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

Simplified specialization constants by limiting to floats2

Browse source
This commit is contained in:
Alejandro Saucedo 2021-02-14 06:55:09 +00:00
parent 7126cc47ff
commit a7801cedd0
5 changed files with 358 additions and 470 deletions
Download patch file Download diff file Expand all files Collapse all files

6
src/include/kompute/operations/OpAlgoBase.hpp
View file

@ -50,7 +50,7 @@ class OpAlgoBase : public OpBase
std::shared_ptr<vk::CommandBuffer> commandBuffer,
std::vector<std::shared_ptr<Tensor>>& tensors,
KomputeWorkgroup komputeWorkgroup = {},
const Algorithm::SpecializationContainer& specializationConstants = {});
const std::vector<float>& specializationConstants = {});
/**
* Constructor that enables a file to be passed to the operation with
@ -70,7 +70,7 @@ class OpAlgoBase : public OpBase
std::vector<std::shared_ptr<Tensor>>& tensors,
std::string shaderFilePath,
KomputeWorkgroup komputeWorkgroup = {},
const Algorithm::SpecializationContainer& specializationConstants = {});
const std::vector<float>& specializationConstants = {});
/**
* Constructor that enables raw shader data to be passed to the main operation
@ -89,7 +89,7 @@ class OpAlgoBase : public OpBase
std::vector<std::shared_ptr<Tensor>>& tensors,
const std::vector<char>& shaderDataRaw,
KomputeWorkgroup komputeWorkgroup = {},
const Algorithm::SpecializationContainer& specializationConstants = {});
const std::vector<float>& specializationConstants = {});
/**
* Default destructor, which is in charge of destroying the algorithm