diff --git a/common/common.cpp b/common/common.cpp index efbdd00e2..55fec9211 100644 --- a/common/common.cpp +++ b/common/common.cpp @@ -655,7 +655,7 @@ bool gpt_params_parse_ex(int argc, char ** argv, gpt_params & params) { params.use_mmap = false; } else if (arg == "--numa") { if (++i >= argc) { - invalid_param = true; + invalid_param = true; break; } else { std::string value(argv[i]); diff --git a/common/common.h b/common/common.h index 9b20c6f6f..2c864c04c 100644 --- a/common/common.h +++ b/common/common.h @@ -76,7 +76,7 @@ struct gpt_params { float yarn_beta_slow = 1.0f; // YaRN high correction dim int32_t yarn_orig_ctx = 0; // YaRN original context length int32_t rope_scaling_type = LLAMA_ROPE_SCALING_UNSPECIFIED; - int32_t numa = LLAMA_NUMA_STRATEGY_DISABLED; + int32_t numa = LLAMA_NUMA_STRATEGY_DISABLED; // // sampling parameters struct llama_sampling_params sparams; diff --git a/ggml.c b/ggml.c index 4aaf193c2..6922934e3 100644 --- a/ggml.c +++ b/ggml.c @@ -16630,7 +16630,7 @@ static void set_numa_thread_affinity(int thread_n, int n_threads) { break; case GGML_NUMA_STRATEGY_NUMACTL: // use the cpuset that numactl gave us - rv = pthread_setaffinity_np(pthread_self(), setsize, &g_state.numa.cpuset); + rv = pthread_setaffinity_np(pthread_self(), setsize, &g_state.numa.cpuset); if (rv) { fprintf(stderr, "warning: pthread_setaffinity_np() failed: %s\n",strerror(rv)); }