mirror of
https://github.com/ggml-org/llama.cpp.git
synced 2026-09-07 13:31:13 +02:00
a97123e497
* support host pinned mem, ggml_backend_sycl_host_buffer_type_get_max_size, * fix the thread-safe issue