mikhail/llama.cpp-mtp-turboquant
1
0
Fork 0
Code Issues Pull Requests Actions 34 Packages Projects Releases Wiki Activity

ggml : add ggml_build_forward_select (#18550)

Browse Source 
* ggml : add ggml_build_forward_select

* cuda : adapt CUDA graph compat to new feature

* vulkan : update logic to handle command buffer closing

* ggml : check compute for fusion

* ggml : add comment
This commit is contained in:
Georgi Gerganov
2026-01-19 20:03:19 +02:00
committed by GitHub
parent 3d55846a5c
commit 365a3e8c31
17 changed files with 148 additions and 28 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ggml/src/ggml-backend.cpp
+3 -2
View File
@@ -874,9 +874,9 @@ static void ggml_backend_sched_print_assignments(ggml_backend_sched_t sched, str
}
if (sched->debug > 1) {
ggml_backend_t tensor_backend = ggml_backend_sched_get_tensor_backend(sched, node);
GGML_LOG_DEBUG("node #%3d (%10.10s): %20.20s (%5.5s) [%5.5s %8.8s] use=%d:", i, ggml_op_name(node->op), node->name,
GGML_LOG_DEBUG("node #%3d (%10.10s): %20.20s (%5.5s) [%5.5s %8.8s] use=%d,c=%d:", i, ggml_op_name(node->op), node->name,
fmt_size(ggml_nbytes(node)), tensor_backend ? ggml_backend_name(tensor_backend) : "NULL", GET_CAUSE(node),
graph->use_counts[ggml_hash_find(&graph->visited_hash_set, node)]);
graph->use_counts[ggml_hash_find(&graph->visited_hash_set, node)], node->flags & GGML_TENSOR_FLAG_COMPUTE ? 1 : 0);
for (int j = 0; j < GGML_MAX_SRC; j++) {
struct ggml_tensor * src = node->src[j];
if (src == NULL) {
@@ -1922,6 +1922,7 @@ static struct ggml_tensor * graph_copy_dup_tensor(struct ggml_hash_set hash_set,
dst->view_offs = src->view_offs;
}
dst->op = src->op;
dst->flags = src->flags;
memcpy(dst->op_params, src->op_params, sizeof(dst->op_params));
ggml_set_name(dst, src->name);