happyz
/
llama.cpp
mirror of https://github.com/ggerganov/llama.cpp.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

kleidiai : fix MUL_MAT support for batched (3D) inputs (#20620) 

* kleidiai : fix MUL_MAT support for batched (3D) inputs

The supports_op() check incorrectly rejected MUL_MAT operations with 3D
inputs (ne[2] > 1), but the actual compute_forward_qx() implementation
handles batched inputs correctly via a loop over ne12.

This caused models with Q4_0/Q8_0 weights to crash during graph scheduling
when n_seq_max > 1, because weights were placed in KLEIDIAI buffers during
loading (tested with 2D inputs) but the runtime used 3D inputs.

Also relax the buffer check to allow supports_op() to be called during
weight loading when src[0]->buffer is NULL.

Fixes #20608

* Kleidiai support_ops should only return true for 3D inputs, not also 4D
This commit is contained in:
Justin Bradford 2026-03-17 05:03:54 -07:00 committed by GitHub
parent 740a447fc3
commit 627670601a
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
ggml/src/ggml-cpu/kleidiai/kleidiai.cpp
View File

@ -1461,7 +1461,7 @@ class extra_buffer_type : ggml::cpu::extra_buffer_type {
return false;
}
if ((op->src[1]->type == GGML_TYPE_F32 || op->src[1]->type == GGML_TYPE_I32) &&
ggml_ne(op->src[1], 2) == 1 && ggml_ne(op->src[1], 3) == 1) {
ggml_ne(op->src[1], 3) == 1) {
return true;
}
}