Repin GLM-5-Next onto the head carrying the indexer softmax fix - #217
Merged
Merged
Conversation
#214 landed on glm5next/upstream after #216 was cut, so e2738e0 is no longer the head. 86ebfef is that squash on top of it: the k-pool gate logits are reshaped to 2D before ggml_soft_max so n_new_max stops mapping to gridDim.y, which CUDA caps at 65535 and which aborted the launch at n_kv >= 262144 with kpool = 4. Replayed the resolve loop on b10994 with the new pin: 11 clean, 2 additive, 0 hard fails; merge_checks clean, all 13 pins intact, llama + mtmd compile gate passed, test-llama-archs 316 rows 0 failures.
|
You have reached your Codex usage limits for security reviews. Please try again later. |
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Why
#214 merged into
glm5next/upstreamafter #216 was cut, so the pin that #216 landed (e2738e07) is now one commit behind the PR head. It still merges, but it ships without the indexer softmax fix.86ebfef2is #214 squashed on top ofe2738e07. It reshapes the k-pool gate logits to 2D beforeggml_soft_max, son_new_maxno longer maps togridDim.y:ggml-cuda/softmax.culaunches withblock_nums(ne01, ne02, ne03)and CUDA capsgridDim.yat 65535, so atn_kv >= 262144withkpool = 4the launch was rejected outright.Verified
Reproduced the crash and the fix on a B200 with a standalone ggml program, no weights:
SOFT_MAX failed/CUDA error: invalid argumentSoftmax is row-wise over ne0 and the reshape leaves ne0 alone, so the values are unchanged: max abs diff between the control row and the fixed row is 0.0, both rows sum to 1.0.
Replayed the
resolvemerge loop fromunsloth-prebuilt.ymlonb10994, the tag tonight's run selects:additive_merge.py, 0 hard failsmerge_checks.pycleanpin_contract.py: all 13 pins intact in the merged treellama+mtmd) passedtest-llama-archs: 316 rows, 0 failuresOnly the pin line changes.