Hi Satish,

Since you already set "cpu_mode = host-passthrough", there's no need 
to set cpu_model.

BTW, we're not known about the CPU topology a lot. But IME we always 
set "hw_cpu_sockets = 2" in specified image or flavor metadata if running
Windows instance. In default, KVM always allocate all vcpus into sockets 
in CPU topology, and this will affect the Windows VM performance since
Windows only support maximum 2 CPU sockets.

Perhaps you can try limit socket numbers by setting hw_cpu_sockets in
image metadata (or hw:cpu_sockets in flavor metadata.)

Satish Patel <satish.txt@gmail.com> 於 2020年3月5日 週四 下午10:46寫道：

cpu_mode = cpu-passthrough 
cpu_model = none

Do you think cpu_model make difference ?


Sent from my iPhone

On Mar 5, 2020, at 7:18 AM, Satish Patel<satish.txt@gmail.com> wrote:



cpu-passthrough 

Sent from my iPhone

On Mar 4, 2020, at 9:24 PM, rui zang <rui.zang@yandex.com> wrote:


Hi,
 
What is the value for the "cpu_mode" configuration option?
 
Thanks,
Zang, Rui
 
 
05.03.2020, 01:24, "Satish Patel" <satish.txt@gmail.com>:

Folks,

We are running openstack with KVM and i have noticed kvm presenting
wrong CPU Tolopoly to VM and because of that we are seeing bad
performance to our application.

This is openstack compute:

# lstopo-no-graphics --no-io
Machine (64GB total)
  NUMANode L#0 (P#0 32GB) + Package L#0 + L3 L#0 (25MB)
    L2 L#0 (256KB) + L1d L#0 (32KB) + L1i L#0 (32KB) + Core L#0
      PU L#0 (P#0)
      PU L#1 (P#20)
    L2 L#1 (256KB) + L1d L#1 (32KB) + L1i L#1 (32KB) + Core L#1
      PU L#2 (P#1)
      PU L#3 (P#21)

This is VM running on above compute

# lstopo-no-graphics --no-io
Machine (59GB total)
  NUMANode L#0 (P#0 29GB) + Package L#0 + L3 L#0 (16MB)
    L2 L#0 (4096KB) + Core L#0
      L1d L#0 (32KB) + L1i L#0 (32KB) + PU L#0 (P#0)
      L1d L#1 (32KB) + L1i L#1 (32KB) + PU L#1 (P#1)
    L2 L#1 (4096KB) + Core L#1
      L1d L#2 (32KB) + L1i L#2 (32KB) + PU L#2 (P#2)
      L1d L#3 (32KB) + L1i L#3 (32KB) + PU L#3 (P#3)

if you noticed P#0 and P#1 has own (32KB) cache per thread that is
wrong presentation if you compare with physical CPU.

This is a screenshot of AWS vs Openstack CPU Topology and looking at
openstack its presentation is little odd, is that normal?

https://imgur.com/a/2sPwJVC

I am running CentOS7.6 with kvm 2.12 version.