Conversation
Gemma 4 E-series routes a per-layer embedding input into every block; under partial offload (e.g. 3.5GB RENOIR Vulkan + CPU) each one lands in the split input list and GGML_ASSERT(n_inputs < GGML_SCHED_MAX_SPLIT_INPUTS) aborts during memory fitting — the in-app 'model process crashed' report. Reproduced with gemma-4-E4B-it-Q4_K_S on RADV RENOIR; -ngl 0 avoids it, any GPU offload trips it. 30 -> 128 with headroom for 31B-deep stacks.
fix: sched split-input overflow on Gemma 4 E-series partial offload
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Promotes #46 for the b10018-1.1.2 patch release.
Fixes the 'model process crashed' abort in Atomic Chat when loading Gemma 4 E-series models with partial offload on small-VRAM GPUs: GGML_SCHED_MAX_SPLIT_INPUTS raised 30 → 128 (per-layer embedding inputs of the E-series overflow the split input list when the graph splits between backends). Verified on the reporting hardware (RADV RENOIR 3.5GB) with the PR CI artifact: crash on 1.1.1, healthy load and generation with the fix.