From ff0def677959e035a3e12304d5851739ff1d660a Mon Sep 17 00:00:00 2001 From: Katie Horne Date: Wed, 1 Dec 2021 13:17:32 -0600 Subject: [PATCH 1/3] uchore: update GPU doc --- .../workspace-management/gpu-acceleration.md | 26 +++++++++---------- 1 file changed, 13 insertions(+), 13 deletions(-) diff --git a/admin/workspace-management/gpu-acceleration.md b/admin/workspace-management/gpu-acceleration.md index fd55cfe04..47329746f 100644 --- a/admin/workspace-management/gpu-acceleration.md +++ b/admin/workspace-management/gpu-acceleration.md @@ -3,22 +3,22 @@ title: "GPU acceleration" description: Learn how to use GPUs with workspaces. --- -GPUs can be used for performing tasks such as machine learning within -workspaces. +Graphical processing units (GPUs) are useful with compute-intensive workloads, +such as those involved with data science/machine learning projects. -By default, users cannot create workspaces with GPUs until a site manager -configures and enables this feature. +You can allocate GPUs to workspaces once a site manager configures and enables +this feature. Enabling GPU acceleration requires that you've configured your +Kubernetes cluster appropriately and +[scheduled GPUs](https://kubernetes.io/docs/tasks/manage-gpus/scheduling-gpus/). -### Step 1: Configure your Kubernetes cluster +- [Amazon Elastic Container Service](https://docs.aws.amazon.com/AmazonECS/latest/developerguide/ecs-gpu.html) +- [Azure Kubernetes Service](https://docs.microsoft.com/en-us/azure/aks/gpu-cluster) +- [Google Kubernetes Engine](https://cloud.google.com/kubernetes-engine/docs/how-to/gpus) -Configure your Kubernetes cluster -with the available GPUs and all necessary device plugins and drivers. +Once your Kubernetes cluster has been configured, you can enable GPUs in Coder. -### Step 2: Enable GPU vendor in Coder - -Go to **Manage** > **Admin**. On the **Infrastructure** tab, find the **GPU -Vendor** setting, and change it to the GPU vendor of choice (either **AMD** or -**Nvidia**). Click **Save Vendor**. +To do so, go to **Manage** > **Admin**. On the **Infrastructure** tab, find the +**GPU Vendor** setting, and change it to the GPU vendor of choice (either +**AMD** or **Nvidia**). Click **Save Vendor**. ![Enable GPU vendor](../../assets/admin/gpu.png) From 2cd0519da02cf3798b3696678265cb73518a3e24 Mon Sep 17 00:00:00 2001 From: Katie Horne Date: Wed, 1 Dec 2021 13:37:08 -0600 Subject: [PATCH 2/3] edit text --- admin/workspace-management/gpu-acceleration.md | 1 + 1 file changed, 1 insertion(+) diff --git a/admin/workspace-management/gpu-acceleration.md b/admin/workspace-management/gpu-acceleration.md index 47329746f..9f77a3e25 100644 --- a/admin/workspace-management/gpu-acceleration.md +++ b/admin/workspace-management/gpu-acceleration.md @@ -10,6 +10,7 @@ You can allocate GPUs to workspaces once a site manager configures and enables this feature. Enabling GPU acceleration requires that you've configured your Kubernetes cluster appropriately and [scheduled GPUs](https://kubernetes.io/docs/tasks/manage-gpus/scheduling-gpus/). +For instructions on how to schedule GPUs with your specific cloud vendor, see: - [Amazon Elastic Container Service](https://docs.aws.amazon.com/AmazonECS/latest/developerguide/ecs-gpu.html) - [Azure Kubernetes Service](https://docs.microsoft.com/en-us/azure/aks/gpu-cluster) From abc6637f172df1a3e3f167f55b3ceb460e841c34 Mon Sep 17 00:00:00 2001 From: Katie Horne Date: Wed, 1 Dec 2021 13:39:31 -0600 Subject: [PATCH 3/3] clarify sentence --- admin/workspace-management/gpu-acceleration.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/admin/workspace-management/gpu-acceleration.md b/admin/workspace-management/gpu-acceleration.md index 9f77a3e25..0357f6a66 100644 --- a/admin/workspace-management/gpu-acceleration.md +++ b/admin/workspace-management/gpu-acceleration.md @@ -8,7 +8,7 @@ such as those involved with data science/machine learning projects. You can allocate GPUs to workspaces once a site manager configures and enables this feature. Enabling GPU acceleration requires that you've configured your -Kubernetes cluster appropriately and +Kubernetes cluster and [scheduled GPUs](https://kubernetes.io/docs/tasks/manage-gpus/scheduling-gpus/). For instructions on how to schedule GPUs with your specific cloud vendor, see: