You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@yunikorn.apache.org by GitBox <gi...@apache.org> on 2022/10/20 06:07:37 UTC
[GitHub] [yunikorn-site] 0yukali0 commented on a diff in pull request #198: [YUNIKORN-1355] Generic example of GPU scheduling with Yunikorn
0yukali0 commented on code in PR #198:
URL: https://github.com/apache/yunikorn-site/pull/198#discussion_r1000177727
##########
docs/user_guide/workloads/run_nvidia.md:
##########
@@ -0,0 +1,255 @@
+---
+id: run_nvidia
+title: Run NVIDIA GPU Scheduling Jobs
+description: How to run generic example of GPU scheduling with Yunikorn.
+keywords:
+ - NVIDIA GPU
+---
+
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+To know how the GPU scheduling works, please refer to [**Time-Slicing GPUs in Kubernetes | Introduction**](https://docs.nvidia.com/datacenter/cloud-native/gpu-operator/gpu-sharing.html#introduction). This page covers ways to enable GPU scheduling in Yunikorn using [**NVIDIA GPU Operator**](https://catalog.ngc.nvidia.com/orgs/nvidia/containers/gpu-operator).
+
+:::info
+Before following this guide, Yunikorn need to deploy on the [**Kubernetes with GPUs**](https://docs.nvidia.com/datacenter/cloud-native/kubernetes/install-k8s.html#install-kubernetes).
+:::
+
+## Configuration
+Specify multiple configurations in a `ConfigMap` as in the following example.
+```yaml
+# time-slicing-config.yaml
+apiVersion: v1
+kind: ConfigMap
+metadata:
+ name: time-slicing-config
+ namespace: nvidia
+data:
+ a100-40gb: |-
+ version: v1
+ sharing:
+ timeSlicing:
+ resources:
+ - name: nvidia.com/gpu
+ replicas: 8
+ - name: nvidia.com/mig-1g.5gb
+ replicas: 2
+ - name: nvidia.com/mig-2g.10gb
+ replicas: 2
+ - name: nvidia.com/mig-3g.20gb
+ replicas: 3
+ - name: nvidia.com/mig-7g.40gb
+ replicas: 7
+ rtx-3070: |-
+ version: v1
+ sharing:
+ timeSlicing:
+ resources:
+ - name: nvidia.com/gpu
+ replicas: 8
+```
+:::note
+MIG support was added to Kubernetes in 2020. Refer to [**Supporting MIG in Kubernetes**](https://www.google.com/url?q=https://docs.google.com/document/d/1mdgMQ8g7WmaI_XVVRrCvHPFPOMCm5LQD5JefgAh6N8g/edit&sa=D&source=editors&ust=1655578433019961&usg=AOvVaw1F-OezvM-Svwr1lLsdQmu3) for details on how this works.
+:::
+
+Create a `ConfigMap` in the operator namespace.
+```bash
+kubectl create namespace nvidia
+kubectl create -f time-slicing-config.yaml
+```
+
+## Install NVIDIA GPU Operator
+Add the nvidia-gpu-operator helm repository.
+```bash
+helm repo add nvidia https://helm.ngc.nvidia.com/nvidia
+helm repo update
+helm repo list
+```
+
+Enabling shared access to GPUs with the NVIDIA GPU Operator.
+- During fresh install of the NVIDIA GPU Operator with time-slicing enabled.
+ ```bash
+ helm install gpu-operator nvidia/gpu-operator \
+ -n nvidia \
+ --set devicePlugin.config.name=time-slicing-config
+ ```
+
+- For dynamically enabling time-slicing with GPU Operator already installed.
+ ```bash
+ kubectl patch clusterpolicy/cluster-policy \
+ -n nvidia --type merge \
+ -p '{"spec": {"devicePlugin": {"config": {"name": "time-slicing-config"}}}}'
+ ```
+
+## Applying the Time-Slicing Configuration
+There are two methods:
+- Across the Cluster
+
+ Install the GPU Operator by passing the time-slicing `ConfigMap` name and the default configuration.
+ ```bash
+ kubectl patch clusterpolicy/cluster-policy \
+ -n nvidia --type merge \
+ -p '{"spec": {"devicePlugin": {"config": {"name": "time-slicing-config", "default": "rtx-3070"}}}}'
+ ```
+
+- Per Node
Review Comment:
Applying time-slicing configuration on certain nodes
##########
docs/user_guide/workloads/run_nvidia.md:
##########
@@ -0,0 +1,255 @@
+---
+id: run_nvidia
+title: Run NVIDIA GPU Scheduling Jobs
+description: How to run generic example of GPU scheduling with Yunikorn.
+keywords:
+ - NVIDIA GPU
+---
+
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+To know how the GPU scheduling works, please refer to [**Time-Slicing GPUs in Kubernetes | Introduction**](https://docs.nvidia.com/datacenter/cloud-native/gpu-operator/gpu-sharing.html#introduction). This page covers ways to enable GPU scheduling in Yunikorn using [**NVIDIA GPU Operator**](https://catalog.ngc.nvidia.com/orgs/nvidia/containers/gpu-operator).
+
+:::info
+Before following this guide, Yunikorn need to deploy on the [**Kubernetes with GPUs**](https://docs.nvidia.com/datacenter/cloud-native/kubernetes/install-k8s.html#install-kubernetes).
+:::
+
+## Configuration
+Specify multiple configurations in a `ConfigMap` as in the following example.
Review Comment:
Add a description to explain how to change the yaml context according to the GPU types.
::: note
If the GPU type in nodes do not include the a100-40gb or rtx-3070, you could modify the yaml file based on existing GPU types.
For example, there are only multiple rtx-2080ti in the local kubernetes cluster.
MIG is not supported by rtx-2080ti, so it could not replace the a100-40gb.
Time slicing is supported by rtx-2080ti, so it could replace rtx-3070.
:::
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: reviews-unsubscribe@yunikorn.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org