whisper.cpp/ggml
Pascal 60894f1e37 mtmd/ggml: add ggml_build_forward_order (llama/26649)
* ggml: add ggml_build_forward_order

ggml_build_forward_expand marks the tensor and all its ancestors for
compute, so using it as a pure ordering hint (keeping q, k and v
together) defeats ggml_build_forward_select: the unselected branch is
forced to run with inputs that were never uploaded. In the mtmd audio
graph this makes GEN_WAV calls execute the GEN_CODE branch with a
stale inp_code0, hitting the get_rows bound assert on CPU.

Add ggml_build_forward_order, which inserts nodes without the compute
flag; the flag is restored when the branch is actually selected.
Switch the q/k/v hints in clip_graph::build_attn to it.

* nit: reduce comments (AGENTS.md)
2026-08-07 21:59:49 +03:00
..
cmake ggml : Parallelize quant LUT init (llama/23595) 2026-05-25 12:26:07 +03:00
include mtmd/ggml: add ggml_build_forward_order (llama/26649) 2026-08-07 21:59:49 +03:00
src mtmd/ggml: add ggml_build_forward_order (llama/26649) 2026-08-07 21:59:49 +03:00
.gitignore whisper : reorganize source code + improve CMake (#2256) 2024-06-26 19:34:09 +03:00
CMakeLists.txt ggml : bump version to 0.18.1 (ggml/1578) 2026-08-04 13:37:47 +03:00