mirror of
https://github.com/ggerganov/llama.cpp.git
synced 2024-12-26 14:20:31 +01:00
7c7836d9d4
* Refactor shaders, extract GLSL code from ggml_vk_generate_shaders.py into vulkan-shaders directory * Improve debug log code * Add memory debug output option * Fix flake8 * Fix unnecessary high llama-3 VRAM use
14 lines
315 B
Plaintext
14 lines
315 B
Plaintext
#version 450
|
|
|
|
#include "types.comp"
|
|
#include "generic_unary_head.comp"
|
|
|
|
void main() {
|
|
if (gl_GlobalInvocationID.x >= p.ne) {
|
|
return;
|
|
}
|
|
|
|
const FLOAT_TYPE val = FLOAT_TYPE(data_a[src0_idx(gl_GlobalInvocationID.x)]);
|
|
data_d[p.d_offset + dst_idx(gl_GlobalInvocationID.x)] = D_TYPE(val * val);
|
|
}
|