Lumpiasty/llama.cpp
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

vulkan: top_k radix select for k >= 1024 for Qwen 3.8 Flash Next (#28032)

Browse Source 
* vulkan: add top-k radix sort shader for k >= 1024

* add Qwen 3.8 Flash Next top-k tests

* add top-k qsa fusion

* clean up code
This commit is contained in:
Ruben Ortlam
2026-08-31 07:04:34 +02:00
committed by GitHub
parent 9723942adc
commit daef7b6874
4 changed files with 471 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ggml/src/ggml-vulkan/vulkan-shaders/vulkan-shaders-gen.cpp
+1
View File
@@ -1028,6 +1028,7 @@ void process_shaders() {
string_to_spv("topk_argsort_f32", "topk_argsort.comp", {{"A_TYPE", "float"}});
string_to_spv("topk_nary_search_f32", "topk_nary_search.comp", {{"A_TYPE", "float"}});
string_to_spv("topk_radix_select_f32", "topk_radix_select.comp", {{"A_TYPE", "float"}});
string_to_spv("argmax_f32", "argmax.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"D_TYPE", "int"}}));
string_to_spv("sum_rows_f32", "sum_rows.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"D_TYPE", "float"}}));