mirror of
https://github.com/ggml-org/llama.cpp.git
synced 2025-11-19 11:57:07 +00:00
vulkan: fuse rms_norm + mul + rope (+ view + set_rows) (#16977)
This change combines the rms_norm+mul and rope+view+set_rows fusions to allow fusing the whole sequence together. This comes up in Qwen3, Bailing, and some other models.
This commit is contained in:
@@ -695,6 +695,8 @@ void process_shaders() {
|
||||
string_to_spv("group_norm_f32", "group_norm.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"D_TYPE", "float"}}));
|
||||
string_to_spv("rms_norm_f32", "rms_norm.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"B_TYPE", "float"}, {"D_TYPE", "float"}}));
|
||||
string_to_spv("rms_norm_partials_f32", "rms_norm_partials.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"B_TYPE", "float"}, {"D_TYPE", "float"}}));
|
||||
string_to_spv("rms_norm_mul_rope_f32_f32", "rms_norm.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"B_TYPE", "float"}, {"D_TYPE", "float"}, {"ROPE_D_TYPE", "float"}, {"RMS_NORM_ROPE_FUSION", "1"}}));
|
||||
string_to_spv("rms_norm_mul_rope_f32_f16_rte", "rms_norm.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"B_TYPE", "float"}, {"D_TYPE", "float"}, {"ROPE_D_TYPE", "float16_t"}, {"RMS_NORM_ROPE_FUSION", "1"}, {"RTE16", "1"}}));
|
||||
string_to_spv("rms_norm_back_f32", "rms_norm_back.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"B_TYPE", "float"}, {"D_TYPE", "float"}}));
|
||||
string_to_spv("l2_norm_f32", "l2_norm.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"D_TYPE", "float"}}));
|
||||
|
||||
@@ -840,25 +842,25 @@ void process_shaders() {
|
||||
string_to_spv("soft_max_f32_f16", "soft_max.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"B_TYPE", "float16_t"}, {"D_TYPE", "float"}}));
|
||||
string_to_spv("soft_max_back_f32", "soft_max_back.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"B_TYPE", "float"}, {"D_TYPE", "float"}}));
|
||||
|
||||
string_to_spv("rope_norm_f32", "rope_norm.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}});
|
||||
string_to_spv("rope_norm_f16", "rope_norm.comp", {{"A_TYPE", "float16_t"}, {"D_TYPE", "float16_t"}});
|
||||
string_to_spv("rope_norm_f16_rte", "rope_norm.comp", {{"A_TYPE", "float16_t"}, {"D_TYPE", "float16_t"}, {"RTE16", "1"}});
|
||||
string_to_spv("rope_norm_f32_f16", "rope_norm.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float16_t"}});
|
||||
string_to_spv("rope_norm_f32_f16_rte", "rope_norm.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float16_t"}, {"RTE16", "1"}});
|
||||
string_to_spv("rope_norm_f32", "rope_norm.comp", {{"A_TYPE", "float"}, {"ROPE_D_TYPE", "float"}});
|
||||
string_to_spv("rope_norm_f16", "rope_norm.comp", {{"A_TYPE", "float16_t"}, {"ROPE_D_TYPE", "float16_t"}});
|
||||
string_to_spv("rope_norm_f16_rte", "rope_norm.comp", {{"A_TYPE", "float16_t"}, {"ROPE_D_TYPE", "float16_t"}, {"RTE16", "1"}});
|
||||
string_to_spv("rope_norm_f32_f16", "rope_norm.comp", {{"A_TYPE", "float"}, {"ROPE_D_TYPE", "float16_t"}});
|
||||
string_to_spv("rope_norm_f32_f16_rte", "rope_norm.comp", {{"A_TYPE", "float"}, {"ROPE_D_TYPE", "float16_t"}, {"RTE16", "1"}});
|
||||
|
||||
string_to_spv("rope_neox_f32", "rope_neox.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}});
|
||||
string_to_spv("rope_neox_f16", "rope_neox.comp", {{"A_TYPE", "float16_t"}, {"D_TYPE", "float16_t"}});
|
||||
string_to_spv("rope_neox_f16_rte", "rope_neox.comp", {{"A_TYPE", "float16_t"}, {"D_TYPE", "float16_t"}, {"RTE16", "1"}});
|
||||
string_to_spv("rope_neox_f32_f16", "rope_neox.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float16_t"}});
|
||||
string_to_spv("rope_neox_f32_f16_rte", "rope_neox.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float16_t"}, {"RTE16", "1"}});
|
||||
string_to_spv("rope_neox_f32", "rope_neox.comp", {{"A_TYPE", "float"}, {"ROPE_D_TYPE", "float"}});
|
||||
string_to_spv("rope_neox_f16", "rope_neox.comp", {{"A_TYPE", "float16_t"}, {"ROPE_D_TYPE", "float16_t"}});
|
||||
string_to_spv("rope_neox_f16_rte", "rope_neox.comp", {{"A_TYPE", "float16_t"}, {"ROPE_D_TYPE", "float16_t"}, {"RTE16", "1"}});
|
||||
string_to_spv("rope_neox_f32_f16", "rope_neox.comp", {{"A_TYPE", "float"}, {"ROPE_D_TYPE", "float16_t"}});
|
||||
string_to_spv("rope_neox_f32_f16_rte", "rope_neox.comp", {{"A_TYPE", "float"}, {"ROPE_D_TYPE", "float16_t"}, {"RTE16", "1"}});
|
||||
|
||||
string_to_spv("rope_multi_f32", "rope_multi.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}});
|
||||
string_to_spv("rope_multi_f16", "rope_multi.comp", {{"A_TYPE", "float16_t"}, {"D_TYPE", "float16_t"}});
|
||||
string_to_spv("rope_multi_f16_rte", "rope_multi.comp", {{"A_TYPE", "float16_t"}, {"D_TYPE", "float16_t"}, {"RTE16", "1"}});
|
||||
string_to_spv("rope_multi_f32", "rope_multi.comp", {{"A_TYPE", "float"}, {"ROPE_D_TYPE", "float"}});
|
||||
string_to_spv("rope_multi_f16", "rope_multi.comp", {{"A_TYPE", "float16_t"}, {"ROPE_D_TYPE", "float16_t"}});
|
||||
string_to_spv("rope_multi_f16_rte", "rope_multi.comp", {{"A_TYPE", "float16_t"}, {"ROPE_D_TYPE", "float16_t"}, {"RTE16", "1"}});
|
||||
|
||||
string_to_spv("rope_vision_f32", "rope_vision.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}});
|
||||
string_to_spv("rope_vision_f16", "rope_vision.comp", {{"A_TYPE", "float16_t"}, {"D_TYPE", "float16_t"}});
|
||||
string_to_spv("rope_vision_f16_rte", "rope_vision.comp", {{"A_TYPE", "float16_t"}, {"D_TYPE", "float16_t"}, {"RTE16", "1"}});
|
||||
string_to_spv("rope_vision_f32", "rope_vision.comp", {{"A_TYPE", "float"}, {"ROPE_D_TYPE", "float"}});
|
||||
string_to_spv("rope_vision_f16", "rope_vision.comp", {{"A_TYPE", "float16_t"}, {"ROPE_D_TYPE", "float16_t"}});
|
||||
string_to_spv("rope_vision_f16_rte", "rope_vision.comp", {{"A_TYPE", "float16_t"}, {"ROPE_D_TYPE", "float16_t"}, {"RTE16", "1"}});
|
||||
|
||||
string_to_spv("argsort_f32", "argsort.comp", {{"A_TYPE", "float"}});
|
||||
|
||||
|
||||
Reference in New Issue
Block a user