2 CPU's is a no go for PM unless someone can fix the kernel.

I get very good scores in Cinebench and Geekbench though, so not complaining

***/(edit)***

 

After this it was business as usual with my Z820.

[bilbo]

 

It happened twice.

 

 

*** Panic Report ***

panic(cpu 8 caller 0xffffff801f9ff2dd): Kernel trap at 0xffffff7fa035c592, type 0=divide error, registers:

CR0: 0x000000008001003b, CR2: 0x00007ff1b0c0e7c0, CR3: 0x0000000034604000, CR4: 0x00000000000626e0

RAX: 0x000000000000005d, RBX: 0xffffff813b6ea800, RCX: 0x0000000000000000, RDX: 0x0000000000000000

RSP: 0xffffff8e30963da0, RBP: 0xffffff8e30963e40, RSI: 0xffffff814e1bae00, RDI: 0xffffff814e1baf00

R8:  0x00000000045d2001, R9:  0x0000000000000000, R10: 0x000000000ec00000, R11: 0xffffff8e30716058

R12: 0x0000000000000000, R13: 0xffffff81238bc000, R14: 0x0000000000000000, R15: 0x000000000000005e

...

 

 

When I wrote the guide, I forgot mention that , from every and now, I have to boot rebulding caches (using clover options). Have you tried to do it?

 

-------

 

 

 

hugo, I guess that after plug-in your cpus, you repeated the process of PM (extract the original ssdt, recreate the new one with pike's script, etc etc) , didn`t you?

[theroadw]

hugo, I guess that after plug-in your cpus, you repeated the process of PM (extract the original ssdt, recreate the new one with pike's script, etc etc) , didn`t you?

 

Yes recreated the script with many different versions and options, with no success. I think the problem lies in the kernel

[theroadw]

Ok, I tweaked the DSDT CPU details further and that allowed me to create a better SSDT script and got PM to work on 10.11.16!! (And as far as my tests go it worked and was stable but no P states on Mavericks)

I get performance that is close to Windows, so I'm happy there.

Not the easiest upgrade, so be ready if you want to try these CPU's that are now being sold cheaply on eBay.

 

Some pics.

 

[bilbo]

Ok, I tweaked the DSDT CPU ....

 

 

What kind of tweaks? I´m interested in this issue, because I'm interested in upgrade my CPU as well.

 

 

 

 

To clarify (and sorry if I'm a bit thick) .To execute only the pike's script from older dsdt and ssdt is not enough.

 

The steps are: (apart of use the right kexts and clover patches)

 

-Extract the originals SSDTs - (important, using clover at boot,  or using some utility from windows or linux. Not from OsX,

-Add the content of the new ssdt containing the ACPI Processor declarations

-reboot (important too!)

-Execute the pike's script to generate the extra ssdt.

[theroadw]

Bilbo, I understand, but here's a brief recollection of what I went through.

After changing CPU's OSX would freeze on CPU declaration. (Obviously I had removed the old DSDT and SSDT) I used my other mac to create a new USB clover disk to start testing, and extracted the fresh SSDT's and DSDT.

After performing all your tweaks from this thread on the DSDT and SSDT, it would still not work, so I researched the problem and there was some mention somewhere of using a bare minimum info for CPU part of SSDT for a similar kind of error. (also the SSDT-1 CPU info would get populated in a non-sequential order and using CPU numbers above what OSX likes)

That did not work, so I used what I could extract from the new SSDT-1 and compared it to your working E5-v2 DSDT (the CPU info at the bottom)

After cleaning up more stuff and removing all but the main declarations, and re-ordering and re-numbering the CPU's it worked, I managed to boot.

I then proceeded to test more by going back to my Mavericks install, and it worked, so I started messing around with PM.

Mavericks would freeze after 30 seconds or so, (like before adding the modified -maybe rolled back- AppleAcpiPlatform.kext) and 10.11.16 would only get me 2 or 3 states, so I methodically reintroduced the information I removed from the SSDT until I found that it is the If statements on each CPU that would for some reason get the system a kernel panic, so I removed the option and just left in the result of a positive.

Then, voila! PM is working at least on 10.11.16, but I don't know if it isn't getting all the states up to 4GHz or HardwareMonitor doesn't report very accurately. After some investigation, it looks like these CPU's clock a maximum of 4GHZ on single core and 3.66 on multicore, which is what I'm getting, so all is good.

The performance feels a lot faster though, and the benchmarks are very close or better than their Windows counterparts.

Would be interested in checking in 10.12, but I use my Z820 for audio and there are a lot of incompatibilities on that OS, so will wait to see what others get.

My SSDT edit may not be "perfect", but it works, it is stable, and I've spent enough time on it for now.

 

My clover config.plist is for all settings a copy of yours, using smbios macpro 6,1.

I had a patched AIPM kext from before but since I was using sandy CPU, I had to revert to original and re-patch.

My Bios is the most current 3.91 rev A.

Everything else is the same as in this guide, and the only trick to get it to boot is to fix the DSDT with the stripped CPU info from SSDT-1. (info has to be in the DSDT or you get a KP)

One important thing is the name or numbering of each CPU core. In my case they would get weird numbers that OSX didn't like, so I modified each one so they were sequential from CPU0 to CPUF and so on... and then added the socket 0 and socket 1, as they would not get assigned from the original SSDT-1.

 

I've attached my DSDT and SSDT files.

z820.zip

[bluejj1]

Hi Bilbo,

Thanks for the great guide! I was able to successfully install 10.12.6 on my HP z620 after I tried to follow the guide to the details.

Setup:

Z620

Dual E5-2643v1  >>> single E5-2667v2

Clover_v2.4k_r4200

Clover configurator classic edition v4.51.1.0

Kext Wizard and/or KextBea.....

 

Long story, first, I got the z620 with dual E5-2643v1 and decided to install 10.12.6 on it and I was able to make it mostly working by following your guide. Except only two out of the five USB2.0 were working after part VI of installation (fixing DSDT). After some attempts, I fixed this issue by installing RehabMan’s USBInjectAll kext (v2017-0724) into /library/extensions folder (5/5 USB2.0 working correctly). Also, I used VoodooHDA_v2.8.7 for the audio since the fix for 10.11 did not seem to work for 10.12.6. I believed I also got PM to work for E5-2643v1 using fakeSMC from multibea…….

 

At the moment, I got an E5-2667v2 and decided to move onto this CPU. However, I could not boot into the 10.12.6 because I got kernel panics (should have 16 threads but only found 17 for die 0) every time and none of the boot flags helped in this situation. BIG THANKS to @hugodlc!! I was able to solve my problem after asked him and carefully read his replies to this thread. Since we are using the same CPU (his is dual E5-2667v2 z820 and mine is single E5-2667v2 z620), I carefully compared his DSDT and SSDT to mine and make modifications on mine. Eventually, I was able to boot into the system and re-install 10.12.6 last night. So far, system seems pretty stable and everything seems to work fine (except USB3.0).

Thanks again to everyone who makes this possible!

 

Best,

[kuroshe]

Hi Bilbo,

Thanks for the great guide! I was able to successfully install 10.12.6 on my HP z620 after I tried to follow the guide to the details.

Setup:

Z620

Dual E5-2643v1  >>> single E5-2667v2

Clover_v2.4k_r4200

Clover configurator classic edition v4.51.1.0

Kext Wizard and/or KextBea.....

 

Long story, first, I got the z620 with dual E5-2643v1 and decided to install 10.12.6 on it and I was able to make it mostly working by following your guide. Except only two out of the five USB2.0 were working after part VI of installation (fixing DSDT). After some attempts, I fixed this issue by installing RehabMan’s USBInjectAll kext (v2017-0724) into /library/extensions folder (5/5 USB2.0 working correctly). Also, I used VoodooHDA_v2.8.7 for the audio since the fix for 10.11 did not seem to work for 10.12.6. I believed I also got PM to work for E5-2643v1 using fakeSMC from multibea…….

 

At the moment, I got an E5-2667v2 and decided to move onto this CPU. However, I could not boot into the 10.12.6 because I got kernel panics (should have 16 threads but only found 17 for die 0) every time and none of the boot flags helped in this situation. BIG THANKS to @hugodlc!! I was able to solve my problem after asked him and carefully read his replies to this thread. Since we are using the same CPU (his is dual E5-2667v2 z820 and mine is single E5-2667v2 z620), I carefully compared his DSDT and SSDT to mine and make modifications on mine. Eventually, I was able to boot into the system and re-install 10.12.6 last night. So far, system seems pretty stable and everything seems to work fine (except USB3.0).

Thanks again to everyone who makes this possible!

 

Best,

Hi bluejj1, did u got lsi working too? 

[bluejj1]

Hi bluejj1, did u got lsi working too? 

 

Hi,

I do not have lsi controller in my system so I totally skipped that part of installation. 

 

GL~

[milsabor]

hello everybody

don't know what happened to me, but Sierra forced me to do an update of my system to it's last update of 10.12....i could boot but had problems with my Nvidia 1080....and finally can't boot anymore....
Tried some fixes here or there, but always stuck ate the beginning of the boot as if it was the cache....


So finally, trying to install High Sierra on my Z620...and it reboots at the install...don't know why...


Has anyone tried this tutorial with High Sierra ?

[bluejj1]

hello everybody

 

don't know what happened to me, but Sierra forced me to do an update of my system to it's last update of 10.12....i could boot but had problems with my Nvidia 1080....and finally can't boot anymore....

Tried some fixes here or there, but always stuck ate the beginning of the boot as if it was the cache....

 

 

So finally, trying to install High Sierra on my Z620...and it reboots at the install...don't know why...

 

 

Has anyone tried this tutorial with High Sierra ?

 

I have not tried 10.13 by myself but I remember 1 or 2 threads mentioned some ppl could not install 10.13 beta onto their z600/z620 a while back. Don't know if the situation has changed or not.

 

I currently use a Nvidia 1050ti on my z600 (not z620) which is also running 10.12.6. I installed the last security update about 7-10 days ago and I remembered that the screen resolution became lower right after the update. But Nvidia driver manager asked to update the driver for 1050ti and it's working fine after the driver update. Hope you have found some solution with your situation.

[bennals]

Hi,

 

I'm just going through this guide for the first time and it's also the first time I've used Clover (done plenty of Hackintoshes with Chameleon/Enoch).

 

In part iv of the guide bilbo writes:

 

-Cpu.

 

Add the right CPU ID. For my CPU is 0x0a01

 

My cpu is an e5-2650v2 and the only reference I can find after some searching to it's cpuid is on cpu-upgrade.com where it's stated as "0306E4".

 

Is this likely to be the correct value, and if it is, do I enter it in the 'Type' Field?  If it isn't correct, how do I find the correct value?

 

Thanks,

 

Bill.

[milsabor]

I have the same question as bengals !
It's a 2690....and i have 206D6


i don't even remember how i managed to solve it 6 months ago

 

Hi,

 

I'm just going through this guide for the first time and it's also the first time I've used Clover (done plenty of Hackintoshes with Chameleon/Enoch).

 

In part iv of the guide bilbo writes:

 

-Cpu.

 

Add the right CPU ID. For my CPU is 0x0a01

 

My cpu is an e5-2650v2 and the only reference I can find after some searching to it's cpuid is on cpu-upgrade.com where it's stated as "0306E4".

 

Is this likely to be the correct value, and if it is, do I enter it in the 'Type' Field?  If it isn't correct, how do I find the correct value?

 

Thanks,

 

Bill.

@bennals i think yours is 0x406F1 

Hi,

 

I'm just going through this guide for the first time and it's also the first time I've used Clover (done plenty of Hackintoshes with Chameleon/Enoch).

 

In part iv of the guide bilbo writes:

 

-Cpu.

 

Add the right CPU ID. For my CPU is 0x0a01

 

My cpu is an e5-2650v2 and the only reference I can find after some searching to it's cpuid is on cpu-upgrade.com where it's stated as "0306E4".

 

Is this likely to be the correct value, and if it is, do I enter it in the 'Type' Field?  If it isn't correct, how do I find the correct value?

 

Thanks,

 

Bill.

[bennals]

@milsabor - Thanks for the response.  Where did you find that cpuid (0x406F1)?  I'll give it a try and hope it doesn't mess anything up.  After reading a little bit more about Clover though it seems like entering a value in the 'Type' field might only have cosmetic consequences and affect what is displayed in the 'About This Mac' screen?

[bennals]

I'm now at the dsdt cleaning stage (11) and have hit three syntax errors that all seem to be the same problem.  The first one is:

 

syntax error, unexpected PARSEOP_ONE, expecting '('

 

for the following bit of code

 

 If (WMIT)
            {
                ^^^WMID.WMBB
                One

            }

 

 

The second is:

 

syntax error, unexpected PARSEOP_INTEGER, expecting '('

 

for:

 

If (WMIF)
            {
                ^^^WMID.WMBB
                0x02
            }
 

 

and the third is the same as the second error for:

 

 If (WMIH)
            {
                ^^^WMID.WMBB
                0x03
            }

 

 

Hopefully it's an easy thing to correct by sticking some parentheses in the right place.  Can anyone help?

 

Thanks,

 

Bill.

[bilbo]

Bennals,  0x0a01 is an ID for all Xeon E5 v2 processors , BUT, now, I´m not sure 100%.

 

Anyway, yes, it seems only a cosmetic data. But I always set up clover , just in case. It's only a kind of superstition.

 

My advice: Use 0x0a01 or left it blank. (or ask to people who really know about this )

 

About your syntax errors, my bet is you did some wrong fix. Do you have that error on a original dsdt? Try to fix it again from scrach, step by step.

 

Maybe it was done when you tried to fix the "Not all control paths return a value (WHB_) " . Adding  the “Return (Zero)” at the right place sometimes is very tricky. Pay your attention on the last example of my guide.

[bennals]

Ok, I'll try again.    While you're here Bilbo, I have the exact same graphics card as you (Quadro K4000) and even after installing the web drivers the performance is very glitchy.  Moving windows results in big drags and mousing across a Firefox window creates a big block of blinking pixels under the pointer.

 

Did you experience anything similar in the early stages after installation, and if so, at what stage of the process did it resolve?

 

Bill.

 

edit: Whoops, I see your gfx card is a K5000 - my bad. 

[bilbo]

                         ---------------NOTICE ---------------

 

Hey mates.

 

After one year beeing happy with El Capitán, the idea of update to Sierra is shaping in my head. Looking for more perfomance , these days I have been thinking in buy a 1080 Ti , but I've read that the drivers are only compatible with last Sierra version number. So update the SO is increasingly necessary. (I'd like update the cpus also)

 

Moreover, with the release of the first beta of high sierra 10.13 , we would say that Sierra is finally stable .

 

I plan make a new and documented guide, but now with a little help. It's not a plain in short term, because I'm busy. Maybe at the end of December or at the new year.

So , I'd like ask for a little brief of your experience from those with Sierra running in your Z820, Installing, updating the kexts, clover versions, usb issues, etc, and the differences with the current guide.

 

 

thanks

[asiansurf]

Hello Bilbo,

 

Your guide has been of great help to me as I followed your instructions to the letter for my Z620 but using Sierra.

 

I have since upgraded the CPU to an E5-2690, installed the HP Thunderbolt 2 card and a GTX1080Ti and it is now a speed demon!  It is super stable and super useful.  It had never froze on me.  Like all MacOSes before, it just works.

 

Do note that for the Z620, when the Thunderbolt 2 card is installed in PCIe slot 3, you need to squeeze the GPU to the last slot.  My GTX1080Ti was too fat, so I pinched a slimmer GTX960 to fit into the last slot.  I hope the Z820 should have slots wide enough to accommodate both cards?

 

A few months ago, I chanced upon a Dell T7910 with a E5-2623 v3 and tried to replicate a Sierra install.  But without a guide like yours, I gave up without ever getting it to boot at all.  There isn't any instruction for a Dell workstation available.... yet.

 

One lesson I have learnt recently. I updated the Clover app without updating Clover at boot.  This has caused a very perculiar symptom where Sierra would kernel panic at exactly 1 minute 20 seconds and reboot!

 

It is good to have an exact bootable clone of Sierra and keep a backup of the Clover folder in EFI.  This has saved me a lot of recovery time when I feel adventurous to perform an OS update or upgrade or install some unstable app or drivers.

 

The MacOS running on the Z620 started out as a curiosity but it is now my main "Mac Pro" workstation!  Thanks to you, Bilbo!

 

Out of curiosity, I tried upgrading to High Sierra on the Z620 but it didn't boot.  I suspect that I did not install clover into the High Sierra boot partition.  Will try it again some other time to get it to work.  Do you think that copying the whole Clover folder from the existing functional EFI could do the trick?

[bilbo]

Asiansurf, I'm glad to hear that. No. I don't think that copy the efi folders works. A lot of kexts have to be update... I think that the most conflictive is AICPM, applehda, usb... nope. In fact, is a surprise that the same kexts for El capitan of my guide, work in Sierra. Do you have native sound with my AppleHDA kext patched?

 

 

Otherwise, I'm not interested in High Sierra . I think it's in Beta version (in hackintosh and in native hardware). I can't understand why people rush to update it. For me, the cycle of the release has a gap of one year. TIme of the final release of Sierra starts now. High Sierra is in beta, with the free testing of the people.

 

I'm aware of the problem of de pci- slots space. There is no enough room in z280 either. The EVGA SC2 gaming has similar size of my current Quadro.

 

Interesting to know that you are using a E5-2690. My other option is the E5-2687W but I'm not sure about E5-XXX7 models and the powermagement.

 

.

.

.

 

 

 

Ok, I'll try again.    While you're here Bilbo, I have the exact same graphics card as you (Quadro K4000) and even after installing the web drivers the performance is very glitchy.  Moving windows results in big drags and mousing across a Firefox window creates a big block of blinking pixels under the pointer.

 

Did you experience anything similar in the early stages after installation, and if so, at what stage of the process did it resolve?

 

 

Well, looks similar to one old problem: In old versions, (Mavericks, Yosemite) there were an important graphic issue due to a bug between dual cpus and quadro graphic cards. Was necessary rollback a pair of kexts from snow leopard, (with some cons)

 

But that problem dissapeared on El capitan. I've never have that kind of issue. Try to post your config and your install steps.

[amadeusex]

Hi everybody !

 

Proud z820 owner here and thanks to bilbo's excellent guidance up and running with El Capitan.

 

Just one problem i'm trying to solve.

 

After applying the power management fixes my geekbench results dropped from around 24k to 18-19k.

Same with cinebench 15 (from 1700 to 1300-1400).

 

Any clues why this happens ?

 

Before the pm patch my cpus where constantly running @2.6 GHz and now they go up to the 3.3 GHz mark but don't perform as well.

 

FYI my z820 is equipped with 2 E5 2670 Xeons and all 16 cores (and Hyperthread) are visible from within the OS.

 

Thanks again for all Your effort ! It's highly appreciated

 

P.S. Attached are my patched DSDT AND SSDT files from CLOVER/ACPI/PATCHED

I also forgot to mention that i have BIOS version 3.91

Archive.zip

[amadeusex]

I'm now at the dsdt cleaning stage (11) and have hit three syntax errors that all seem to be the same problem.  The first one is:

 

syntax error, unexpected PARSEOP_ONE, expecting '('

 

for the following bit of code

 

 If (WMIT)

            {

                ^^^WMID.WMBB

                One

            }

 

 

The second is:

 

syntax error, unexpected PARSEOP_INTEGER, expecting '('

 

for:

 

If (WMIF)

            {

                ^^^WMID.WMBB

                0x02

            }

 

 

and the third is the same as the second error for:

 

 If (WMIH)

            {

                ^^^WMID.WMBB

                0x03

            }

 

 

Hopefully it's an easy thing to correct by sticking some parentheses in the right place.  Can anyone help?

 

Thanks,

 

Bill.

Hi Bill

 

Try to place the values in parentheses 

 

If (WMIH)

            {

                ^^^WMID.WMBB

                0x03

            }

 

to

 

If (WMIH)

            {

                ^^^WMID.WMBB

                (0x03)

            }

 

and see how it goes

 

Getting my DSDT clean was a challenge and did drive me crazy at some times.

 

The way i did it was to open bilbos and compare all the entries one by one  (the syntax).

It's very easy to screw up especially the "return zero"-edits.

 

 

 

[amadeusex]

Allright Guys ..

 

Good news.

 

After switching the machine on today i tried again some benchmarks and the results look good.

 

Geekbench 3 32000

http://browser.geekbench.com/geekbench3/8491237

 

Cinebench 15 1829 and 55 fps (Nvidia GT 740)

 

Can't really say why but can't complain either.

 

Greetings from Greece !

[bilbo]

Cool! Always is great read good news 

 

... and bennals,  yes, I've just check out my dsdt patched and the original dsdt, and the mistake seems to be the parentheses in one, 0x02 and 0x03 values.

 

This is my dsdt:

            If (WMIT)
            {
                ^^^WMID.WMBB (One)
            }

            If (WMIF)
            {
                ^^^WMID.WMBB (0x02)
            }

            If (WMIH)
            {
                ^^^WMID.WMBB (0x03)
            }

[bennals]

That's great, thanks guys.  I'll try that as soon as I can.

 

On the graphics card front, after stuffing up the dsdt I booted from a cloned backup I made of the initial 11.0 installation.  I was surprised to see the graphics glitches were gone and the K4000 properly recognised in 'About This Mac' (after updating to 11.6 it was reported as Nvidia Graphics Chip 8MB or something).  This made me wonder if the update to 11.6 caused the problem, so I copied the nvidia related kexts from the 11.0 installation to the 11.6 installation and voila, the card is now properly recognised in 11.6.

 

It's a bit weird because the nvidia kexts from both the 11.0 and 11.6 installations seemed to have the same creation dates so I would have assumed they were identical, but maybe that's something I don't understand properly.