fabiosun

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 ;-)

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...( ;( )

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

ciao ti chiedevo anche i software in quanto per la scheda grafica ancora oggi e' possibile scegliere Nvidia o AMD anche se il supporto apple e' orientato su AMD

uhm for me today is enough if you can try boot strap togheter fakecpuid..and test if XCPM and Frequency Vectors are set to 1 good evening :) Ha scritto dopo 40 secondi: tomorrow we will try other solution :)

ok in that condition could you flag in config KernelXCPM and see

If you have a dual system with also old High Sierra you can use both usb patches specifying matchOS accordly have you test latest config?

https://www.gigabyte.com/FileUpload/global/MicroSite/473/x299.html Strumento interessante per il posizionamento delle periferiche (Gpu/Nvme) per sfruttare al massimo le "lanes" della propria cpu per possessori di gigabyte X299

interesting tools to configure properly lanes cpu/GPU/Nvme To achieve optimal performance for pciexpress slot for X299 Gigabyte owners https://www.gigabyte.com/FileUpload/global/MicroSite/473/x299.html

for 10.14.x AppleUSBXHCI find 83FB0F0F 83030500 00 Replace 83FB0F90 90909090 90 MatchOS 10.14.x sorry I am concentrated on CPU..I haven't seen you are not using correct one for Mojave Ha scritto dopo 5 minuti 52 secondi: also you can find here (in Italian language) a summary of changement USB port limit thread

It is the one you use in your config.plist use it togheter boot strap and see (without fakecpuid obviusly) :) Ha scritto dopo 1 minuto 35 secondi: if possible put also here actual efi you are using in the tries Thanks Ha scritto dopo 5 minuti 53 secondi: obus, if you can download from here this tool https://www.macos86.it/MacOS86_Downloads/Applicazioni/Info/Hack_Info.zip">MacOS86_Downloads/Applicazioni/Info/Hack_Info.zip">https://www.macos86.it/MacOS86_Downloads/Applicazioni/Info/Hack_Info.zip post zip it produce. in this way is it possible to check other aspect

Ciao Perdu per prima cosa prova a spostare tutti i kext in Other magari e' solo il solito problema di clover che si ripropone ciclicamente;)

in processor tab you see unknown or your processors? obviously if you have not modified it manually..as I do in my case

ok from from clover boot menu press O and disable alternatively kernel patches above leave only and always active boot strap and cycle the others ie bootstrap scope bootstrap core boot strap MSR boot strap pkg and see also, when system starts try to open terminal window and run: sysctl -n machdep.xcpm.mode and sysctl -n machdep.xcpm.vectors_loaded_count post results

first try in config 1 state flag Kernel PM on config.plist with clover configuration Ha scritto dopo 4 minuti 18 secondi: then try this kernel patches togheter bootstrap patch find 31D2E87E FCFFFF replace 31D29090 909090 comment_xcpm_smt_scope find 31D2E891 FCFFFF replace 31D29090 909090 comment_xcpm_core_scope find 31D2E8AE FCFFFF replace 31D29090 909090 comment_pkg_core_scope Ha scritto dopo 1 minuto 16 secondi: do not do any changes in bios during this test also we know using fakecpuid is helping..but we have a goal to avoid it :)

ciao obus Have you tried these two config .plist?

config2.plist.zip config1.plist.zip

fatto :) uSb stessa patch dell'ultima beta..

Grazie mille :) comodissimo tool per i più' pigri e non :)

Che ne pensate dei nuovi thread ripper? Mi aspettavo di più' (molto di più') dal modello top..ma quelli intermedi sembrano essere ottimi

beta 6 :)))))

vedi che accade al cabinet quando fanno delle saldature sui ponticelli delle VGA nvidia? :P :)

dai che @Baio77[/mention] non se la e' cavata con il latino :)