make_bfloat162

Applicability

Product	Supported
Atlas 350 Accelerator Card	√
Atlas A3 training product/Atlas A3 inference product	x
Atlas A2 training product/Atlas A2 inference product	x
Atlas 200I/500 A2 inference product	x
Atlas inference product AI Core	x
Atlas inference product Vector Core	x
Atlas training product	x

Function Usage

Creates a bfloat16x2_t vector from two given bfloat16_t values. For 16-bit data types, the underlying instructions support performing computations on two data elements simultaneously. Computational efficiency can be improved through vector type construction.

Prototype

inline bfloat16x2_t make_bfloat162(bfloat16_t x, bfloat16_t y)

Parameters

**Table 1** Parameters
Parameter	Input/Output	Description
x	Input	Source operand.
y	Input	Source operand.

Returns

A new bfloat16x2_t vector consisting of the two bfloat16 values

Restrictions

For SIMT programming, this API is not supported.

Header Files to Be Included

To use this API, the simt_api/asc_bf16.h header file must be included.

#include "simt_api/asc_bf16.h"

Examples

For SIMD and SIMT programming:

__simt_vf__ __launch_bounds__(1024) inline void kernel_make_bfloat162(__gm__ bfloat16x2_t* dst, __gm__ bfloat16_t* x, __gm__ bfloat16_t* y)
{
    int idx = threadIdx.x + blockIdx.x * blockDim.x;
    dst[idx] = make_bfloat162(x[idx], y[idx]);
}

Parent topic: Vector Type Constructor