thek0tyara/llama-cpp-turboquant
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 4

Added capabilities for specialization data based on the size of the vectors passed

Browse source
This commit is contained in:
Alejandro Saucedo 2020-08-29 17:03:24 +01:00
parent b80548ac3d
commit 23cf43e231
5 changed files with 140 additions and 109 deletions
Download patch file Download diff file Expand all files Collapse all files

4
shaders/glsl/opmult.comp
View file

@ -12,6 +12,10 @@ layout(set = 0, binding = 2) buffer tensorOutput {
uint valuesOutput[ ];
};
layout (constant_id = 0) const uint LEN_LHS = 0;
layout (constant_id = 1) const uint LEN_RHS = 0;
layout (constant_id = 2) const uint LEN_OUT = 0;
// TODO: Explore how to make layout inside shader dynamic
layout (local_size_x = 1, local_size_y = 1, local_size_z = 1) in;