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

ggml-cuda: check for srcs outside the cgraph (#18583) 

* ggml-cuda: check for srcs outside the cgraph

* review: use leafs instead
This commit is contained in:
Aman Gupta 2026-01-05 22:46:36 +08:00 committed by GitHub
parent da143b9940
commit f47edb8c19
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 15 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
ggml/src/ggml-cuda/common.cuh
View File

@ -1065,6 +1065,7 @@ struct ggml_cuda_graph {
int number_consecutive_updates = 0; int number_consecutive_updates = 0;
bool cuda_graphs_enabled = false; bool cuda_graphs_enabled = false;
std::vector<ggml_graph_node_properties> ggml_graph_properties; std::vector<ggml_graph_node_properties> ggml_graph_properties;
std::vector<ggml_graph_node_properties> extraneous_srcs_properties;
#endif #endif
}; };

16
ggml/src/ggml-cuda/ggml-cuda.cu
View File

@ -2973,15 +2973,16 @@ static bool is_cuda_graph_update_required(ggml_backend_cuda_context * cuda_ctx,
} }
// Check if the graph size has changed // Check if the graph size has changed
if (cuda_ctx->cuda_graph->ggml_graph_properties.size() != (size_t)cgraph->n_nodes) { if (cuda_ctx->cuda_graph->ggml_graph_properties.size() != (size_t)cgraph->n_nodes + cgraph->n_leafs) {
cuda_graph_update_required = true; cuda_graph_update_required = true;
cuda_ctx->cuda_graph->ggml_graph_properties.resize(cgraph->n_nodes); cuda_ctx->cuda_graph->ggml_graph_properties.resize(cgraph->n_nodes + cgraph->n_leafs);
} }
// Loop over nodes in GGML graph to determine if CUDA graph update is required // Loop over nodes in GGML graph to determine if CUDA graph update is required
// and store properties to allow this comparison for the next token // and store properties to allow this comparison for the next token
for (int i = 0; i < cgraph->n_nodes; i++) { for (int i = 0; i < cgraph->n_nodes; i++) {
bool has_matching_properties = true; bool has_matching_properties = true;
if (!cuda_graph_update_required) { if (!cuda_graph_update_required) {
has_matching_properties = ggml_graph_node_has_matching_properties(cgraph->nodes[i], &cuda_ctx->cuda_graph->ggml_graph_properties[i]); has_matching_properties = ggml_graph_node_has_matching_properties(cgraph->nodes[i], &cuda_ctx->cuda_graph->ggml_graph_properties[i]);
} }
@ -2991,6 +2992,17 @@ static bool is_cuda_graph_update_required(ggml_backend_cuda_context * cuda_ctx,
set_ggml_graph_node_properties(cgraph->nodes[i], &cuda_ctx->cuda_graph->ggml_graph_properties[i]); set_ggml_graph_node_properties(cgraph->nodes[i], &cuda_ctx->cuda_graph->ggml_graph_properties[i]);
} }
for (int i = 0; i < cgraph->n_leafs; i++) {
bool has_matching_properties = true;
if (!cuda_graph_update_required) {
has_matching_properties = ggml_graph_node_has_matching_properties(cgraph->leafs[i], &cuda_ctx->cuda_graph->ggml_graph_properties[cgraph->n_nodes + i]);
}
if (!has_matching_properties) {
cuda_graph_update_required = true;
}
set_ggml_graph_node_properties(cgraph->leafs[i], &cuda_ctx->cuda_graph->ggml_graph_properties[cgraph->n_nodes + i]);
}
return cuda_graph_update_required; return cuda_graph_update_required;
} }