chore: update NVIDIA GRID driver to v18.6 (570.211.01)#154
Merged
Conversation
Bump the GRID driver from 550.144.06 (vGPU 17.55) to 570.211.01 (vGPU 18.6). Version + URL sourced from the Azure HPC NVIDIA driver manifest: https://github.com/Azure/azhpc-extensions/blob/master/NvidiaGPU/resources.json Note: auto_update.py currently reads Nvidia-GPU-Linux-Resources.json which is stuck at 17.55. A follow-up PR will update the auto-updater to pull v18.x from the resources.json file (filtered by vGPUVersion major == "18"). Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
djsly
approved these changes
May 19, 2026
ganeshkumarashok
added a commit
that referenced
this pull request
May 27, 2026
Builds a second GRID image, `aks-gpu-grid-v20`, alongside the existing `aks-gpu-grid` (v18). The new image carries NVIDIA GRID v20 driver `595.58.03`, which is required for the RTX PRO 6000 Blackwell Server Edition v6 SKU family. Existing GRID v18 image and pin are unchanged. - `driver_config.yml`: add `grid_v20` block. - CI/release workflows: matrix-build both image repos from their respective `driver_config.yml` entries; `DRIVER_KIND=grid` for both; cache keys and image tags scope by image repo. - `justfile`: add `buildgridv20`/`pushgridv20`; include v20 in `pushallgrid`. Note: `auto_update.py` is left unchanged; v20 can be bumped via the same manual flow currently used for v18 (PR #154). Extending the auto-updater to handle both majors can be a separate, focused PR. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
ganeshkumarashok
added a commit
that referenced
this pull request
May 28, 2026
Adds two new NVIDIA driver container images alongside the existing ones, following the same matrix-build pattern as the existing aks-gpu-cuda / aks-gpu-cuda-arm64 split: - `aks-gpu-grid-v20` (GRID v20, driver `595.58.03`): required for the RTX PRO 6000 Blackwell Server Edition v6 SKU family. Existing `aks-gpu-grid` (v18, 570.211.01) is unchanged. - `aks-gpu-cuda-lts` and `aks-gpu-cuda-lts-arm64` (CUDA LTS, NVIDIA R580 LTSB `580.159.04`): a Long Term Support Branch variant alongside the existing Production Branch `aks-gpu-cuda` (595.71.05). R580 LTSB is supported by NVIDIA through Aug 2028. Files: - `driver_config.yml`: add `grid_v20` and `cuda_lts` blocks. - `main.yaml` / `ci.yaml`: matrix-include both branches in each of the grid, cuda, and cuda-arm64 jobs. `DRIVER_KIND=grid` / `cuda` is literal; cache keys and image tags scope by image repo. - `justfile`: add `buildgridv20` / `pushgridv20` and `buildcudalts` / `pushcudalts`. Out of scope: - `auto_update.py` — new variants pinned manually for now, same flow as the existing v18 pin (PR #154). A focused follow-up can extend the updater to handle each branch within its own major. - AgentBaker consumption of the new images — separate PR. - Mariner/Azure Linux/ACL GRID install paths — use separate package/sysext flows, unchanged. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
ganeshkumarashok
added a commit
that referenced
this pull request
May 29, 2026
Adds two new NVIDIA driver container images alongside the existing ones, following the same matrix-build pattern as the existing aks-gpu-cuda / aks-gpu-cuda-arm64 split: - `aks-gpu-grid-v20` (GRID v20, driver `595.58.03`): required for the RTX PRO 6000 Blackwell Server Edition v6 SKU family. Existing `aks-gpu-grid` (v18, 570.211.01) is unchanged. - `aks-gpu-cuda-lts` and `aks-gpu-cuda-lts-arm64` (CUDA LTS, NVIDIA R580 LTSB `580.159.04`): a Long Term Support Branch variant alongside the existing Production Branch `aks-gpu-cuda` (595.71.05). R580 LTSB is supported by NVIDIA through Aug 2028. Files: - `driver_config.yml`: add `grid_v20` and `cuda_lts` blocks. - `main.yaml` / `ci.yaml`: matrix-include both branches in each of the grid, cuda, and cuda-arm64 jobs. `DRIVER_KIND=grid` / `cuda` is literal; cache keys and image tags scope by image repo. - `justfile`: add `buildgridv20` / `pushgridv20` and `buildcudalts` / `pushcudalts`. Out of scope: - `auto_update.py` — new variants pinned manually for now, same flow as the existing v18 pin (PR #154). A focused follow-up can extend the updater to handle each branch within its own major. - AgentBaker consumption of the new images — separate PR. - Mariner/Azure Linux/ACL GRID install paths — use separate package/sysext flows, unchanged. Co-authored-by: ganeshkumarashok <ganeshkumarashok@users.noreply.github.com> Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Bumps the GRID driver from 550.144.06 (vGPU 17.55) to 570.211.01 (vGPU 18.6).
Source
Version + URL pulled from the Azure HPC NVIDIA driver manifest:
Azure/azhpc-extensions
NvidiaGPU/resources.json— the v18.6 Linux GRID entry:{ "Num": "570.211.01", "DirLink": "https://download.microsoft.com/download/2a04ca6a-9eec-40d9-9564-9cdea1ab795f/NVIDIA-Linux-x86_64-570.211.01-grid-azure.run", "vGPUVersion": "18.6", "SupportedOSVersions": "Ubuntu 20.04 LTS, 22.04 LTS, 24.04 LTS, Red Hat Enterprise Linux 8.6, 8.8, 8.9" }URL verified
HTTP 200.Auto-updater note
auto_update.pyreadsNvidia-GPU-Linux-Resources.jsonwhich the HPC team stopped updating at v17.55 — so the daily auto-PR job won't surface v18.x bumps without changes. Follow-up PR will updateauto_update.pyto read fromresources.jsonand filter byvGPUVersionmajor ==18.