llama : drop the redundant cnv from llm_attn_mask (#31558)

The dspark block mask is now derived at input-fill time: llm_attn_mask::block_uniform
fills every token of a sequence with the mask row of its first token, so the
graph-level reshape/repeat in dflash is gone and cnv always aliases raw.

Assisted-by: dsh:llama.cpp/ggml-org/GLM-5.3-Flash-GGUF (Mac Studio, M5 Ultra, 256GB)
This commit is contained in:
Georgi Gerganov authored and GitHub committed 2026-11-03 08:11:06 +02:00
1 parent 0f663d2471
commit ca8daad199
14 files changed
+46 -69

No files matched your search

+2 -2
View File
@@ -256,7 +256,7 @@ llama_model_dots3note::graph::graph(const llama_model & model, const llm_graph_p
ggml_tensor * indexer_score = nullptr;
if (cparams.fused_lid) {
indexer_score = ggml_lightning_indexer(ctx0, indexer_q, indexer_k, indexer_weights, inp_attn->get_dsa()->self_attn_mask_lid.cnv);
indexer_score = ggml_lightning_indexer(ctx0, indexer_q, indexer_k, indexer_weights, inp_attn->get_dsa()->self_attn_mask_lid.raw);
cb(indexer_score, "indexer_score", il);
res->add_fused_node({LLM_FUSED_OP_LIGHTNING_INDEXER, indexer_score, il});
} else {
@@ -286,7 +286,7 @@ llama_model_dots3note::graph::graph(const llama_model & model, const llm_graph_p
indexer_score = ggml_cont(ctx0, ggml_permute(ctx0, indexer_score, 2, 1, 0, 3));
cb(indexer_score, "indexer_score", il);
ggml_tensor * indexer_kq_mask = inp_attn->get_dsa()->self_attn_mask_lid.cnv;
ggml_tensor * indexer_kq_mask = inp_attn->get_dsa()->self_attn_mask_lid.raw;
indexer_score = ggml_add(ctx0, indexer_score, indexer_kq_mask);
cb(indexer_score, "indexer_score", il);
}