Skip to content

Add RTX Pro 6000 #61388

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
nickhil-g wants to merge 1 commit into
base: master
Choose a base branch
from
Open

Add RTX Pro 6000 #61388

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions python/ray/util/accelerators/accelerators.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -12,6 +12,7 @@
NVIDIA_H20 = "H20"
NVIDIA_B200 = "B200"
NVIDIA_B300 = "B300"
NVIDIA_RTX_PRO_6000 = "RTX"
Copy link
Contributor

@gemini-code-assist gemini-code-assist bot Feb 27, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

medium

The value assigned to NVIDIA_RTX_PRO_6000 is currently "RTX". This value is quite generic and could lead to ambiguity if other RTX series GPUs are added in the future (e.g., RTX 3090, RTX 4090). For consistency with other specific NVIDIA GPU constants like NVIDIA_A100_40G = "A100-40G", it would be more precise to use a value that reflects the full model name, such as "RTX_PRO_6000" or "RTX6000".

Suggested change
NVIDIA_RTX_PRO_6000 = "RTX"
NVIDIA_RTX_PRO_6000 = "RTX_PRO_6000"

Copy link

@cursor cursor bot Feb 27, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Accelerator type value "RTX" is ambiguous across GPUs

Medium Severity

The value "RTX" assigned to NVIDIA_RTX_PRO_6000 is too generic to serve as a unique identifier for this specific GPU. The auto-detection regex in nvidia_gpu.py (r"\w+\s+([A-Z0-9]+)") extracts only the second word from the NVML device name, so any NVIDIA RTX professional card (e.g., RTX A6000, RTX 6000 Ada) would also produce "RTX". A user requesting NVIDIA_RTX_PRO_6000 could be matched to a completely different RTX GPU. Other constants in this file (like "A100", "H100", "L4") are unique identifiers, but "RTX" is a brand prefix shared across many products.

Fix in Cursor Fix in Web

Copy link
Contributor

@MengjinYan MengjinYan Feb 27, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think that's a valid feedback.

If other RTX type GPU also generate the same string, we will update function to get the device type to make the string unique.

INTEL_MAX_1550 = "Intel-GPU-Max-1550"
INTEL_MAX_1100 = "Intel-GPU-Max-1100"
INTEL_GAUDI = "Intel-GAUDI"
Expand Down