Add different cpus worker node into the same mcp, you will see the warn message "Profiles with bootcmdline conflict" when executing <oc get co/node-tuning>, the message will disappear when removing one worker node from mcp. but sometime the warn message won't disappear. 

1. label current one worker node as <node-role.kubernetes.io/worker-diffcpus=> as first node

2. add a new machineset 
MACHINE_SET=`oc get machineset -n openshift-machine-api |grep worker | awk '{print $1}' | sort | tail -1`
oc get machineset $MACHINE_SET -n openshift-machine-api -o json>./differnode.json
cat ./differnode.json |grep $MACHINE_SET
echo "Replace name "
sed -i "s/$MACHINE_SET/openshift-psap-qe-gpu-node01/g"./differnode.json
grep openshift-psap-qe-gpu-node01 ./differnode.json
sed -i 's/"instanceType":.*/"instanceType": "m6i.2xlarge",/' ./differnode.json
sed -i 's/"replicas": 2,/"replicas": 1,/' ./differnode.json
grep instanceType ./differnode.json
oc get machines -n openshift-machine-api |grep m6i.2xlarge 

3. oc create -f ./differnode.json

4.label the new worker node as <node-role.kubernetes.io/worker-diffcpus=> as second worker node

5. Create mcp 
oc apply f<<EOF
apiVersion: machineconfiguration.openshift.io/v1
kind: MachineConfigPool
metadata:
  name: worker-diffcpus
  labels:
    worker-diffcpus: ""
spec:
  machineConfigSelector:
    matchExpressions:
      - {key: machineconfiguration.openshift.io/role, operator: In, values: [worker,worker-diffcpus]}
  nodeSelector:
    matchLabels:
      node-role.kubernetes.io/worker-diffcpus: ""
EOF
6. Create custom profile 
apiVersion: tuned.openshift.io/v1
kind: Tuned
metadata:
  name: openshift-bootcmdline-cpu
  namespace: openshift-cluster-node-tuning-operator
spec:
  profile:
  - data: |
      [main]
      summary=Custom OpenShift profile
      [bootloader]
      cmdline=+cpus=${f:exec:/usr/bin/bash:-c:nproc|tr -d '\n'}
    name: openshift-bootcmdline-cpu  recommend:
  - machineConfigLabels:
      machineconfiguration.openshift.io/role: "worker-diffcpus"
    priority: 20
    profile: openshift-bootcmdline-cpu

7. check NTO operator pod and oc get co/node-tuning, you will see the WARN message "Profiles with bootcmdline conflict"

8. unlabel second worker. It will remove from mcp
9. wait for status of mcp is ready
10. unabel first worker node and wait for mcp is ready
11. delete mcp 
Repeat step 1-11 again, the warn message "Profiles with bootcmdline conflict" will not disappear when executing oc get co/node-tuning 

the warn message "Profiles with bootcmdline conflict" will not disappear when executing oc get co/node-tuning 

the warn message "Profiles with bootcmdline conflict" will disappear when executing oc get co/node-tuning