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

Vulkan Shader Refactor, Memory Debugging Option (#7947)

Browse source 
* 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
This commit is contained in:
0cc4m 2024-06-16 07:17:31 +02:00 • committed by GitHub
parent 0c7b3595b9
commit 7c7836d9d4
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
54 changed files with 25266 additions and 21885 deletions
Download patch file Download diff file Expand all files Collapse all files

12
vulkan-shaders/mul.comp Normal file
View file

@ -0,0 +1,12 @@
#version 450
#include "types.comp"
#include "generic_binary_head.comp"
void main() {
if (gl_GlobalInvocationID.x >= p.ne) {
return;
}
data_d[p.d_offset + dst_idx(gl_GlobalInvocationID.x)] = D_TYPE(FLOAT_TYPE(data_a[src0_idx(gl_GlobalInvocationID.x)]) * FLOAT_TYPE(data_b[src1_idx(gl_GlobalInvocationID.x)]));
}