ggml : fix unaligned access in AMX code #16315
Merged
+1
−0
Conversation
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
github-actions
bot
added
the
ggml
ggerganov
commented
Sep 28, 2025
| if (op->op == GGML_OP_MUL_MAT && is_contiguous_2d(op->src[0]) && // src0 must be contiguous | ||
| is_contiguous_2d(op->src[1]) && // src1 must be contiguous | ||
| op->src[0]->buffer && op->src[0]->buffer->buft == ggml_backend_amx_buffer_type() && | ||
| op->src[0]->ne[0] % (TILE_K * 2 * 32) == 0 && // TODO: not sure if correct (https://github.com/ggml-org/llama.cpp/pull/16315) |
There was a problem hiding this comment.
Adding this fixes the CI workflows, though I'm not sure about the reasoning. It effectively requires the row size to be multiple of 2048.
ggerganov
force-pushed
the
gg/fix-amx
branch
from
fed48ec to
96f9f39
Compare
|
I could not reproduce this error with:
Any suggestions? What garbage are you getting out of the AMX backend? |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
The AMX backend produces garbage with
Q4_0andQ4_1quantizations:https://github.com/ggml-org/llama.cpp/actions/runs/18070620247/job/51419855639#step:3:13646
Repro:
The only problem that I was able to find using the address sanitizer is this unaligned access of the quants. However, this does not fix the incorrect results - there is some remaining issue somewhere.
cc @mingfeima