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

minor: rename ctx as plan; const

Browse source
This commit is contained in:
mqy 2023-07-03 16:22:52 +08:00
parent b11ac01f6b
commit a37de23953
2 changed files with 32 additions and 32 deletions
Download patch file Download diff file Expand all files Collapse all files

50
ggml.c
View file

@ -15941,13 +15941,13 @@ void clear_numa_thread_affinity(void) {}
#endif #endif
struct ggml_compute_state_shared { struct ggml_compute_state_shared {
struct ggml_cgraph * cgraph; const struct ggml_cgraph * cgraph;
struct ggml_graph_compute_plan * cgraph_ctx; const struct ggml_graph_compute_plan * plan;
int64_t perf_node_start_cycles; int64_t perf_node_start_cycles;
int64_t perf_node_start_time_us; int64_t perf_node_start_time_us;
int n_threads; const int n_threads;
// synchronization primitives // synchronization primitives
atomic_int n_active; // num active threads atomic_int n_active; // num active threads
@ -15971,10 +15971,10 @@ static void ggml_graph_compute_perf_stats_node(struct ggml_tensor * node, const
static thread_ret_t ggml_graph_compute_thread(void * data) { static thread_ret_t ggml_graph_compute_thread(void * data) {
struct ggml_compute_state * state = (struct ggml_compute_state *) data; struct ggml_compute_state * state = (struct ggml_compute_state *) data;
struct ggml_cgraph * cgraph = state->shared->cgraph; const struct ggml_cgraph * cgraph = state->shared->cgraph;
struct ggml_graph_compute_plan * ctx = state->shared->cgraph_ctx; const struct ggml_graph_compute_plan * plan = state->shared->plan;
const int *n_tasks_arr = ctx->n_tasks; const int *n_tasks_arr = plan->n_tasks;
const int n_threads = state->shared->n_threads; const int n_threads = state->shared->n_threads;
set_numa_thread_affinity(state->ith, n_threads); set_numa_thread_affinity(state->ith, n_threads);
@ -15989,8 +15989,8 @@ static thread_ret_t ggml_graph_compute_thread(void * data) {
/*.type =*/ GGML_TASK_FINALIZE, /*.type =*/ GGML_TASK_FINALIZE,
/*.ith =*/ 0, /*.ith =*/ 0,
/*.nth =*/ 0, /*.nth =*/ 0,
/*.wsize =*/ ctx->work_size, /*.wsize =*/ plan->work_size,
/*.wdata =*/ ctx->work_data, /*.wdata =*/ plan->work_data,
}; };
if (node_n != -1) { if (node_n != -1) {
@ -16059,8 +16059,8 @@ static thread_ret_t ggml_graph_compute_thread(void * data) {
/*.type =*/ GGML_TASK_COMPUTE, /*.type =*/ GGML_TASK_COMPUTE,
/*.ith =*/ state->ith, /*.ith =*/ state->ith,
/*.nth =*/ n_tasks, /*.nth =*/ n_tasks,
/*.wsize =*/ ctx->work_size, /*.wsize =*/ plan->work_size,
/*.wdata =*/ ctx->work_data, /*.wdata =*/ plan->work_data,
}; };
if (state->ith < n_tasks) { if (state->ith < n_tasks) {
@ -16077,9 +16077,9 @@ struct ggml_graph_compute_plan ggml_graph_compute_make_plan(struct ggml_cgraph *
n_threads = GGML_DEFAULT_N_THREADS; n_threads = GGML_DEFAULT_N_THREADS;
} }
struct ggml_graph_compute_plan ctx; struct ggml_graph_compute_plan plan;
memset(&ctx, 0, sizeof(struct ggml_graph_compute_plan)); memset(&plan, 0, sizeof(struct ggml_graph_compute_plan));
int * n_tasks = ctx.n_tasks; int * n_tasks = plan.n_tasks;
size_t work_size = 0; size_t work_size = 0;
// initialize tasks + work buffer // initialize tasks + work buffer
@ -16403,35 +16403,35 @@ struct ggml_graph_compute_plan ggml_graph_compute_make_plan(struct ggml_cgraph *
work_size += CACHE_LINE_SIZE*(n_threads - 1); work_size += CACHE_LINE_SIZE*(n_threads - 1);
} }
ctx.n_threads = n_threads; plan.n_threads = n_threads;
ctx.work_size = work_size; plan.work_size = work_size;
ctx.work_data = NULL; plan.work_data = NULL;
return ctx; return plan;
} }
void ggml_graph_compute(struct ggml_graph_compute_plan * ctx, struct ggml_cgraph * cgraph) { void ggml_graph_compute(struct ggml_graph_compute_plan * plan, struct ggml_cgraph * cgraph) {
{ {
GGML_ASSERT(ctx); GGML_ASSERT(plan);
GGML_ASSERT(ctx->n_threads > 0); GGML_ASSERT(plan->n_threads > 0);
if (ctx->work_size > 0) { if (plan->work_size > 0) {
GGML_ASSERT(ctx->work_data); GGML_ASSERT(plan->work_data);
} }
for (int i = 0; i < cgraph->n_nodes; ++i) { for (int i = 0; i < cgraph->n_nodes; ++i) {
if (cgraph->nodes[i]->op != GGML_OP_NONE) { if (cgraph->nodes[i]->op != GGML_OP_NONE) {
GGML_ASSERT(ctx->n_tasks[i] > 0); GGML_ASSERT(plan->n_tasks[i] > 0);
} }
} }
} }
const int n_threads = ctx->n_threads; const int n_threads = plan->n_threads;
struct ggml_compute_state_shared state_shared = { struct ggml_compute_state_shared state_shared = {
/*.cgraph =*/ cgraph, /*.cgraph =*/ cgraph,
/*.cgraph_ctx =*/ ctx, /*.cgraph_plan =*/ plan,
/*.perf_node_start_cycles =*/ 0, /*.perf_node_start_cycles =*/ 0,
/*.perf_node_start_time_us =*/ 0, /*.perf_node_start_time_us =*/ 0,
/*.n_threads =*/ n_threads, /*.n_threads =*/ n_threads,

14
ggml.h
View file

@ -66,14 +66,14 @@
// ggml_set_f32(b, 4.0f); // ggml_set_f32(b, 4.0f);
// //
// const int n_threads = 1; // const int n_threads = 1;
// struct ggml_graph_compute_plan ctx = ggml_graph_compute_make_plan(&gf, n_threads); // struct ggml_graph_compute_plan plan = ggml_graph_compute_make_plan(&gf, n_threads);
// if (ctx.work_size > 0) { // if (plan.work_size > 0) {
// ctx.work_data = malloc(ctx.work_size); // plan.work_data = malloc(plan.work_size);
// GGML_ASSERT(ctx.work_data); // GGML_ASSERT(plan.work_data);
// } // }
// ggml_graph_compute(&ctx, &gf); // ggml_graph_compute(&plan, &gf);
// if (ctx.work_data) { // if (plan.work_data) {
// free(ctx.work_data); // free(plan.work_data);
// } // }
// //
// printf("f = %f\n", ggml_get_f32_1d(f, 0)); // printf("f = %f\n", ggml_get_f32_1d(f, 0));