vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

vulkan: implement GGML_OP_COUNT_EQUAL

Browse source
This commit is contained in:
Rémy O 2025-02-08 14:01:27 +01:00
parent 148f58681b
commit 095f8d17ac
4 changed files with 61 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

31
ggml/src/ggml-vulkan/vulkan-shaders/count_equal.comp Normal file
View file

@ -0,0 +1,31 @@
#version 450
#extension GL_EXT_control_flow_attributes : enable
#include "types.comp"
#include "generic_head.comp"
layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
layout (binding = 0) readonly buffer X {A_TYPE data_a[];};
layout (binding = 1) readonly buffer Y {B_TYPE data_b[];};
layout (binding = 2) buffer D {D_TYPE data_d[];};
const uint CHUNK_SIZE = 512;
void main() {
const uint base = gl_WorkGroupID.x * CHUNK_SIZE;
const uint col = gl_LocalInvocationID.x;
uint count = 0;
[[unroll]]
for (uint i = 0; i < CHUNK_SIZE; i += gl_WorkGroupSize.x) {
const uint idx = base + i + col;
if (idx >= p.KX) {
break;
}
count += uint(data_a[idx] == data_b[idx]);
}
atomicAdd(data_d[0], D_TYPE(count));
}

1
ggml/src/ggml-vulkan/vulkan-shaders/vulkan-shaders-gen.cpp
View file

@ -488,6 +488,7 @@ void process_shaders() {
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"}}));
string_to_spv("count_equal_i32", "count_equal.comp", merge_maps(base_dict, {{"A_TYPE", "int"}, {"B_TYPE", "int"}, {"D_TYPE", "int"}}));
string_to_spv("im2col_f32", "im2col.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"D_TYPE", "float"}}));
string_to_spv("im2col_f32_f16", "im2col.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"D_TYPE", "float16_t"}}));