Combo PR for GPU runners, 2 vCPU linux runners, and hardware acceleration (#49959)

Co-authored-by: Larissa Fortuna <[email protected]>
Co-authored-by: Laura Coursen <[email protected]>

Author: 3 people

content/actions/learn-github-actions/usage-limits-billing-and-administration.md

@@ -60,13 +60,17 @@ There are some limits on {% data variables.product.prodname_actions %} usage whe
 
   **{% data variables.product.prodname_dotcom %}-hosted {% data variables.actions.hosted_runner %}s**
 
-  | GitHub plan | Total concurrent jobs | Maximum concurrent macOS jobs |
-  |---|---|---|
-  | All | 1000 | The limit is based on your {% data variables.product.prodname_dotcom %} plan. |
+  | GitHub plan | Total concurrent jobs | Maximum concurrent macOS jobs | Maximum concurrent GPU jobs |
+  |---|---|---|---|
+  | Team | 1000 | 5 | 100 |
+  | Enterprise | 1000 | 50 | 100 |
 
   {% note %}
 
-  **Note:** If required, customers on enterprise plans can request a higher limit for concurrent jobs. For more information, contact us through the {% data variables.contact.contact_support_portal %}, or contact your sales representative.
+  **Notes:**
+
+  - If required, customers on enterprise plans can request a higher limit for concurrent jobs. For more information, contact us through the {% data variables.contact.contact_support_portal %}, or contact your sales representative.
+  - The maximum concurrent macOS jobs is shared across standard {% data variables.product.prodname_dotcom %}-hosted runner and {% data variables.product.prodname_dotcom %}-hosted {% data variables.actions.hosted_runner %}s.
 
   {% endnote %}
 

content/actions/using-github-hosted-runners/about-github-hosted-runners/about-github-hosted-runners.md

@@ -120,6 +120,8 @@ While the job runs, the logs and output can be viewed in the {% data variables.p
 
 {% data variables.product.prodname_dotcom %}-hosted runners are available for use in both public and private repositories.
 
+{% data variables.product.prodname_dotcom %}-hosted Linux runners support hardware acceleration for Android SDK tools, which makes running Android tests much faster and consumes fewer minutes. For more information on Android hardware acceleration, see [Configure hardware acceleration for the Android Emulator](https://developer.android.com/studio/run/emulator-acceleration) in the Android Developers documentation.
+
 {% note %}
 
 **Note:** The `-latest` runner images are the latest stable images that {% data variables.product.prodname_dotcom %} provides, and might not be the most recent version of the operating system available from the operating system vendor.
@@ -178,9 +180,7 @@ You can install additional software on {% data variables.product.prodname_dotcom
 
 {% data variables.product.prodname_dotcom %} hosts Linux and Windows runners on virtual machines in Microsoft Azure with the {% data variables.product.prodname_actions %} runner application installed. The {% data variables.product.prodname_dotcom %}-hosted runner application is a fork of the Azure Pipelines Agent. Inbound ICMP packets are blocked for all Azure virtual machines, so ping or traceroute commands might not work. {% data variables.product.prodname_dotcom %} hosts macOS runners in Azure data centers.
 
-{% data variables.product.prodname_dotcom %} provides different Linux and Windows runners for public and private repositories.
-- For Linux and Windows runners in **public** repositories, GitHub uses `Standard_D4ads_v5` virtual machines. For more information, see [Dasv5 and Dadsv5-series](https://learn.microsoft.com/en-us/azure/virtual-machines/dasv5-dadsv5-series#dadsv5-series) in the Microsoft Azure documentation.
-- For Linux and Windows runners in **private** repositories, GitHub uses `Standard_DS2_v2` virtual machines. For more information, see [Dv2 and DSv2-series](https://learn.microsoft.com/en-us/azure/virtual-machines/dv2-dsv2-series#dsv2-series) in the Microsoft Azure documentation.
+For Linux and Windows runners, GitHub uses `Dadsv5-series` virtual machines. For more information, see [Dasv5 and Dadsv5-series](https://learn.microsoft.com/en-us/azure/virtual-machines/dasv5-dadsv5-series#dadsv5-series) in the Microsoft Azure documentation.
 
 ## Workflow continuity
 

content/actions/using-github-hosted-runners/about-larger-runners/about-larger-runners.md

@@ -1,7 +1,7 @@
 ---
 title: About larger runners
 shortTitle: About larger runners
-intro: '{% data variables.product.prodname_dotcom %} offers runners with more RAM, CPU, and disk space.'
+intro: '{% data variables.product.prodname_dotcom %} offers runners with advanced features to support more customized use cases.'
 permissions: '{% data reusables.actions.larger-runner-permissions %}'
 versions:
   feature: actions-hosted-runners
@@ -13,7 +13,7 @@ redirect_from:
 
 {% data reusables.actions.about-larger-runners %}
 
-{% data variables.product.prodname_dotcom %} offers {% data variables.actions.hosted_runners %} with macOS, Ubuntu, or Windows operating systems, and different features are available depending on which operating system you use. For more information, see "[Additional features for {% data variables.actions.hosted_runners %}](#additional-features-for-larger-runners)."
+{% data variables.product.prodname_dotcom %} offers {% data variables.actions.hosted_runners %} with macOS, Ubuntu, or Windows operating systems, and different features and sizes are available depending on which operating system you use. For more information, see "[Additional features for {% data variables.actions.hosted_runners %}](#additional-features-for-larger-runners)."
 
 ### About Ubuntu and Windows {% data variables.actions.hosted_runners %}
 
@@ -45,7 +45,6 @@ Compared to standard {% data variables.product.prodname_dotcom %}-hosted runners
 
 | Operating system                             | Ubuntu | Windows | macOS |
 | -------------------------------------------- | ------ | ------- | ----- |
-| Hardware acceleration for Android SDK tools  | {% octicon "check" aria-label="Supported" %} | {% octicon "x" aria-label="Not supported" %} | {% octicon "x" aria-label="Not supported" %} |
 | Static IP addresses | {% octicon "check" aria-label="Supported" %} | {% octicon "check" aria-label="Supported" %} | {% octicon "x" aria-label="Not supported" %} |
 | Autoscaling | {% octicon "check" aria-label="Supported" %} | {% octicon "check" aria-label="Supported" %} | {% octicon "x" aria-label="Not supported" %} |
 | Runner groups | {% octicon "check" aria-label="Supported" %} | {% octicon "check" aria-label="Supported" %} | {% octicon "x" aria-label="Not supported" %} |
@@ -54,12 +53,11 @@ Compared to standard {% data variables.product.prodname_dotcom %}-hosted runners
 
 These features can enhance your CI/CD pipelines in the following ways.
 
-- Hardware acceleration for the Android SDK tools makes running Android tests much faster and consumes fewer minutes. For more information on Android hardware acceleration, see [Configure hardware acceleration for the Android Emulator](https://developer.android.com/studio/run/emulator-acceleration) in the Android Developers documentation.
 - Assigning {% data variables.actions.hosted_runners %}  static IP addresses from a specific range enables you to use this range to configure a firewall allowlist. For more information, see "[Networking for {% data variables.actions.hosted_runners %}](#networking-for-larger-runners)."
 - Autoscaling enables {% data variables.actions.hosted_runners %}  to scale up to a maximum limit set by you, so your workflows can run concurrently. For more information, see "[Autoscaling {% data variables.actions.hosted_runners %}](#autoscaling-larger-runners)."
 - Runner groups allow you to control access to {% data variables.actions.hosted_runners %} for your organizations, repositories, and workflows. For more information, see "[AUTOTITLE](/actions/using-github-hosted-runners/controlling-access-to-larger-runners)."
 
-For a full list of included tools for each runner operating system, see the [{% data variables.product.prodname_actions %} Runner Images](https://github.com/actions/runner-images) repository.
+For a full list of included tools for each runner operating system (excluding GPU runners), see the [{% data variables.product.prodname_actions %} Runner Images](https://github.com/actions/runner-images) repository. For GPU runners, see [NVIDIA GPU-Optimized VMI](https://azuremarketplace.microsoft.com/en-us/marketplace/apps/nvidia.ngc_azure_17_11) and [Data Science Virtual Machine - Windows 2019](https://azuremarketplace.microsoft.com/en-us/marketplace/apps/microsoft-dsvm.dsvm-win-2019?tab=overview).
 
 ### Understanding billing
 
@@ -73,15 +71,28 @@ Compared to standard {% data variables.product.prodname_dotcom %}-hosted runners
 
 ## Machine sizes for {% data variables.actions.hosted_runners %}
 
-| Processor (CPU)| Memory (RAM)  | Storage (SSD) | Operating system (OS) |
-| -------------- | ------------- | ------------- | --------------------- |
-| 6              | 14 GB         | 14 GB         | macOS                 |
-| 12             | 30 GB         | 14 GB         | macOS                 |
-| 4              | 16 GB         | 150 GB        | Ubuntu                |
-| 8              | 32 GB         | 300 GB        | Ubuntu, Windows       |
-| 16             | 64 GB         | 600 GB        | Ubuntu, Windows       |
-| 32             | 128 GB        | 1200 GB       | Ubuntu, Windows       |
-| 64             | 256 GB        | 2040 GB       | Ubuntu, Windows       |
+You can choose from several specifications for {% data variables.actions.hosted_runners %}.
+
+### Specifications for general {% data variables.actions.hosted_runners %}
+
+| CPU | Memory (RAM)  | Storage (SSD) | Operating system (OS) |
+| --- | ------------- | ------------- | --------------------- |
+| 6   | 14 GB         | 14 GB         | macOS                 |
+| 12  | 30 GB         | 14 GB         | macOS                 |
+| 2   | 8 GB          | 14 GB         | Ubuntu                |
+| 4   | 16 GB         | 150 GB        | Ubuntu, Windows                |
+| 8   | 32 GB         | 300 GB        | Ubuntu, Windows       |
+| 16  | 64 GB         | 600 GB        | Ubuntu, Windows       |
+| 32  | 128 GB        | 1200 GB       | Ubuntu, Windows       |
+| 64  | 256 GB        | 2040 GB       | Ubuntu, Windows       |
+
+### Specifications for GPU {% data variables.actions.hosted_runners %}
+
+{% data reusables.actions.gpu-runner-beta-note %}
+
+| CPU | GPU | GPU card | Memory (RAM) | GPU memory (VRAM) | Storage (SSD) | Operating system (OS) |
+| --- | --- | -------- | ------------ | ----------------- | ------------- | --------------------- |
+| 4   | 1   | Tesla T4 | 28 GB        | 16 GB             | 176 GB        | Ubuntu, Windows       |
 
 ## About runner groups
 
@@ -123,7 +134,7 @@ In the following diagram, a class of hosted runner named `ubuntu-20.04-16core` h
 
 {% data variables.actions.hosted_runner_caps %}s can automatically scale to suit your needs. You can provision machines to run a specified maximum number of jobs when jobs are submitted for processing. Each machine only handles one job at a time, so these settings effectively determine the number of jobs that can be run concurrently.
 
-You can configure the maximum job concurrency, which allows you to control your costs by setting the maximum parallel number of jobs that can be run using this set. A higher value here can help avoid workflows being blocked due to parallelism. For more information, see "[AUTOTITLE](/actions/using-github-hosted-runners/managing-larger-runners#configuring-autoscaling-for-larger-runners)."
+You can configure the maximum job concurrency, which allows you to control your costs by setting the maximum parallel number of jobs that can be run using this set. A higher value here can help avoid workflows being blocked due to parallelism. For more information on how to set limits, see "[AUTOTITLE](/actions/using-github-hosted-runners/managing-larger-runners#configuring-autoscaling-for-larger-runners)". For more information on the maximum auto-scaling limits for {% data variables.product.company_short %}-hosted runners, see "[AUTOTITLE](/actions/learn-github-actions/usage-limits-billing-and-administration#usage-limits)."
 
 ## Networking for {% data variables.actions.hosted_runners %}
 

content/actions/using-github-hosted-runners/about-larger-runners/managing-larger-runners.md

@@ -11,6 +11,8 @@ redirect_from:
 
 {% data reusables.actions.windows-linux-larger-runners-note %}
 
+{% data reusables.actions.gpu-runner-beta-note %}
+
 {% ifversion ghec %}
 
 ## Adding a {% data variables.actions.hosted_runner %} to an enterprise

content/billing/managing-billing-for-github-actions/about-billing-for-github-actions.md

@@ -70,7 +70,33 @@ If your account's usage surpasses these limits and you have set a spending limit
 
 ### Per-minute rates
 
-{% data reusables.billing.billing-hosted-runners %}
+#### Per-minute rates for standard runners
+
+| Operating system            | Per-minute rate (USD) |
+|-----------------------------| ----------------------|
+|  Linux 2-core               |   $0.008              |
+|  Windows 2-core             |   $0.016              |
+|  macOS 3 or 4 (M1 or Intel) |   $0.08               |
+
+#### Per-minute rates for {% data variables.actions.hosted_runners %}
+
+| Operating system    | Per-minute rate (USD) |
+|---------------------| -----------|
+|  Linux 2-core       |   $0.008   |
+|  Linux 4-core       |   $0.016   |
+|  Linux 8-core       |   $0.032   |
+|  Linux 16-core      |   $0.064   |
+|  Linux 32-core      |   $0.128   |
+|  Linux 64-core      |   $0.256   |
+|  Linux 4-core GPU   |   $0.07    |
+|  Windows 4-core     |   $0.032   |
+|  Windows 8-core     |   $0.064   |
+|  Windows 16-core    |   $0.128   |
+|  Windows 32-core    |   $0.256   |
+|  Windows 64-core    |   $0.512   |
+|  Windows 4-core GPU |   $0.014   |
+|  macOS 12-core      |   $0.12    |
+|  macOS 6-core (M1)  |   $0.16    |
 
 - The number of jobs you can run concurrently across all repositories in your user or organization account depends on your GitHub plan. For more information, see "[AUTOTITLE](/actions/learn-github-actions/usage-limits-billing-and-administration)" for {% data variables.product.prodname_dotcom %}-hosted runners and "[AUTOTITLE](/actions/hosting-your-own-runners/managing-self-hosted-runners/about-self-hosted-runners#usage-limits)" for self-hosted runner usage limits.
 - {% data reusables.user-settings.context_switcher %}

data/reusables/actions/about-larger-runners.md

@@ -1 +1,8 @@
-In addition to the [standard {% data variables.product.prodname_dotcom %}-hosted runners](/actions/using-github-hosted-runners/about-github-hosted-runners#supported-runners-and-hardware-resources), {% data variables.product.prodname_dotcom %} offers customers on {% data variables.product.prodname_team %} and {% data variables.product.prodname_ghe_cloud %} plans a range of managed virtual machines with more RAM, CPU, and disk space. These {% data variables.actions.hosted_runners %} are hosted by {% data variables.product.prodname_dotcom %} and have the runner application and other tools preinstalled.
+In addition to the [standard {% data variables.product.prodname_dotcom %}-hosted runners](/actions/using-github-hosted-runners/about-github-hosted-runners#supported-runners-and-hardware-resources), {% data variables.product.prodname_dotcom %} offers customers on {% data variables.product.prodname_team %} and {% data variables.product.prodname_ghe_cloud %} plans a range of managed virtual machines with advanced features such as:
+
+- More RAM, CPU, and disk space
+- Static IP addresses
+- The ability to group runners
+- Autoscaling to support concurrent workflows
+
+These {% data variables.actions.hosted_runners %} are hosted by {% data variables.product.prodname_dotcom %} and have the runner application and other tools preinstalled.

data/reusables/actions/add-hosted-runner.md

@@ -2,9 +2,12 @@
 1. Complete the required details to configure your new runner:
 
     - **Name**: Enter a name for your new runner. For easier identification, this should indicate its hardware and operating configuration, such as `ubuntu-20.04-16core`.
-    - **Runner image**: Choose an operating system from the available options. Once you've selected an operating system, you will be able to choose a specific version.
-    - **Runner size**: Choose a hardware configuration from the drop-down list of available options.
-    - **Auto-scaling**: Choose the maximum number of jobs that can be active at any time.
+    - **Platform**: Choose a platform from the available options. Once you've selected a platform, you will be able to choose a specific image.
+    - **Image**: Choose an image from the available options. Once you've selected an image, you will be able to choose a specific size.
+      - **GitHub-owned**: For images managed by GitHub, select an image under this tab.
+      - **Partner**: For images managed by a partner, select an image under this tab. GPU-optimized images are located under this tab.
+    - **Size**: Choose a hardware configuration from the list of available options. For GPU runners, select a size under the **GPU-powered** tab.
+    - **Maximum concurrency**: Choose the maximum number of jobs that can be active at any time.
     - **Runner group**: Choose the group that your runner will be a member of. This group will host multiple instances of your runner, as they scale up and down to suit demand. {% ifversion ghec %}
     - **Networking**: Choose whether static IP address ranges will be assigned to instances of the {% data variables.actions.hosted_runner %}. You can use up to 10 {% data variables.actions.hosted_runner %}s with static IP addresses in total. {% endif %}
 

data/reusables/actions/gpu-runner-beta-note.md

@@ -0,0 +1,5 @@
+{% note %}
+
+**Note:** GPU runners are currently in beta and subject to change.
+
+{% endnote %}

data/reusables/actions/supported-github-runners.md

@@ -105,7 +105,6 @@ macOS
 <td>
 7 GB
 </td>
-N/A
 <td>
 14 GB
 </td>
@@ -223,7 +222,6 @@ macOS
 <td>
 7 GB
 </td>
-N/A
 <td>
 14 GB
 </td>