Add GPU + CPU TEE requirements article to Fundamentals - #431
Conversation
WalkthroughAdded comprehensive documentation guide for GPU and CPU hardware requirements when operating in Trusted Execution Environment mode. The guide describes supported configurations for Hopper and Blackwell GPU architectures, explains TEE operation modes and their implications, documents driver capabilities, CPU generation constraints, validation practices, and provides reference links. ChangesGPU + CPU TEE Requirements Guide
Estimated code review effort🎯 1 (Trivial) | ⏱️ ~5 minutes Poem
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
There was a problem hiding this comment.
Actionable comments posted: 2
🧹 Nitpick comments (1)
docs/fundamentals/gpu-cpu-tee-requirements.md (1)
16-245: ⚡ Quick winAdd custom anchor tags for stable cross-references.
This long page defines many targetable sections but does not include explicit custom anchor tags. Add anchors to key headings (requirements, modes, unsupported configs, driver releases) to match repo cross-referencing conventions and avoid link drift when titles change.
As per coding guidelines, "Follow the established cross-referencing pattern with custom anchor tags".
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@docs/fundamentals/gpu-cpu-tee-requirements.md` around lines 16 - 245, The document lacks stable custom anchor tags for key headings, so add explicit HTML anchors following the repo's cross-referencing pattern to the major sections: "Core TEE Requirements", "TEE Mode Compatibility", "Supported GPU + CPU Configurations", "NVIDIA Confidential Computing Driver Releases", and "Unsupported Configurations" (and sub-sections like "Hopper", "Blackwell", "RTX PRO 6000"); insert stable id attributes or explicit <a id="..."></a> anchors immediately before those headings so other docs can reference them even if titles change, using consistent slug-style names (e.g., core-tee-requirements, tee-mode-compatibility, supported-gpu-cpu-configs, driver-releases, unsupported-configs, hopper, blackwell, rtx-pro-6000) to match the project's established cross-referencing convention.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@docs/fundamentals/gpu-cpu-tee-requirements.md`:
- Around line 10-12: Replace the hardcoded header string "As of May 22, 2026"
with a neutral, maintainable indicator such as a frontmatter "last_updated"
field or a non-specific label "Last updated: <date>" that is driven by
frontmatter or the changelog; update the document header in
docs/fundamentals/gpu-cpu-tee-requirements.md where the phrase appears so future
edits update the date centrally rather than embedding a static date in the body.
- Around line 135-136: The two bullets conflict: make the current-state claim
that on "RTX PRO 6000 Blackwell SE" only SPT CC mode is validated (limiting TEE
workloads to 1 GPU per CVM under R595) and remove or reword the assertion that
"1 to 8 GPUs" can operate in a single TEE instance; instead add a clearly
labeled "Planned / Future support" note stating MPT CC (multi-GPU passthrough)
for RTX PRO 6000 SE is not yet validated in R595 and may enable multi-GPU (1–8)
TEEs in later driver releases when NVLink/requirements are met, and mention that
efficiency for intermediate configurations (e.g., 2 or 3 GPUs) is still under
evaluation.
---
Nitpick comments:
In `@docs/fundamentals/gpu-cpu-tee-requirements.md`:
- Around line 16-245: The document lacks stable custom anchor tags for key
headings, so add explicit HTML anchors following the repo's cross-referencing
pattern to the major sections: "Core TEE Requirements", "TEE Mode
Compatibility", "Supported GPU + CPU Configurations", "NVIDIA Confidential
Computing Driver Releases", and "Unsupported Configurations" (and sub-sections
like "Hopper", "Blackwell", "RTX PRO 6000"); insert stable id attributes or
explicit <a id="..."></a> anchors immediately before those headings so other
docs can reference them even if titles change, using consistent slug-style names
(e.g., core-tee-requirements, tee-mode-compatibility, supported-gpu-cpu-configs,
driver-releases, unsupported-configs, hopper, blackwell, rtx-pro-6000) to match
the project's established cross-referencing convention.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: e76508d8-96fa-41b7-881c-d2c13433964b
📒 Files selected for processing (1)
docs/fundamentals/gpu-cpu-tee-requirements.md
| As of May 22, 2026 | ||
|
|
||
| This document is subject to periodic updates as new hardware, drivers, and validation data become available — or as such information is confirmed by or reported to Super Protocol. |
There was a problem hiding this comment.
Avoid hardcoded recency phrasing that goes stale.
Line 10’s fixed “As of May 22, 2026” can age quickly and make the guidance look current when it is not. Prefer a neutral “Last updated” label in frontmatter or a versioned changelog reference.
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@docs/fundamentals/gpu-cpu-tee-requirements.md` around lines 10 - 12, Replace
the hardcoded header string "As of May 22, 2026" with a neutral, maintainable
indicator such as a frontmatter "last_updated" field or a non-specific label
"Last updated: <date>" that is driven by frontmatter or the changelog; update
the document header in docs/fundamentals/gpu-cpu-tee-requirements.md where the
phrase appears so future edits update the date centrally rather than embedding a
static date in the body.
| 1. Currently, only SPT CC mode is validated on the RTX PRO 6000 Blackwell SE. Multi-GPU passthrough (MPT CC) on RTX PRO 6000 SE is not yet validated in R595, limiting TEE workloads to 1 GPU per CVM until future driver releases enable MPT CC for this SKU. | ||
| 2. RTX PRO 6000 systems do not include NVLink, allow 1 to 8 GPUs (1, 2, 3, 4, 5, 6, 7, or 8) to operate in a single TEE instance, provided all components meet Confidential Computing requirements. The efficiency of some configurations (e.g., 2 or 3 GPUs) is still to be evaluated. |
There was a problem hiding this comment.
Conflicting guidance on RTX PRO 6000 multi-GPU TEE support.
Line 135 states current validation is SPT CC only (1 GPU per CVM), while Line 136 states 1–8 GPUs in a single TEE instance. These conflict and can mislead deployment decisions. Please align the section to one current-state claim and clearly mark future/expected support separately.
Proposed clarification
-1. Currently, only SPT CC mode is validated on the RTX PRO 6000 Blackwell SE. Multi-GPU passthrough (MPT CC) on RTX PRO 6000 SE is not yet validated in R595, limiting TEE workloads to 1 GPU per CVM until future driver releases enable MPT CC for this SKU.
-2. RTX PRO 6000 systems do not include NVLink, allow 1 to 8 GPUs (1, 2, 3, 4, 5, 6, 7, or 8) to operate in a single TEE instance, provided all components meet Confidential Computing requirements. The efficiency of some configurations (e.g., 2 or 3 GPUs) is still to be evaluated.
+1. Currently, only SPT CC mode is validated on RTX PRO 6000 Blackwell SE in R595, which means 1 GPU per CVM.
+2. Multi-GPU TEE operation on RTX PRO 6000 (including potential 2–8 GPU topologies) is expected in future releases and should be treated as unvalidated until NVIDIA and Super validation confirms support.📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| 1. Currently, only SPT CC mode is validated on the RTX PRO 6000 Blackwell SE. Multi-GPU passthrough (MPT CC) on RTX PRO 6000 SE is not yet validated in R595, limiting TEE workloads to 1 GPU per CVM until future driver releases enable MPT CC for this SKU. | |
| 2. RTX PRO 6000 systems do not include NVLink, allow 1 to 8 GPUs (1, 2, 3, 4, 5, 6, 7, or 8) to operate in a single TEE instance, provided all components meet Confidential Computing requirements. The efficiency of some configurations (e.g., 2 or 3 GPUs) is still to be evaluated. | |
| 1. Currently, only SPT CC mode is validated on RTX PRO 6000 Blackwell SE in R595, which means 1 GPU per CVM. | |
| 2. Multi-GPU TEE operation on RTX PRO 6000 (including potential 2–8 GPU topologies) is expected in future releases and should be treated as unvalidated until NVIDIA and Super validation confirms support. |
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@docs/fundamentals/gpu-cpu-tee-requirements.md` around lines 135 - 136, The
two bullets conflict: make the current-state claim that on "RTX PRO 6000
Blackwell SE" only SPT CC mode is validated (limiting TEE workloads to 1 GPU per
CVM under R595) and remove or reword the assertion that "1 to 8 GPUs" can
operate in a single TEE instance; instead add a clearly labeled "Planned /
Future support" note stating MPT CC (multi-GPU passthrough) for RTX PRO 6000 SE
is not yet validated in R595 and may enable multi-GPU (1–8) TEEs in later driver
releases when NVLink/requirements are met, and mention that efficiency for
intermediate configurations (e.g., 2 or 3 GPUs) is still under evaluation.
Add GPU + CPU TEE requirements article to Fundamentals
Summary by CodeRabbit