Add RTX PRO 6000 Blackwell Server Edition to GPU_VRAM_GB - #2
Open
aizorr0 wants to merge 1 commit into
Open
Conversation
Registers the model (96 GB) plus aliases so validators resolve its proof GPU hash to the canonical name instead of tagging "Unknown GPU". Enables the RTX PRO 6000 timing row (testnet-468 calibration, counts 1/2/4). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What
Adds
NVIDIA RTX PRO 6000 Blackwell Server Edition(96 GB) toGPU_VRAM_GBincommon/config.py, plus two aliases (RTX PRO 6000,RTX PRO 6000 Blackwell).Why
Without the entry, validators resolve the model's proof GPU hash to
Unknown GPU, so its calibrated timing row can't take effect. This addition lets the hash resolve to the canonical name.Context
Part of the testnet-468 timing calibration for netuid 106. The RTX PRO 6000 timing row (counts 1/2/4, n=146) is ready to publish via the subnet-config admin PUT, but only resolves once this lands and validators redeploy. The other calibrated rows (H100, H200, RTX 5090, B200) need no code change.
Testing
py_compileclean; the miner-reported hash for the canonical name resolves viagpu_model_hash_lookup().