opencl: refactor some ops, concat, repeat, tanh and scale (#19226)
* opencl: refactor concat * opencl: refactor repeat * opencl: refactor tanh * opencl: enable fp16 for tanh * opencl: refactor scale * opencl: fix unused variables
This commit is contained in:
@@ -1,9 +1,19 @@
|
||||
#pragma OPENCL EXTENSION cl_khr_fp16 : enable
|
||||
|
||||
//------------------------------------------------------------------------------
|
||||
// scale
|
||||
//------------------------------------------------------------------------------
|
||||
kernel void kernel_scale(
|
||||
kernel void kernel_scale_f32(
|
||||
global float * src0,
|
||||
ulong offset0,
|
||||
global float * dst,
|
||||
ulong offsetd,
|
||||
float scale,
|
||||
float bias
|
||||
) {
|
||||
src0 = (global float*)((global char*)src0 + offset0);
|
||||
dst = (global float*)((global char*)dst + offsetd);
|
||||
dst[get_global_id(0)] = src0[get_global_id(0)] * scale + bias;
|
||||
}
|
||||
|
||||
kernel void kernel_scale_f32_4(
|
||||
global float4 * src0,
|
||||
ulong offset0,
|
||||
global float4 * dst,
|
||||
|
||||
Reference in New Issue
Block a user