glm53 vulkan: clamp SiLU*Up in the shared gate_up shader - #1566
Merged
JustVugg merged 1 commit intoSep 18, 2026
Merged
Conversation
CPU and Metal already cap that product. The Vulkan path used the same qmatmul_gate_up.comp without a limit, so GLM-5.3 decode could NaN on fp16 overflow. limit<=0 stays unclamped for GLM-5.2, per dispatch. Fixes JustVugg#1520
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Fixes #1520
CPU and Metal already cap
silu(gate)*upwith the checkpointswiglu_limit. The Vulkan fused kernel (qmatmul_gate_up.comp) did the unclamped product, so GLM-5.3 output depended on which experts were tier-resident.The limit rides in the gate_up push constants per dispatch (
PCGU).limit <= 0stays unclamped, which is the correct value for GLM-5.2 /colibri.con the same shader.glm53publishesm->c.swiglu_limitaftercoli_vk_init. Source contract inc/tests/test_glm53_vulkan_clamped_source.py.