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

ggml-virtgpu: backend: don't continue if couldn't allocate the tensor memory

This commit is contained in:
Kevin Pouget 2026-01-29 09:25:09 +01:00
parent 171ab8b4ad
commit 3864be3e5d
1 changed files with 6 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
ggml/src/ggml-virtgpu/backend/shared/apir_cs_ggml.h
View File

@ -39,11 +39,17 @@ static inline void apir_encode_ggml_tensor(apir_encoder * enc, const ggml_tensor
static inline const ggml_tensor * apir_decode_ggml_tensor(apir_decoder * dec) {
const apir_rpc_tensor * apir_rpc_tensor = apir_decode_apir_rpc_tensor_inplace(dec);
if (!apir_rpc_tensor) {
return NULL;
}
ggml_init_params params{
/*.mem_size =*/ ggml_tensor_overhead(),
/*.mem_buffer =*/ NULL,
/*.no_alloc =*/ true,
};
ggml_context * ctx = ggml_init(params);
const ggml_tensor * tensor = apir_deserialize_tensor(ctx, apir_rpc_tensor);