upstream-archive/stable-diffusion.cpp
1
0
Fork 0
mirror of https://github.com/leejet/stable-diffusion.cpp.git synced 2026-08-07 02:30:40 -05:00
Code Issues Packages Projects Releases Wiki Activity

feat: support for cross-device row split (#1735)

Browse Source
This commit is contained in:
Piotr Wilkin (ilintar)
2026-07-04 10:31:18 +02:00
committed by GitHub
parent 7bcd189639
commit 68f3d6df9f
12 changed files with 322 additions and 21 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
src/core/layer_split_partition.cpp
View File

@@ -9,7 +9,7 @@
namespace sd {
static int tensor_block_index(const std::string& name) {
int layer_split_tensor_block_index(const std::string& name) {
static const char* block_keywords[] = {"transformer_blocks.", "joint_blocks.", "double_blocks.",
"single_blocks.", "blocks.", "block.", "layers."};
for (const char* keyword : block_keywords) {
@@ -92,7 +92,7 @@ namespace sd {
int n_blocks = 0;
for (const auto& kv : tensors) {
int64_t bytes = (int64_t)ggml_nbytes(kv.second);
int idx = split_tensors.count(kv.first) != 0 ? tensor_block_index(kv.first) : -1;
int idx = split_tensors.count(kv.first) != 0 ? layer_split_tensor_block_index(kv.first) : -1;
if (idx >= 0) {
block_bytes[idx] += bytes;
total_block_bytes += bytes;
@@ -160,7 +160,7 @@ namespace sd {
for (const auto& kv : tensors) {
size_t target = 0;
int idx = split_tensors.count(kv.first) != 0 ? tensor_block_index(kv.first) : -1;
int idx = split_tensors.count(kv.first) != 0 ? layer_split_tensor_block_index(kv.first) : -1;
if (idx >= 0) {
while (target < boundaries.size() && idx >= boundaries[target]) {
target++;