mikhail/llama.cpp-mtp-turboquant
1
0
Fork 0
Code Issues Pull Requests Actions 31 Packages Projects Releases Wiki Activity

vulkan: add more num_blocks instantiations in rms_norm (#17701)

Browse Source
This commit is contained in:
Jeff Bolz
2025-12-05 15:08:56 -06:00
committed by GitHub
parent a0f3897d53
commit 933414c0b6
1 changed files with 5 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ggml/src/ggml-vulkan/vulkan-shaders/rms_norm.comp
+5 -1
View File
@@ -131,8 +131,12 @@ void main() {
rms_norm(num_blocks);
} else if (num_blocks > 16) {
rms_norm(32);
} else if (num_blocks > 8) {
} else if (num_blocks > 12) {
rms_norm(16);
} else if (num_blocks > 10) {
rms_norm(12);
} else if (num_blocks > 8) {
rms_norm(10);
} else if (num_blocks > 4) {
rms_norm(8);
} else if (num_blocks == 4) {