Skip to content

Commit 45afa94

Browse files
authored
Merge pull request GoogleCloudPlatform#4817 from parulbajaj01/parul/a4x
Update nccl-rdma manifest in gke a4x
2 parents 3c55837 + 92f5fb2 commit 45afa94

2 files changed

Lines changed: 1 addition & 15 deletions

File tree

examples/gke-a4x/gke-a4x.yaml

Lines changed: 0 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -289,8 +289,6 @@ deployment_groups:
289289
guest_accelerator:
290290
- type: nvidia-gb200
291291
count: 4
292-
gpu_driver_installation_config:
293-
gpu_driver_version: DEFAULT
294292
reservation_affinity:
295293
consume_reservation_type: SPECIFIC_RESERVATION
296294
specific_reservations:

examples/gke-a4x/nccl-rdma-installer.yaml

Lines changed: 1 addition & 13 deletions
Original file line numberDiff line numberDiff line change
@@ -62,19 +62,7 @@ spec:
6262
path: /home/kubernetes/bin/nvidia
6363
type: Directory
6464
initContainers:
65-
- name: disable-log-martian
66-
image: alpine:latest
67-
command: ["/bin/sh"]
68-
securityContext:
69-
privileged: true
70-
args:
71-
- -c
72-
- |
73-
sysctl -w net.ipv4.conf.gpu0rdma0.log_martians=0
74-
sysctl -w net.ipv4.conf.gpu1rdma0.log_martians=0
75-
sysctl -w net.ipv4.conf.gpu2rdma0.log_martians=0
76-
sysctl -w net.ipv4.conf.gpu3rdma0.log_martians=0
77-
- image: us-docker.pkg.dev/gce-ai-infra/gpudirect-gib/nccl-plugin-gib-arm64:v1.0.4
65+
- image: us-docker.pkg.dev/gce-ai-infra/gpudirect-gib/nccl-plugin-gib-arm64:v1.0.7
7866
name: nccl-rdma-installer
7967
resources:
8068
requests:

0 commit comments

Comments
 (0)