[object Object][object Object][object Object]undefined[object Object]

Converts the vector_float type to the vector_bfloat16_t type, writes to the upper half or lower half of the destination operand, and supports multiple rounding modes and saturation modes:

• RINT rounding mode: round half to even.

• ROUND rounding mode: round half up.

• FLOOR rounding mode: round toward negative infinity.

• CEIL rounding mode: round toward positive infinity.

• TRUNC rounding mode: round toward zero.

Taking the asc_float2bfloat16_rd interface as an example:

• asc_float2bfloat16_rd: non-saturation mode, writes to the upper half of the destination operand.

• asc_float2bfloat16_rd_sat: saturation mode, writes to the upper half of the destination operand.

• asc_float2bfloat16_rd_v2: non-saturation mode, writes to the lower half of the destination operand.

• asc_float2bfloat16_rd_sat_v2: saturation mode, writes to the lower half of the destination operand.

[object Object][object Object][object Object]undefined

For detailed descriptions of vector data registers and mask registers, see .

[object Object]

None

[object Object]

PIPE_V

[object Object]

• To enable the saturation mode and non-saturation mode, the ctrl register must be configured. For detailed descriptions of the ctrl register, see .

[object Object]