huggingface/candle
1
0
Fork 0
mirror of https://github.com/huggingface/candle.git synced 2025-06-21 20:22:49 +00:00
Code Issues Packages Projects Releases Wiki Activity

Refactor the hierarchy.

Browse Source
This commit is contained in:
Nicolas Patry
2023-06-27 11:57:27 +02:00
parent 6c4a960b15
commit d7f729fb8f
41 changed files with 35 additions and 33 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
candle-kernels/src/fill.cu Normal file
View File

@ -0,0 +1,11 @@
#include "cuda_fp16.h"
template<typename T>
__device__ void fill_with(T *buf, T value, const size_t numel) {
for (unsigned int i = blockIdx.x * blockDim.x + threadIdx.x; i < numel; i += blockDim.x * gridDim.x) {
buf[i] = value;
}
}
extern "C" __global__ void fill_f16(__half *buf, __half value, const size_t numel) { fill_with(buf, value, numel); }
extern "C" __global__ void fill_f32(float *buf, float value, const size_t numel) { fill_with(buf, value, numel); }
extern "C" __global__ void fill_f64(double *buf, double value, const size_t numel) { fill_with(buf, value, numel); }