LLM inference in C/C++
C++ ggml
#pragma OPENCL EXTENSION cl_khr_fp16 : enable
//------------------------------------------------------------------------------
// clamp
//------------------------------------------------------------------------------
kernel void kernel_clamp(
global float * src0,
ulong offset0,
global float * dst,
ulong offsetd,
float min,
float max
) {
src0 = (global float*)((global char*)src0 + offset0);
dst = (global float*)((global char*)dst + offsetd);
dst[get_global_id(0)] = src0[get_global_id(0)] < min ?
min :
(src0[get_global_id(0)] > max ? max : src0[get_global_id(0)]);
}
kernel void kernel_clamp_nc(
global const char * src0,
ulong offset0,
global char * dst,
ulong offsetd,
float min,
float max,
int ne00,
ulong nb00,
ulong nb01,
ulong nb02,
ulong nb03,
ulong nb0,
ulong nb1,
ulong nb2,
ulong nb3
) {
src0 = src0 + offset0;
dst = dst + offsetd;
const int i3 = get_group_id(2);
const int i2 = get_group_id(1);
const int i1 = get_group_id(0);
for (int i0 = get_local_id(0); i0 < ne00; i0 += get_local_size(0)) {
const float x = *(global const float *)(src0 + i3*nb03 + i2*nb02 + i1*nb01 + i0*nb00);
*(global float *)(dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0) = x < min ? min : (x > max ? max : x);
}
}