Merge pull request #293415 from xinlaoda/update-vm-extension

add install driver by VM extensions description.

Author: prmerger-automator[bot]

articles/batch/batch-pool-compute-intensive-sizes.md

@@ -3,7 +3,7 @@ title: Use compute-intensive Azure VMs with Batch
 description: How to take advantage of HPC and GPU virtual machine sizes in Azure Batch pools. Learn about OS dependencies and see several scenario examples.
 ms.topic: how-to
 ms.custom: linux-related-content
-ms.date: 06/07/2024
+ms.date: 02/04/2025
 ---
 # Use RDMA or GPU instances in Batch pools
 
@@ -33,7 +33,7 @@ The RDMA or GPU capabilities of compute-intensive sizes in Batch are supported o
 | -------- | -------- | ----- |  -------- | ----- |
 | [H16r, H16mr](/azure/virtual-machines/sizes-hpc)<br/>[NC24r, NC24rs_v2, NC24rs_v3, ND24rs<sup>*</sup>](/azure/virtual-machines/linux/n-series-driver-setup#rdma-network-connectivity) | RDMA | Ubuntu 22.04 LTS <br/> (Azure Marketplace) | Intel MPI 5<br/><br/>Linux RDMA drivers | Enable inter-node communication, disable concurrent task execution |
 | [NCv3, NDv2, NDv4, NDv5 series](/azure/virtual-machines/linux/n-series-driver-setup) | NVIDIA Tesla GPU (varies by series) | Ubuntu 22.04 LTS  <br/> (Azure Marketplace) | NVIDIA CUDA or CUDA Toolkit drivers | N/A |
-| [NVv3, NVv4, NVv5 series](/azure/virtual-machines/linux/n-series-driver-setup) | Accelerated Visualization GPU | Ubuntu 22.04 LTS <br/> (Azure Marketplace) | NVIDIA GRID drivers (if required) | N/A |
+| [NVv3, NVv4, NVv5 series](/azure/virtual-machines/linux/n-series-driver-setup) | Accelerated Visualization GPU | Ubuntu 22.04 LTS <br/> (Azure Marketplace) | NVIDIA GRID drivers or AMD GPU drivers | N/A |
 
 <sup>*</sup>RDMA-capable N-series sizes also include NVIDIA Tesla GPUs
 
@@ -69,14 +69,16 @@ To configure a specialized VM size for your Batch pool, you have several options
 
 * For pools in the virtual machine configuration, choose a preconfigured [Azure Marketplace](https://azuremarketplace.microsoft.com/marketplace/) VM image that has drivers and software preinstalled. Examples:
 
-* [Data Science Virtual Machine](/azure/machine-learning/data-science-virtual-machine/overview) for Linux or Windows - includes NVIDIA CUDA drivers
+    * [Data Science Virtual Machine](/azure/machine-learning/data-science-virtual-machine/overview) for Linux or Windows - includes NVIDIA CUDA drivers
 
-* Linux images for Batch container workloads that also include GPU and RDMA drivers:
+    * Linux images for Batch container workloads that also include GPU and RDMA drivers:
 
-* [Ubuntu Server (with GPU and RDMA drivers) for Azure Batch container pools](https://azuremarketplace.microsoft.com/marketplace/apps/microsoft-azure-batch.ubuntu-server-container-rdma?tab=Overview)
+    * [Ubuntu Server (with GPU and RDMA drivers) for Azure Batch container pools](https://azuremarketplace.microsoft.com/marketplace/apps/microsoft-azure-batch.ubuntu-server-container-rdma?tab=Overview)
 
 * Create a [custom Windows or Linux VM image](batch-sig-images.md) with installed drivers, software, or other settings required for the VM size.
 
+* [Install GPU and RDMA drivers by VM extension](create-pool-extensions.md).
+
 * Create a Batch [application package](batch-application-packages.md) from a zipped driver or application installer. Then, configure Batch to deploy this package to pool nodes and install once when each node is created. For example, if the application package is an installer, create a [start task](jobs-and-tasks.md#start-task) command line to silently install the app on all pool nodes. Consider using an application package and a pool start task if your workload depends on a particular driver version.
 
   > [!NOTE]