Run NCCL on custom GKE clusters that use A3 Mega or A3 High

This page describes how to run NVIDIA Collective Communications Library (NCCL) tests on custom GKE clusters that use GPUDirect-TCPXO and GPUDirect-TCPX networking protocols. A custom GKE cluster is a cluster that you create by using gcloud commands.

You can use the tests that are described on this page for the following scenarios:

Before you begin

The tests on this page use JobSet and Kueue with Topology Aware Scheduling (TAS). Before running any tests, you must set up your cluster and do the following:

  1. Install JobSet.

  2. Install Kueue.

    kubectl apply --server-side -f https://github.com/kubernetes-sigs/kueue/releases/download/v0.16.5/manifests.yaml
    

Set up your cluster with Jobset and Kueue

After you install JobSet and Kueue, take the following steps:

  1. Save the following manifest as kueue-config.yaml:

    A3 High

    apiVersion: kueue.x-k8s.io/v1beta2
    kind: Topology
    metadata:
      name: "gke-default"
    spec:
      levels:
      - nodeLabel: "cloud.google.com/gce-topology-block"
      - nodeLabel: "cloud.google.com/gce-topology-subblock"
      - nodeLabel: "cloud.google.com/gce-topology-host"
      - nodeLabel: "kubernetes.io/hostname"
    ---
    apiVersion: kueue.x-k8s.io/v1beta2
    kind: ResourceFlavor
    metadata:
      name: a3-high-flavor
    spec:
      nodeLabels:
        cloud.google.com/gke-accelerator: nvidia-h100-80gb
      topologyName: "gke-default"
    ---
    apiVersion: kueue.x-k8s.io/v1beta2
    kind: ResourceFlavor
    metadata:
      name: a3-high-dws-flavor
    spec:
      nodeLabels:
        cloud.google.com/gke-accelerator: nvidia-h100-80gb
      topologyName: "gke-default"
      tolerations:
      - key: "cloud.google.com/gke-queued"
        operator: "Exists"
        effect: NoSchedule
    ---     
    apiVersion: kueue.x-k8s.io/v1beta2
    kind: AdmissionCheck
    metadata:
      name: dws-prov
    spec:
      controllerName: kueue.x-k8s.io/provisioning-request
      parameters:
        apiGroup: kueue.x-k8s.io
        kind: ProvisioningRequestConfig
        name: dws-config
    ---
    apiVersion: kueue.x-k8s.io/v1beta2
    kind: ProvisioningRequestConfig
    metadata:
      name: dws-config
    spec:
      provisioningClassName: queued-provisioning.gke.io
      podSetUpdates:
      - key: autoscaling.gke.io/provisioning-request
        valueFromProvisioningClassDetail: ResizeRequestName
      managedResources:
      - nvidia.com/gpu
    ---
    apiVersion: kueue.x-k8s.io/v1beta2
    kind: ClusterQueue
    metadata:
      name: cq-tas
    spec:
      namespaceSelector: {}
      clusterQueueingStrategy: BestEffortFIFO
      resourceGroups:
      - flavors:
        - name: a3-high-flavor
          resources:
          - name: "cpu"
            nominalQuota: 1000
          - name: "memory"
            nominalQuota: 1000Ti
          - name: "nvidia.com/gpu"
            nominalQuota: 1000
        - name: a3-high-dws-flavor
          resources:
          - name: "cpu"
            nominalQuota: 1000
          - name: "memory"
            nominalQuota: 1000Ti
          - name: "nvidia.com/gpu"
            nominalQuota: 1000
      admissionChecksStrategy:
        admissionChecks:
        - name: "dws-prov"
          onFlavors: [a3-high-dws-flavor]
    ---
    apiVersion: kueue.x-k8s.io/v1beta2
    kind: LocalQueue
    metadata:
      namespace: default
      name: lq-tas
    spec:
      clusterQueue: cq-tas
    

    A3 Mega

    apiVersion: kueue.x-k8s.io/v1beta2
    kind: Topology
    metadata:
      name: "gke-default"
    spec:
      levels:
      - nodeLabel: "cloud.google.com/gce-topology-block"
      - nodeLabel: "cloud.google.com/gce-topology-subblock"
      - nodeLabel: "cloud.google.com/gce-topology-host"
      - nodeLabel: "kubernetes.io/hostname"
    ---
    apiVersion: kueue.x-k8s.io/v1beta2
    kind: ResourceFlavor
    metadata:
      name: a3-mega-flavor
    spec:
      nodeLabels:
        cloud.google.com/gke-accelerator: nvidia-h100-mega-80gb
      topologyName: "gke-default"
    ---
    apiVersion: kueue.x-k8s.io/v1beta2
    kind: ResourceFlavor
    metadata:
      name: a3-mega-dws-flavor
    spec:
      nodeLabels:
        cloud.google.com/gke-accelerator: nvidia-h100-mega-80gb
      topologyName: "gke-default"
      tolerations:
      - key: "cloud.google.com/gke-queued"
        operator: "Exists"
        effect: NoSchedule
    ---     
    apiVersion: kueue.x-k8s.io/v1beta2
    kind: AdmissionCheck
    metadata:
      name: dws-prov
    spec:
      controllerName: kueue.x-k8s.io/provisioning-request
      parameters:
        apiGroup: kueue.x-k8s.io
        kind: ProvisioningRequestConfig
        name: dws-config
    ---
    apiVersion: kueue.x-k8s.io/v1beta2
    kind: ProvisioningRequestConfig
    metadata:
      name: dws-config
    spec:
      provisioningClassName: queued-provisioning.gke.io
      podSetUpdates:
      - key: autoscaling.gke.io/provisioning-request
        valueFromProvisioningClassDetail: ResizeRequestName
      managedResources:
      - nvidia.com/gpu
    ---
    apiVersion: kueue.x-k8s.io/v1beta2
    kind: ClusterQueue
    metadata:
      name: cq-tas
    spec:
      namespaceSelector: {}
      clusterQueueingStrategy: BestEffortFIFO
      resourceGroups:
      - flavors:
        - name: a3-mega-flavor
          resources:
          - name: "cpu"
            nominalQuota: 1000
          - name: "memory"
            nominalQuota: 1000Ti
          - name: "nvidia.com/gpu"
            nominalQuota: 1000
        - name: a3-mega-dws-flavor
          resources:
          - name: "cpu"
            nominalQuota: 1000
          - name: "memory"
            nominalQuota: 1000Ti
          - name: "nvidia.com/gpu"
            nominalQuota: 1000
      admissionChecksStrategy:
        admissionChecks:
        - name: "dws-prov"
          onFlavors: [a3-mega-dws-flavor]
    ---
    apiVersion: kueue.x-k8s.io/v1beta2
    kind: LocalQueue
    metadata:
      namespace: default
      name: lq-tas
    spec:
      clusterQueue: cq-tas
    
  2. Apply the manifest:

    kubectl apply -f kueue-config.yaml
    

When running workloads with TAS enabled, you can specify how strictly topology constraints are enforced by using one of the following annotations in your workload manifest:

  • kueue.x-k8s.io/podset-required-topology: If you use this annotation, Kueue blocks scheduling until the workload can be scheduled within the requested topology constraint. Use this annotation to ensure that pods are placed together for optimal performance.

  • kueue.x-k8s.io/podset-preferred-topology: If you use this annotation, Kueue attempts to schedule pods within the requested topology constraint, but if that's not possible, it admits the workload without meeting topology constraints.

Note: Avoid using the required mode with DWS Flex-start. Because Flex-start provisions nodes dynamically, the resulting nodes might not satisfy strict topological requirements, which can result in unschedulable workloads. For these configurations, use podset-preferred-topology instead.

For either annotation, specify one of the following values as the topology constraint:

  • cloud.google.com/gce-topology-block: Schedules pods within the same network block.
  • cloud.google.com/gce-topology-subblock: Schedules pods within the same rack.
  • cloud.google.com/gce-topology-host: Schedules pods on the same physical host.

Test on two Flex-start nodes

To run NCCL tests on a GKE cluster that uses A3 Mega or A3 High Flex-start VMs, use the following procedure. This procedure uses a JobSet manifest to run an NCCL test on two nodes.

  1. Save the following manifest as nccl-tas-jobset.yaml:

    A3 Mega

    apiVersion: v1
    kind: ConfigMap
    metadata:
      name: nccl-configmap
    data:
      allgather.sh: |
        #!/bin/bash
        service ssh restart;
        /scripts/init_ssh.sh ${@};
        pushd /scripts;
        /scripts/gen_hostfiles.sh ${@};
        popd;
        # Set up environment variables for GPUDirect-TCPXO
        export LD_LIBRARY_PATH=/usr/local/nvidia/lib64
        export NCCL_FASTRAK_CTRL_DEV=eth0
        export NCCL_FASTRAK_IFNAME=eth1,eth2,eth3,eth4,eth5,eth6,eth7,eth8
        export NCCL_SOCKET_IFNAME=eth0
        export NCCL_CROSS_NIC=0
        export NCCL_ALGO=Ring,Tree
        export NCCL_PROTO=Simple
        export NCCL_NET_GDR_LEVEL=PIX
        # Run the benchmark
        /scripts/demo-run-nccl-test-tcpxo-via-mpi.sh
    ---
    apiVersion: jobset.x-k8s.io/v1alpha2
    kind: JobSet
    metadata:
      name: nccl-tas-test
      labels:
        kueue.x-k8s.io/queue-name: lq-tas
    spec:
      ttlSecondsAfterFinished: 1200
      suspend: true
      network:
        enableDNSHostnames: true
      replicatedJobs:
        - name: worker
          replicas: 2
          template:
            spec:
              parallelism: 1
              completions: 1
              template:
                metadata:
                  annotations:
                    kueue.x-k8s.io/podset-preferred-topology: "cloud.google.com/gce-topology-block"
                    networking.gke.io/default-interface: 'eth0'
                    networking.gke.io/interfaces: |
                      [
                        {"interfaceName":"eth0","network":"default"},
                        {"interfaceName":"eth1","network":"vpc0"},
                        {"interfaceName":"eth2","network":"vpc1"},
                        {"interfaceName":"eth3","network":"vpc2"},
                        {"interfaceName":"eth4","network":"vpc3"},
                        {"interfaceName":"eth5","network":"vpc4"},
                        {"interfaceName":"eth6","network":"vpc5"},
                        {"interfaceName":"eth7","network":"vpc6"},
                        {"interfaceName":"eth8","network":"vpc7"}
                      ]
                spec:
                  activeDeadlineSeconds: 3600
                  restartPolicy: Never
                  nodeSelector:
                    cloud.google.com/gke-accelerator: nvidia-h100-mega-80gb
                  tolerations:
                  - key: cloud.google.com/gke-queued
                    effect: NoSchedule
                    value: "true"
                  - key: "nvidia.com/gpu"
                    operator: "Exists"
                    effect: "NoSchedule"
                  setHostnameAsFQDN: true
                  volumes:
                  - name: nvidia
                    hostPath:
                      path: /home/kubernetes/bin/nvidia
                  - name: lib64
                    hostPath:
                      path: /lib64
                  - name: proc
                    hostPath:
                      path: /proc
                  - name: shared-memory
                    emptyDir:
                      medium: "Memory"
                      sizeLimit: 250Gi
                  - name: nccl-config
                    configMap:
                      name: nccl-configmap
                      defaultMode: 0755
                  containers:
                  - name: nccl-test
                    image: us-docker.pkg.dev/gce-ai-infra/gpudirect-tcpxo/nccl-plugin-gpudirecttcpx-dev:v1.0.15
                    stdin: true
                    tty: true
                    securityContext:
                      privileged: true
                    env:
                    - name: LD_LIBRARY_PATH
                      value: /usr/local/nvidia/lib64
                    volumeMounts:
                    - name: nvidia
                      mountPath: /usr/local/nvidia
                    - name: shared-memory
                      mountPath: /dev/shm
                    - name: nccl-config
                      mountPath: /configs
                    resources:
                      limits:
                        cpu: "200"
                        memory: "3700Gi"
                        nvidia.com/gpu: 8
                      requests:
                        cpu: "200"
                        memory: "3700Gi"
                        nvidia.com/gpu: 8
                  - name: tcpxo-daemon
                    image: us-docker.pkg.dev/gce-ai-infra/gpudirect-tcpxo/tcpgpudmarxd-dev:v1.0.21
                    imagePullPolicy: Always
                    command: ["/bin/sh", "-c"]
                    args:
                      - |
                        set -ex
                        chmod 755 /fts/entrypoint_rxdm_container.sh
                        /fts/entrypoint_rxdm_container.sh --num_hops=2 --num_nics=8 --uid= --alsologtostderr
                    securityContext:
                      privileged: true
                      capabilities:
                        add:
                          - NET_ADMIN
                          - NET_BIND_SERVICE
                    volumeMounts:
                    - name: nvidia
                      mountPath: /usr/local/nvidia/lib64
                    - name: proc
                      mountPath: /proc
                    env:
                    - name: LD_LIBRARY_PATH
                      value: /usr/local/nvidia/lib64
    

    A3 High

    apiVersion: v1
    kind: ConfigMap
    metadata:
      name: nccl-config
    data:
      allgather.sh: |
        #!/bin/bash
        for script in /configs/*; do
          name=$(basename $script)
          cp $script "/scripts/$name"
          chmod +x "/scripts/$name"
        done
        /scripts/init_ssh.sh ${@};
        pushd /scripts;
        /scripts/gen_hostfiles.sh ${@};
        popd;
        /scripts/run-allgather.sh 8 eth1,eth2,eth3,eth4 1M 512M ${#};
    ---
    apiVersion: jobset.x-k8s.io/v1alpha2
    kind: JobSet
    metadata:
      name: nccl-tas-test
      labels:
        kueue.x-k8s.io/queue-name: lq-tas
    spec:
      suspend: true
      network:
        enableDNSHostnames: true
      replicatedJobs:
      - name: worker
        replicas: 2
        template:
          spec:
            parallelism: 1
            completions: 1
            template:
              metadata:
                annotations:
                  kueue.x-k8s.io/podset-preferred-topology: "cloud.google.com/gce-topology-block"
                  networking.gke.io/default-interface: 'eth0'
                  networking.gke.io/interfaces: |
                    [
                      {"interfaceName":"eth0","network":"default"},
                      {"interfaceName":"eth1","network":"vpc0"},
                      {"interfaceName":"eth2","network":"vpc1"},
                      {"interfaceName":"eth3","network":"vpc2"},
                      {"interfaceName":"eth4","network":"vpc3"}
                    ]
              spec:
                terminationGracePeriodSeconds: 0
                nodeSelector:
                  cloud.google.com/gke-accelerator: nvidia-h100-80gb
                tolerations:
                - key: cloud.google.com/gke-queued
                  effect: NoSchedule
                  value: "true"
                - key: "nvidia.com/gpu"
                  operator: "Exists"
                  effect: "NoSchedule"
                setHostnameAsFQDN: true
                containers:
                - name: tcpx-daemon
                  image: us-docker.pkg.dev/gce-ai-infra/gpudirect-tcpx/tcpgpudmarxd-dev:v2.0.11
                  command:
                    - /tcpgpudmarxd/build/app/tcpgpudmarxd
                    - --gpu_nic_preset
                    - a3vm
                    - --gpu_shmem_type
                    - fd
                    - --uds_path
                    - /run/tcpx
                    - --setup_param
                    - "--verbose 128 2 0 "
                  securityContext:
                    privileged: true
                    capabilities:
                      add:
                        - NET_ADMIN
                  volumeMounts:
                    - name: libraries
                      mountPath: /usr/local/nvidia/lib64
                    - name: tcpx-socket
                      mountPath: /run/tcpx
                    - name: sys
                      mountPath: /hostsysfs
                    - name: proc-sys
                      mountPath: /hostprocsysfs
                  env:
                    - name: LD_LIBRARY_PATH
                      value: /usr/local/nvidia/lib64
                - name: nccl-test
                  image: us-docker.pkg.dev/gce-ai-infra/gpudirect-tcpx/nccl-plugin-gpudirecttcpx-dev:v3.1.8
                  command:
                    - bash
                    - -c
                    - |
                      /scripts/container_entry.sh daemon;
                      sleep infinity;
                  securityContext:
                    privileged: true
                  volumeMounts:
                    - name: tcpx-socket
                      mountPath: /tmp
                    - name: libraries
                      mountPath: /usr/local/nvidia/lib64
                    - name: nccl-config
                      mountPath: /configs
                    - name: shared-memory
                      mountPath: /dev/shm
                  resources:
                    limits:
                      cpu: "200"
                      memory: "1800Gi"
                      nvidia.com/gpu: 8
                    requests:
                      cpu: "200"
                      memory: "1800Gi"
                      nvidia.com/gpu: 8
                volumes:
                - name: libraries
                  hostPath:
                    path: /home/kubernetes/bin/nvidia/lib64
                - name: tcpx-socket
                  emptyDir: {}
                - name: sys
                  hostPath:
                    path: /sys
                - name: proc-sys
                  hostPath:
                    path: /proc/sys
                - name: shared-memory
                  emptyDir:
                    medium: Memory
                    sizeLimit: 250Gi
                - name: nccl-config
                  configMap:
                    name: nccl-config
                    defaultMode: 0777
    
  2. Apply the manifest to your cluster:

    kubectl apply -f nccl-tas-jobset.yaml
    
  3. Check that the JobSet is admitted and running:

    kubectl get jobset nccl-tas-test
    

    Wait for the JobSet to be unsuspended and Pods to reach the Running status.

  4. Trigger the NCCL test by executing the allgather.sh script from the first worker Pod:

    kubectl exec --stdin --tty --container=nccl-test nccl-tas-test-worker-0-0 -- /configs/allgather.sh nccl-tas-test-worker-0-0 nccl-tas-test-worker-1-0
    

    The output for a two-node test is similar to the following:

    A3 Mega

    #                                                              out-of-place                       in-place
    #        size         count      type   redop    root     time   algbw   busbw #wrong     time   algbw   busbw #wrong
    #         (B)    (elements)                               (us)  (GB/s)  (GB/s)            (us)  (GB/s)  (GB/s)
        0                 0         float    none      -1     0.24    0.00    0.00      0     0.18    0.00    0.00      0
        ...
        8589934592     134217728    float    none      -1    42603  201.63  189.03      0    42670  201.31  188.73      0
    # Out of bounds values : 0 OK
    # Avg bus bandwidth    : 45.7587
    

    A3 High

    #                                                              out-of-place                       in-place
    #       size         count      type   redop    root     time   algbw   busbw #wrong     time   algbw   busbw #wrong
    #        (B)    (elements)                               (us)  (GB/s)  (GB/s)            (us)  (GB/s)  (GB/s)
        1048576         16384     float    none      -1    696.8    1.50    1.41      0    729.0    1.44    1.35      0
        ...
        536870912       8388608     float    none      -1   7101.7   75.60   70.87      0   7060.9   76.03   71.28      0
    # Out of bounds values : 0 OK
    # Avg bus bandwidth    : 29.8293
    

Deploy an NCCL test workload with TAS

If you have more than two nodes, we recommend using the following test, which uses Topology Aware Scheduling (TAS). To run NCCL tests with TAS on a GKE cluster that uses A3 Mega or A3 High Flex-start VMs, use the following procedure.

  1. Save the following manifest as nccl-jobset-test.yaml. Replace NUM_NODES with the number of nodes in the node pool:

    A3 Mega

    apiVersion: jobset.x-k8s.io/v1alpha2
    kind: JobSet
    metadata:
      name: nccl-ag
      labels:
        kueue.x-k8s.io/queue-name: lq-tas
    spec:
      ttlSecondsAfterFinished: 1200
      suspend: true
      network:
        enableDNSHostnames: true
      replicatedJobs:
        - name: worker
          template:
            spec:
              parallelism: NUM_NODES
              completions: NUM_NODES
              template:
                metadata:
                  annotations:
                    kueue.x-k8s.io/podset-preferred-topology: "cloud.google.com/gce-topology-subblock"
                    networking.gke.io/default-interface: 'eth0'
                    networking.gke.io/interfaces: |
                      [
                        {"interfaceName":"eth0","network":"default"},
                        {"interfaceName":"eth1","network":"vpc0"},
                        {"interfaceName":"eth2","network":"vpc1"},
                        {"interfaceName":"eth3","network":"vpc2"},
                        {"interfaceName":"eth4","network":"vpc3"},
                        {"interfaceName":"eth5","network":"vpc4"},
                        {"interfaceName":"eth6","network":"vpc5"},
                        {"interfaceName":"eth7","network":"vpc6"},
                        {"interfaceName":"eth8","network":"vpc7"}
                      ]
                spec:
                  activeDeadlineSeconds: 3600
                  restartPolicy: Never
                  nodeSelector:
                    cloud.google.com/gke-accelerator: nvidia-h100-mega-80gb
                  tolerations:
                  - key: "nvidia.com/gpu"
                    operator: "Exists"
                    effect: "NoSchedule"
                  setHostnameAsFQDN: true
                  volumes:
                  - name: proc
                    hostPath:
                      path: /proc
                  - name: nvidia
                    hostPath:
                      path: /home/kubernetes/bin/nvidia
                  - name: lib64
                    hostPath:
                      path: /lib64
                  - name: shared-memory
                    emptyDir:
                      medium: "Memory"
                      sizeLimit: 250Gi
                  containers:
                  - name: nccl-test
                    stdin: true
                    tty: true
                    image: us-docker.pkg.dev/gce-ai-infra/gpudirect-tcpxo/nccl-plugin-tcpxo-diagnostic:v1.0.6
                    securityContext:
                      privileged: true
                    env:
                    - name: MY_NODE_NAME
                      valueFrom:
                        fieldRef:
                          fieldPath: spec.nodeName
                    - name: OMPI_ALLOW_RUN_AS_ROOT
                      value: "1"
                    - name: OMPI_ALLOW_RUN_AS_ROOT_CONFIRM
                      value: "1"
                    - name: N_NODES
                      value: "NUM_NODES"
                    - name: NCCL_SOCKET_IFNAME
                      value: eth0
                    - name: NCCL_FASTRAK_CTRL_DEV
                      value: eth0
                    - name: NCCL_FASTRAK_IFNAME
                      value: eth1,eth2,eth3,eth4,eth5,eth6,eth7,eth8
                    - name: NCCL_CROSS_NIC
                      value: "0"
                    - name: NCCL_ALGO
                      value: Ring,Tree
                    - name: NCCL_PROTO
                      value: Simple
                    - name: NCCL_NET_GDR_LEVEL
                      value: PIX
                    - name: LD_LIBRARY_PATH
                      value: /usr/local/nvidia/lib64
                    command:
                    - bash
                    - -c
                    - |
                      set -x
                      /scripts/container_entry.sh daemon &
                      export POSTFIX=$(hostname | cut -d . -f 2-)
                      export WORKERS_BASENAME=$(hostname | cut -d . -f 1 | rev | cut -d - -f 2- | rev )
                      export NODE_RANK=$JOB_COMPLETION_INDEX
                      for i in `seq 0 $(($N_NODES-1))`; do
                        OTHER=<span class="katex"><span class="katex-html" aria-hidden="true"><span class="base"><span class="strut" style="height:0.8333em;vertical-align:-0.15em;"></span><span class="mord"><span class="mord mathnormal" style="margin-right:0.13889em;">W</span><span class="mord mathnormal" style="margin-right:0.00773em;">OR</span><span class="mord mathnormal" style="margin-right:0.07153em;">K</span><span class="mord mathnormal" style="margin-right:0.00773em;">ER</span><span class="mord"><span class="mord mathnormal" style="margin-right:0.05764em;">S</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height:0.3283em;"><span style="top:-2.55em;margin-left:-0.0576em;margin-right:0.05em;"><span class="pstrut" style="height:2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mathnormal mtight" style="margin-right:0.05017em;">B</span></span></span></span><span class="vlist-s">​</span></span><span class="vlist-r"><span class="vlist" style="height:0.15em;"><span></span></span></span></span></span></span><span class="mord mathnormal">A</span><span class="mord mathnormal" style="margin-right:0.10903em;">SEN</span><span class="mord mathnormal">A</span><span class="mord mathnormal" style="margin-right:0.05764em;">ME</span></span><span class="mord">−</span></span></span></span>{i}.${POSTFIX}
                        until ssh -p 222 -o StrictHostKeyChecking=no $OTHER hostname; do
                          sleep 10
                        done
                        echo ${OTHER} port=222 slots=8 | tee -a /tmp/hostfile;
                      done
                      if [[ "${NODE_RANK}" -eq "0" ]]; then
                          export NCCL_TESTS_SPLIT_MASK="0x0";
                          ENV_VARS=$(echo ${!NCCL*} ${!OMPI*} LD_LIBRARY_PATH PATH | sed 's/ / -x /g')
                          mpirun --hostfile /tmp/hostfile \
                            -x $ENV_VARS  \
                            -mca plm_rsh_no_tree_spawn 1 \
                            --mca orte_keep_fqdn_hostnames 1 \
                            --mca btl self,tcp \
                            --mca btl_tcp_if_include eth0 \
                            --bind-to none \
                            --mca plm_rsh_agent "ssh -q -o LogLevel=ERROR -o StrictHostKeyChecking=no -p 222" \
                            /third_party/nccl-tests/build/all_gather_perf -b 1K -e 8G -f 2 -g 1 -w 5 --iters 100 -c 1
                      else
                          while ping -c 1 <span class="katex"><span class="katex-html" aria-hidden="true"><span class="base"><span class="strut" style="height:0.8333em;vertical-align:-0.15em;"></span><span class="mord"><span class="mord mathnormal" style="margin-right:0.13889em;">W</span><span class="mord mathnormal" style="margin-right:0.00773em;">OR</span><span class="mord mathnormal" style="margin-right:0.07153em;">K</span><span class="mord mathnormal" style="margin-right:0.00773em;">ER</span><span class="mord"><span class="mord mathnormal" style="margin-right:0.05764em;">S</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height:0.3283em;"><span style="top:-2.55em;margin-left:-0.0576em;margin-right:0.05em;"><span class="pstrut" style="height:2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mathnormal mtight" style="margin-right:0.05017em;">B</span></span></span></span><span class="vlist-s">​</span></span><span class="vlist-r"><span class="vlist" style="height:0.15em;"><span></span></span></span></span></span></span><span class="mord mathnormal">A</span><span class="mord mathnormal" style="margin-right:0.10903em;">SEN</span><span class="mord mathnormal">A</span><span class="mord mathnormal" style="margin-right:0.05764em;">ME</span></span><span class="mspace" style="margin-right:0.2222em;"></span><span class="mbin">−</span><span class="mspace" style="margin-right:0.2222em;"></span></span><span class="base"><span class="strut" style="height:0.6444em;"></span><span class="mord">0.</span></span></span></span>{POSTFIX}; do
                          sleep 5
                      done
                      fi
                      exit 0
                    volumeMounts:
                    - name: nvidia
                      mountPath: /usr/local/nvidia
                    - name: lib64
                      mountPath: /lib64
                    - name: shared-memory
                      mountPath: /dev/shm
                    resources:
                      limits:
                        cpu: "200"
                        memory: "3700Gi"
                        nvidia.com/gpu: 8
                      requests:
                        cpu: "200"
                        memory: "3700Gi"
                        nvidia.com/gpu: 8
                  - name: tcpxo-daemon
                    image: us-docker.pkg.dev/gce-ai-infra/gpudirect-tcpxo/tcpxo-daemon:v1.0.1
                    imagePullPolicy: Always
                    command:
                    - bash
                    - -c
                    - |
                      /usr/bin/tcpxo_daemon
                    securityContext:
                      privileged: true
                    volumeMounts:
                    - name: nvidia
                      mountPath: /usr/local/nvidia
                    - name: proc
                      mountPath: /proc
                    env:
                    - name: LD_LIBRARY_PATH
                      value: /usr/local/nvidia/lib64
    

    A3 High

    apiVersion: jobset.x-k8s.io/v1alpha2
    kind: JobSet
    metadata:
      name: nccl-ag
      labels:
        kueue.x-k8s.io/queue-name: lq-tas
    spec:
      ttlSecondsAfterFinished: 1200
      suspend: true
      network:
        enableDNSHostnames: true
      replicatedJobs:
        - name: worker
          template:
            spec:
              parallelism: NUM_NODES
              completions: NUM_NODES
              template:
                metadata:
                  annotations:
                    kueue.x-k8s.io/podset-preferred-topology: "cloud.google.com/gce-topology-subblock"
                    networking.gke.io/default-interface: 'eth0'
                    networking.gke.io/interfaces: |
                      [
                        {"interfaceName":"eth0","network":"default"},
                        {"interfaceName":"eth1","network":"vpc0"},
                        {"interfaceName":"eth2","network":"vpc1"},
                        {"interfaceName":"eth3","network":"vpc2"},
                        {"interfaceName":"eth4","network":"vpc3"}
                      ]
                spec:
                  activeDeadlineSeconds: 3600
                  restartPolicy: Never
                  nodeSelector:
                    cloud.google.com/gke-accelerator: nvidia-h100-80gb
                  tolerations:
                  - key: "nvidia.com/gpu"
                    operator: "Exists"
                    effect: "NoSchedule"
                  setHostnameAsFQDN: true
                  volumes:
                  - name: proc
                    hostPath:
                      path: /proc
                  - name: nvidia
                    hostPath:
                      path: /home/kubernetes/bin/nvidia
                  - name: libraries
                    hostPath:
                      path: /home/kubernetes/bin/nvidia/lib64
                  - name: tcpx-socket
                    emptyDir: {}
                  - name: shared-memory
                    emptyDir:
                      medium: "Memory"
                      sizeLimit: 250Gi
                  containers:
                  - name: tcpx-daemon
                    image: us-docker.pkg.dev/gce-ai-infra/gpudirect-tcpx/tcpgpudmarxd-dev:v2.0.11
                    command:
                      - /tcpgpudmarxd/build/app/tcpgpudmarxd
                      - --gpu_nic_preset
                      - a3vm
                      - --uds_path
                      - /run/tcpx
                    securityContext:
                      privileged: true
                    volumeMounts:
                      - name: tcpx-socket
                        mountPath: /run/tcpx
                      - name: libraries
                        mountPath: /usr/local/nvidia/lib64
                  - name: nccl-test
                    stdin: true
                    tty: true
                    image: us-docker.pkg.dev/gce-ai-infra/gpudirect-tcpx/nccl-plugin-gpudirecttcpx-dev:v3.1.8
                    securityContext:
                      privileged: true
                    env:
                    - name: MY_NODE_NAME
                      valueFrom:
                        fieldRef:
                          fieldPath: spec.nodeName
                    - name: OMPI_ALLOW_RUN_AS_ROOT
                      value: "1"
                    - name: OMPI_ALLOW_RUN_AS_ROOT_CONFIRM
                      value: "1"
                    - name: N_NODES
                      value: "NUM_NODES"
                    - name: LD_LIBRARY_PATH
                      value: /usr/local/nvidia/lib64
                    command:
                    - bash
                    - -c
                    - |
                      /scripts/container_entry.sh daemon &
                      export POSTFIX=$(hostname | cut -d . -f 2-)
                      export WORKERS_BASENAME=$(hostname | cut -d . -f 1 | rev | cut -d - -f 2- | rev )
                      export NODE_RANK=$JOB_COMPLETION_INDEX
                      for i in `seq 0 $(($N_NODES-1))`; do
                        OTHER=<span class="katex"><span class="katex-html" aria-hidden="true"><span class="base"><span class="strut" style="height:0.8333em;vertical-align:-0.15em;"></span><span class="mord"><span class="mord mathnormal" style="margin-right:0.13889em;">W</span><span class="mord mathnormal" style="margin-right:0.00773em;">OR</span><span class="mord mathnormal" style="margin-right:0.07153em;">K</span><span class="mord mathnormal" style="margin-right:0.00773em;">ER</span><span class="mord"><span class="mord mathnormal" style="margin-right:0.05764em;">S</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height:0.3283em;"><span style="top:-2.55em;margin-left:-0.0576em;margin-right:0.05em;"><span class="pstrut" style="height:2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mathnormal mtight" style="margin-right:0.05017em;">B</span></span></span></span><span class="vlist-s">​</span></span><span class="vlist-r"><span class="vlist" style="height:0.15em;"><span></span></span></span></span></span></span><span class="mord mathnormal">A</span><span class="mord mathnormal" style="margin-right:0.10903em;">SEN</span><span class="mord mathnormal">A</span><span class="mord mathnormal" style="margin-right:0.05764em;">ME</span></span><span class="mord">−</span></span></span></span>{i}.${POSTFIX}
                        until ssh -p 222 -o StrictHostKeyChecking=no $OTHER hostname; do
                          sleep 10
                        done
                        echo ${OTHER} port=222 slots=8 | tee -a /tmp/hostfile;
                      done
                      if [[ "${NODE_RANK}" -eq "0" ]]; then
                          /scripts/run-allgather.sh 8 eth1,eth2,eth3,eth4 1M 512M ${N_NODES}
                      else
                          while ping -c 1 <span class="katex"><span class="katex-html" aria-hidden="true"><span class="base"><span class="strut" style="height:0.8333em;vertical-align:-0.15em;"></span><span class="mord"><span class="mord mathnormal" style="margin-right:0.13889em;">W</span><span class="mord mathnormal" style="margin-right:0.00773em;">OR</span><span class="mord mathnormal" style="margin-right:0.07153em;">K</span><span class="mord mathnormal" style="margin-right:0.00773em;">ER</span><span class="mord"><span class="mord mathnormal" style="margin-right:0.05764em;">S</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height:0.3283em;"><span style="top:-2.55em;margin-left:-0.0576em;margin-right:0.05em;"><span class="pstrut" style="height:2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mathnormal mtight" style="margin-right:0.05017em;">B</span></span></span></span><span class="vlist-s">​</span></span><span class="vlist-r"><span class="vlist" style="height:0.15em;"><span></span></span></span></span></span></span><span class="mord mathnormal">A</span><span class="mord mathnormal" style="margin-right:0.10903em;">SEN</span><span class="mord mathnormal">A</span><span class="mord mathnormal" style="margin-right:0.05764em;">ME</span></span><span class="mspace" style="margin-right:0.2222em;"></span><span class="mbin">−</span><span class="mspace" style="margin-right:0.2222em;"></span></span><span class="base"><span class="strut" style="height:0.6444em;"></span><span class="mord">0.</span></span></span></span>{POSTFIX}; do
                          sleep 5
                      done
                      fi
                      exit 0
                    volumeMounts:
                    - name: nvidia
                      mountPath: /usr/local/nvidia
                    - name: tcpx-socket
                      mountPath: /tmp
                    - name: libraries
                      mountPath: /usr/local/nvidia/lib64
                    - name: shared-memory
                      mountPath: /dev/shm
                    resources:
                      limits:
                        cpu: "200"
                        memory: "1800Gi"
                        nvidia.com/gpu: 8
                      requests:
                        cpu: "200"
                        memory: "1800Gi"
                        nvidia.com/gpu: 8
    
  2. Apply the manifest:

    kubectl apply -f nccl-jobset-test.yaml
    
  3. Check that the workload is admitted and reaches the Completed state.

  4. Fetch logs for the Pod matching nccl-ag-worker-0-0-.* to see the results:

    kubectl logs $(kubectl get pods -o go-template='{{range .items}}{{.metadata.name}}{{"\n"}}{{end}}' | grep nccl-ag-worker-0-0)
    

What's next