Method graph_init

smallthinker/src/llama-context.cpp:1297–1307

}

ggml_cgraph * llama_context::graph_init() {
    ggml_init_params params = {
        /*.mem_size   =*/ buf_compute_meta.size(),
        /*.mem_buffer =*/ buf_compute_meta.data(),
        /*.no_alloc   =*/ true,
    };

    ctx_compute.reset(ggml_init(params));

    return ggml_new_graph_custom(ctx_compute.get(), graph_max_nodes(), false);
}

ggml_cgraph * llama_context::graph_reserve(uint32_t n_tokens, uint32_t n_seqs, uint32_t n_outputs, const llama_memory_state_i * mstate) {
    LLAMA_LOG_DEBUG("%s: reserving a graph for ubatch with n_tokens = %4u, n_seqs = %2u, n_outputs = %4u\n", __func__, n_tokens, n_seqs, n_outputs);