ver4a/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

ggml: Re-enable CUDA graphs in presence of CONT and DUP nodes (#12970)

Browse source
This commit is contained in:
Alan Gray 2025-04-17 14:19:42 +01:00 • committed by GitHub
parent 7a395f67a7
commit 207c22ec2d
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
3 changed files with 7 additions and 6 deletions
Download patch file Download diff file Expand all files Collapse all files

2
ggml/src/ggml-cuda/ggml-cuda.cu
View file

@ -2489,7 +2489,7 @@ static bool check_node_graph_compatibility_and_refresh_copy_ops(ggml_backend_cud
#endif
}
if (node->op == GGML_OP_MUL_MAT_ID || node->op == GGML_OP_CONT || node->op == GGML_OP_DUP) {
if (node->op == GGML_OP_MUL_MAT_ID) {
use_cuda_graph = false; // This node type is not supported by CUDA graph capture
#ifndef NDEBUG
GGML_LOG_DEBUG("%s: disabling CUDA graphs due to unsupported node type\n", __func__);