[SOLVED] NUMA mapping with Threadripper 2950X

[Chais]

I'm having trouble getting NUMA mapping to work. I'm getting the error message

kvm: -numa node,nodeid=0,cpus=4-7,cpus=20-23,memdev=ram-node0: CPU index (20) should be smaller than maxcpus (8)

Here's my topolgy:

Machine (47GB total) + Package L#0
  NUMANode L#0 (P#0 31GB)
    L3 L#0 (8192KB)
      L2 L#0 (512KB) + L1d L#0 (32KB) + L1i L#0 (64KB) + Core L#0
        PU L#0 (P#0)
        PU L#1 (P#16)
      L2 L#1 (512KB) + L1d L#1 (32KB) + L1i L#1 (64KB) + Core L#1
        PU L#2 (P#1)
        PU L#3 (P#17)
      L2 L#2 (512KB) + L1d L#2 (32KB) + L1i L#2 (64KB) + Core L#2
        PU L#4 (P#2)
        PU L#5 (P#18)
      L2 L#3 (512KB) + L1d L#3 (32KB) + L1i L#3 (64KB) + Core L#3
        PU L#6 (P#3)
        PU L#7 (P#19)
    L3 L#1 (8192KB)
      L2 L#4 (512KB) + L1d L#4 (32KB) + L1i L#4 (64KB) + Core L#4
        PU L#8 (P#4)
        PU L#9 (P#20)
      L2 L#5 (512KB) + L1d L#5 (32KB) + L1i L#5 (64KB) + Core L#5
        PU L#10 (P#5)
        PU L#11 (P#21)
      L2 L#6 (512KB) + L1d L#6 (32KB) + L1i L#6 (64KB) + Core L#6
        PU L#12 (P#6)
        PU L#13 (P#22)
      L2 L#7 (512KB) + L1d L#7 (32KB) + L1i L#7 (64KB) + Core L#7
        PU L#14 (P#7)
        PU L#15 (P#23)
    HostBridge L#0
      PCIBridge
        PCI 1022:43b6
          Block(Disk) L#0 "sda"
          Block(Disk) L#1 "sdd"
          Block(Disk) L#2 "sdb"
          Block(Disk) L#3 "sdc"
        PCIBridge
          PCIBridge
            PCI 8086:1539
              Net L#4 "enp5s0"
      PCIBridge
        PCIBridge
          PCIBridge
            PCI 1002:687f
      PCIBridge
        PCI 1022:7901
  NUMANode L#1 (P#1 16GB)
    L3 L#2 (8192KB)
      L2 L#8 (512KB) + L1d L#8 (32KB) + L1i L#8 (64KB) + Core L#8
        PU L#16 (P#8)
        PU L#17 (P#24)
      L2 L#9 (512KB) + L1d L#9 (32KB) + L1i L#9 (64KB) + Core L#9
        PU L#18 (P#9)
        PU L#19 (P#25)
      L2 L#10 (512KB) + L1d L#10 (32KB) + L1i L#10 (64KB) + Core L#10
        PU L#20 (P#10)
        PU L#21 (P#26)
      L2 L#11 (512KB) + L1d L#11 (32KB) + L1i L#11 (64KB) + Core L#11
        PU L#22 (P#11)
        PU L#23 (P#27)
    L3 L#3 (8192KB)
      L2 L#12 (512KB) + L1d L#12 (32KB) + L1i L#12 (64KB) + Core L#12
        PU L#24 (P#12)
        PU L#25 (P#28)
      L2 L#13 (512KB) + L1d L#13 (32KB) + L1i L#13 (64KB) + Core L#13
        PU L#26 (P#13)
        PU L#27 (P#29)
      L2 L#14 (512KB) + L1d L#14 (32KB) + L1i L#14 (64KB) + Core L#14
        PU L#28 (P#14)
        PU L#29 (P#30)
      L2 L#15 (512KB) + L1d L#15 (32KB) + L1i L#15 (64KB) + Core L#15
        PU L#30 (P#15)
        PU L#31 (P#31)
    HostBridge L#8
      PCIBridge
        PCIBridge
          PCIBridge
            PCI 1002:687f
      PCIBridge
        PCI 1022:7901

And here's my config

agent: 1
balloon: 0
bios: ovmf
boot: dc
bootdisk: scsi0
cores: 8
cpu: host
efidisk0: local-lvm:vm-100-disk-0,size=128K
hostpci0: 0b:00,x-vga=1,romfile=RVG64GO.F2,pcie=1
localtime: 1
machine: q35
memory: 12288
name: Chais-Win10
net0: virtio=02:53:BF:83:1D:EB,bridge=vmbr0
net1: virtio=FA:39:D1:ED:ED:A8,bridge=vmbr1
numa: 1
numa0: cpus=4-7;20-23,hostnodes=0,memory=12288,policy=bind
ostype: win10
sata0: none,media=cdrom
scsi0: /dev/sdb,backup=0,replicate=0,size=976762584K
scsihw: virtio-scsi-pci
sockets: 1
usb0: host=045e:0745

I don't quite understand why I'm getting this error. According to the documentation the NUMA configuration should be correct. So what am I missing here?

Edit:
When setting cores to 32 in order to be able to access all cores with the NUMA config I get the following errors:

kvm: warning: CPU(s) not present in any NUMA nodes: CPU 0 [socket-id: 0, core-id: 0, thread-id: 0], CPU 1 [socket-id: 0, core-id: 1, thread-id: 0], CPU 2 [socket-id: 0, core-id: 2, thread-id: 0], CPU 3 [socket-id: 0, core-id: 3, thread-id: 0], CPU 8 [socket-id: 0, core-id: 8, thread-id: 0], CPU 9 [socket-id: 0, core-id: 9, thread-id: 0], CPU 10 [socket-id: 0, core-id: 10, thread-id: 0], CPU 11 [socket-id: 0, core-id: 11, thread-id: 0], CPU 12 [socket-id: 0, core-id: 12, thread-id: 0], CPU 13 [socket-id: 0, core-id: 13, thread-id: 0], CPU 14 [socket-id: 0, core-id: 14, thread-id: 0], CPU 15 [socket-id: 0, core-id: 15, thread-id: 0], CPU 16 [socket-id: 0, core-id: 16, thread-id: 0], CPU 17 [socket-id: 0, core-id: 17, thread-id: 0], CPU 18 [socket-id: 0, core-id: 18, thread-id: 0], CPU 19 [socket-id: 0, core-id: 19, thread-id: 0], CPU 24 [socket-id: 0, core-id: 24, thread-id: 0], CPU 25 [socket-id: 0, core-id: 25, thread-id: 0], CPU 26 [socket-id: 0, core-id: 26, thread-id: 0], CPU 27 [socket-id: 0, core-id: 27, thread-id: 0], CPU 28 [socket-id: 0, core-id: 28, thread-id: 0], CPU 29 [socket-id: 0, core-id: 29, thread-id: 0], CPU 30 [socket-id: 0, core-id: 30, thread-id: 0], CPU 31 [socket-id: 0, core-id: 31, thread-id: 0]
kvm: warning: All CPU(s) up to maxcpus should be described in NUMA config, ability to start up with partial NUMA mappings is obsoleted and will be removed in future

 

[dcsapak]

numa0: cpus=4-7;20-23,hostnodes=0,memory=12288,policy=bind

i believe there is a misunderstanding

the cpus here are the guest cpus, not the host cpus
afair, qemu automatically binds the cpus to the correct numa nodes on the host

 

[Chais]

Ok. Then how do I expose host cores 4-7 and 20-23 to the guest as cores 0-7? And how do I make sure all memory allocated to the VM is on the same host NUMA node?
Not only do I want this for performance reasons, the VM also won't boot properly when given 16GB RAM, but works fine with 8GB. I suspect this is a result of the memory residing on different NUMA nodes.

 

[dcsapak]

Then how do I expose host cores 4-7 and 20-23 to the guest as cores 0-7?

again, afair qemu does this automatically when given a hostnode

eg this config:

numa0: cpus=0-7,hostnodes=0,memory=12288,policy=bind

 

[Chais]

That did it. Thanks a lot.