3.1. Bfloat16 Arithmetic Constants
To use these constants, include the header file cuda_bf16.h in your program.
Macros
- CUDART_INF_BF16
-
Defines floating-point positive infinity value for the
nv_bfloat16data type. - CUDART_MAX_NORMAL_BF16
-
Defines a maximum representable value for the
nv_bfloat16data type. - CUDART_MIN_DENORM_BF16
-
Defines a minimum representable (denormalized) value for the
nv_bfloat16data type. - CUDART_NAN_BF16
-
Defines canonical NaN value for the
nv_bfloat16data type. - CUDART_NEG_ZERO_BF16
-
Defines a negative zero value for the
nv_bfloat16data type. - CUDART_ONE_BF16
-
Defines a value of 1.0 for the
nv_bfloat16data type. - CUDART_ZERO_BF16
-
Defines a positive zero value for the
nv_bfloat16data type.
3.1.1. Macros
-
CUDART_INF_BF16 __ushort_as_bfloat16((unsigned short)0x7F80U)
-
Defines floating-point positive infinity value for the
nv_bfloat16data type.
-
CUDART_MAX_NORMAL_BF16 __ushort_as_bfloat16((unsigned short)0x7F7FU)
-
Defines a maximum representable value for the
nv_bfloat16data type.
-
CUDART_MIN_DENORM_BF16 __ushort_as_bfloat16((unsigned short)0x0001U)
-
Defines a minimum representable (denormalized) value for the
nv_bfloat16data type.
-
CUDART_NAN_BF16 __ushort_as_bfloat16((unsigned short)0x7FFFU)
-
Defines canonical NaN value for the
nv_bfloat16data type.
-
CUDART_NEG_ZERO_BF16 __ushort_as_bfloat16((unsigned short)0x8000U)
-
Defines a negative zero value for the
nv_bfloat16data type.
-
CUDART_ONE_BF16 __ushort_as_bfloat16((unsigned short)0x3F80U)
-
Defines a value of 1.0 for the
nv_bfloat16data type.
-
CUDART_ZERO_BF16 __ushort_as_bfloat16((unsigned short)0x0000U)
-
Defines a positive zero value for the
nv_bfloat16data type.