Leoyzen

I swiched to opencore instead of clover, there is something interesting: the virtio still need to hot-plug, so the fault is not clovers and there is definately something differences between hotplug and pre-defined. There is no more need of using Penryn, not even New Generation Intel CPU Model can be used but also host passthrough works! So maybe it's interesting for those who use host-passthrough and amd-vallina patches to gain some performance. Topology still not work, I don't want wasting my time on this anymore.You have to remove qemu topology. there is no patches compare to clover "FixDisplay" so to get graphics work properly I have to use ssdt, see here according to here, there is definately something we can do to make cpu/gpu xcpm work, will dig in later. Until now, everything works fine for me, so maybe I will use opencore instead of clover for now.

yeah, add the patch will be easy.Maybe new release by limetech already have them included.

@ghost82 That's strange....I will test more.

Do you have vmxnet3 as primary ethernet interface? When I use vmxnet3 as primary ethernet interface then the iperf3 is much slower.....But with another interface then it turns to be correct.

I found that it is clover which breaks vmxnet3 and virtio-net, maybe during clover devices initializing. After hotplug the ethernet devices after clover initialization, I've got a much better throughput performance: 1. virtio-net works(works properly) and get 20Gb/s+ performance(before won't work, osx can't get correct mac address after clover handled the devices) 2. vmxnet3 works properly and get 4Gb/s+ performance (before only 100Mb/s+ and quiet buggy) I will try to find the specific issues where the clover break. Maybe we are closed to get a much more efficient ethernet option which works property Steps to make virtio-net work: 1. do not add virtio-net in vm xnl 2. boot vm, when in clover boot selection section, go to 3 3. [IMPORTANT] virsh attach-device --live Catalina /boot/config/ethernet-virtio.xml 4. [IMPORTANT] add clover boot args: keepsyms=1 debug=0x100 the ethernet-virtio.xml looks like this: <interface type='bridge'> <mac address='52:54:00:d6:ac:37'/> <source bridge='br0.100'/> <model type='virtio'/> </interface>

1. yeah, it is the same whether mojave or catalina, it always get full 2. my grahpics is Sapphire RX460 but I flashed Rx560 bios(I checked the memory manufacture) 3. I'm not facing any problem but the degree is high and the memory seems something wrong for me. 4. I will try other sensors to see any difference. Because I've hear somewhere that after mojave the graphic PM ONLY works with CPU PM (X86PlatformPlugin and etc), which we won't have. So I guess maybe something related.

@ghost82 My RX580 get 22k+ score so your opencl score definately incorrect(maybe nvidia webdirver related, I'm not sure). My GPU memory always get full and fan not spin up untils get 70+ degrees, I don't know if it pm related issues or bios related even AMD related.... I've got VirtIO network work in Catalina and finally get 20Gb+/s network in Hackintosh VM, awesome! Edit: but still unstable for daily use, it become buggy when reboot .... I should find a way to make it work.

@ghost82 I use iStats Menu to monitor graphic die/mem frequency and mem usage, and manually monitor the fan by my eyes =0= @david279 I face the same problem before when installing catalina using iMacPro1,1 then I switched to MacPro7,1 and it works. It is sucks if as you said the newest update got the same problem. I have no idea what the error is, seems some cpu feature check?May need some patch to make it work. The full error log, seems mce support broken? 2019.09.24 20:41 I found the similar issues from here, so maybe use latest Lilu and CPUFriend can avoid this panic. I will try later and report back. @david279 2019.09.24 22:33 Lilu and CPUFriend not work as expected, I dont familar with Lilu/CPUFriend, so maybe someone test again and give some feedback. iMacPro1,1 and MacPro7,1 are important for those who use hackintosh in VM and want to get native GPU HW acceralation. 2019.09.24 23:38 I've found a solution to disable AppleIntelMCEReporter when using iMacPro1,1 or MacPro7,1 by using the kext from here. You can download the kext from attachment then unzip the kext into clover/kext/other folder then reboot MCEReporterDisabler.kext.zip

gpu fan works properly? my gpu fan only runs when get 70C gpu memory works properly? my gpu memory alwary get full and never lower down.... gpu benchmark compared to bare metal computer? BTW, I add cpu features as much as possible, finally I get nearly same score (1200+ singlescore) from my windows. I don't know which feature is more important so I add them all. Penryn,vendor=GenuineIntel,+hypervisor,-erms,+invtsc,kvm=on,+topoext,+svm,+invtsc,+fma,+mmxext,+avx,+avx2,+aes,+xsave,+xsaveopt,+ssse3,+sse4_2,+popcnt,+sse4a,+bmi1,+bmi2,+arat,+abm,+3dnowprefetch,+adx,+clflushopt,+cr8legacy,+fsgsbase,+fxsr_opt,+misalignsse,+movbe,+osvw,+pclmuldq,+pdpe1gb,+rdrand,+rdseed,+rdtscp,+sha-ni,+smap,+smep,+svm,+vme,+xgetbv1,+xsave,+xsavec,+ht,+clwb,+umip,+xsaves,+extapic,+ibs,+skinit,+tce,+wdt,+topoext,+perfctr-core,+perfctr-nb,+amd-ssbd,+wbnoinvd,check Edit 2019.09.26: It is also IMPORTANT to set fakecpuid to get avx2/bmi1/bmi2 etc recognized by MacOS(mention from here). You can just set cpu model from Penryn to IvyBridge and remove topology from xml, this works too. You can check with "sysctl -a|grep hw" or "sysctl -a|grep cpu" or "sysctl -a|grep avx" to see if the feature correctly recognized. It will get better performance (for me, the system boots faster than before).

Try and see, maybe work maybe not work, I dont know. glad to hear that. No, I tried but it seem to complex for me, because there isn't any steps or tutorial to compile the nvidia module. Sorry. BTW, I've built a 5.2.14 kernel lately with those update: with zenpower build-in, somebody said it is accurate than k10temp vega reset bug patch from Vega 10 and 12 reset application navi reset bug patch from Navi Reset Kernel Patch some ALSA HDA crackled patch for X570 from here and here but I don't have neither vega or navi card to test with, so if somebody interested in, let me know. Also, ** this kernel may influence by these bug **: [6.7.X] VERY SLOW ARRAY CONCURRENT PERFORMANCE DOWNGRADED BACK TO 6.6.7 DUE TO SQLITE CORRUPTION I will update when limetech release newest patches until unRaid 6.8+(which use 5.2.x+ kernel).

Interesting, I've been told there is no metal support for Nvidia GFX series after mojave. Does dedicate graphic power manager works for you? @ghost82

Do you have metal acceralation support?MacPro 7,1 added in catalina so there is not config in mojave. BTW, why your process shows 'Sconosciuto'?

You can just upgrade/replace the two efi files (EFI/CLOVER/CLOVERX64.efi and EFI/BOOT/BOOTX64.efi, they are the same file) which download from github instead of install whole directory by pkg or iso.

Glad to hear that

It is strange, I'm using 3700x and passthrough 4 core 4 thread into the VM, and I've got 1100+ single core and 9000+ multicore score. Does "Meta support" enables for you ? You should find the information in "about my mac ---> System Infomation ---> Grahpics" . It may cause the performance problem.

Here U are, those files attached are my clover files for Catalina and Mojave as nearly bare metal VM, both works for VNC or Real Monitor with Graphics passthrough: Something should notice: * Catalina: MacPro7,1(which iMacPro1,1 not works for me it get reboot cyclely with "x86_topo_lock error") * Mojave: iMacPro1,1 ** It do not need any kext to make function work ** Both files contains OOB patch (for AMD RX560/570/580 patch to make hardware acceralation(HEVC & H264) work also to avoid black screen (make graphics in acpi gfx0 position should avoid this)) and Clover as new as possible (v5058, the latest wont boot for me). Remember you should change SMBios UUID to avoid conflict. Using this then you can get an OOB Hackintosh(Mojave or Catalona) works like charm. As I explained above, you should take some patches from clover to make iMacPro1,1 work for Mojave, also MacPro7,1 for Catalina(check my clover.plist). mojave.clover.qcow2 catalina.clover.qcow2 Edit: some steps to make rx560/570/580 works natively in mojave and catalina to support hevc/h264 hardware acceraltion without IGPU: change your SMBios to iMacPro1,1(mojave) or MacPro7,1(Catalina) select ACPI --> Add DTGP in clover：add some method in acpi for later use. select ACPI --> FixDisplay: in order to change your grahpic to gfx0 select KernalAndDriverPathes --> Black Screen Patch Vega 56/64, RX580 etc. (c)Pike R. Alpha: fix blackscreen issues (Optional)If not work, then add LILU/WEG to you kexts directory. Here are some screen shots about my VM: Geekbench 5 with 4 cores/ 8 threads of my 3700x:

For those who have problem with AMD grahpics in hackintosh (DP/HDMI sound not work, reset bug or etc): I've spent days to get my rx560 works well in Mojave(iMacPro1,1)/Catalina(MacPro7,1) both with metal support(also the hevc and h264 acceralation) and DP/HDMI sound, even without Lilu/WEG/AppleALC. It should work properly for RX570/RX580 too(other cards such as nvidia and vms should doing thos too to get rid of problems). You guys have to make sure in vm xml the graphic sound device are in same bus as video but different function.Then the sounds will work OOB. <hostdev mode='subsystem' type='pci' managed='yes'> <driver name='vfio'/> <source> <address domain='0x0000' bus='0x2d' slot='0x00' function='0x0'/> </source> <rom file='/mnt/user/domains/mojave/Sapphire.RX560.4096.170419.rom'/> <address type='pci' domain='0x0000' bus='0x01' slot='0x00' function='0x0' multifunction='on'/> </hostdev> <hostdev mode='subsystem' type='pci' managed='yes'> <driver name='vfio'/> <source> <address domain='0x0000' bus='0x2d' slot='0x00' function='0x1'/> </source> <address type='pci' domain='0x0000' bus='0x01' slot='0x00' function='0x1'/> </hostdev> like the xml I've post , the video (0000:2d:00) and sound (0000:2d:01) should in same bus "0x01" but different function (video is "0x0" and sound is "0x1"). The unraid GUI won't do it automaticly(it just split the two functions of the card into multiple difference bus,maybe we can make it enhanment?) so you guys have to mannually edit the xml. After doing that ,you will find the AMD reset bug hardly appears,the sound works, and hardware acceralation works OOB without any kext, everything works like charm. Edit: I found someone have mentioned before, from here.

I also find virt-net is support in catalina, but when I add a virtio-pci-net in unraid webui xml the vm crashed. So I hope it will support properly in later version then we can get 10G even 100G network support for our VM.

You have to make sure that the graphic audio is using same bus and difference function as graphic gfx video in xml file. Then hdmi should work properly

First, rename bzimage and bzmodules in /boot to backup the stock kernel; Then unzip the archive to /boot/ Finally reboot your machine. I try to add an entry in startmenu but I don't know how to specify the bzmodule to load

Hi I just build a kernel to support X570 motherboard (mine is msi x570 ace) and latest AMD Ryzen 2 3000 family CPU. 6.8.3 is out, here is the new kernel and some tweaks: Add Vega Reset Patch Add Navi Reset Patch Enable NFSv4 in kernel(God damned, we finnaly get nfsv4 to work) Add R8125 out tree driver. AMD onboard audio/usb controller flr patch. Provide two version (linux-5.5.8 and linux-4.19.108) in case of bug. Notice that linux-4.19.108 still don't have AMD Zen 2 suppport. Download url can be found in the latest comments. For those who want to use NFSv4: NFSv4 have some change compared to v2/v3, it must have root share and the nfs-utils version can't handle it well.You must add this script to UserScript Plugin triggered when array start: #!/bin/bash # Add NFSv4 root echo '"/mnt/user/" -async,no_subtree_check,fsid=0 *(sec=sys,rw,insecure,anongid=100,anonuid=99,all_squash,crossmnt)' >> /etc/exports # Load configuration exportfs -ra # Tunning mkdir -p /var/lib/nfs/v4recovery sed -i 's@/usr/sbin/rpc.nfsd 8@/usr/sbin/rpc.nfsd 32@g' /etc/rc.d/rc.nfsd # Restart Services /etc/rc.d/rc.nfsd restart And if you have trouble mount nfsv4, you can specify mount vers=3 on client. Edit at 2020.01.31 The procedure of compiling kernel(you can compile it in other linux distribute or linux VMs): download kernel sources from kernel.org, notice that should be same or related version of unraid(like 4.19.94 which 6.8.2 used) unarchive the kernel source zip, like kernel-4.19.94/ copy the patches and .config (important!) from unraid server which located at /usr/src (like /usr/src/linux-4.19.94-Unraid/) to step 2 source directory (Optional) copy custom patches like navi patches or others to source directory too apply patches: find . -type f -iname '*.patch' -print0|xargs -n1 -0 patch -p 1 -i use old config : make oldconfig compile the kernel and modules: make -j5 bzImage; make -j5; make -j5 modules installing the modules, then you can find the module directory in /lib/modules: sudo make modules_install Copy the kernel image: cp sources/linux-4.19.94/arch/x86_64/boot/bzImage releases/6.8.2-4.19.94/bzimage (Optional) ThirdParty modules compiling (like nic r8125 outtree driver): enter the thirdparty driver directory compile the module: make -C /lib/modules/4.19.94-Unraid//build M=(pwd)/src modules install the module to direcotry: sudo make -C /lib/modules/4.19.94-Unraid/build M=(pwd)/src INSTALL_MOD_DIR=kernel/drivers/net/ethernet/realtek/ modules_install you can check whether the module exists in /lib/modules/4.19.94-Unraid/kernel/drivers/net/ethernet/realtek/ archive the modules to bzimage: mksquashfs /lib/modules/4.19.94-Unraid/ releases/4.19.94/bzmodules -keep-as-directory -noappend Then you get the bzimage and bzmodules, copy it to unraid server: /boot/

I found something interesting according here, you should use qcow2 instead of raw to make new version clover work. But it not seems to recognized my apfs dirver, only hfs+ works