Built-in Macros
The following table lists several commonly used built-in macros, covering special values of bfloat16, half, and float types, as well as frequently used mathematical constants, logarithmic values, and boundary values. When developers need to handle floating-point boundaries or mathematical constants, they can directly reference and use these built-in macros instead of manually defining or looking up specific numeric values, allowing them to focus more on algorithm logic.
|
Macro Name |
Description |
Header Files to Be Included |
|---|---|---|
|
ASCRT_INF_BF16 |
Defines the positive infinity value of type bfloat16. |
asc_bf16.h |
|
ASCRT_MAX_NORMAL_BF16 |
Defines the maximum representable value of type bfloat16. |
asc_bf16.h |
|
ASCRT_MIN_DENORM_BF16 |
Defines the minimum representable value of type bfloat16. |
asc_bf16.h |
|
ASCRT_NAN_BF16 |
Defines the Not-a-Number (NaN) value of type bfloat16. |
asc_bf16.h |
|
ASCRT_NEG_ZERO_BF16 |
Defines the negative zero value of type bfloat16. |
asc_bf16.h |
|
ASCRT_ONE_BF16 |
Defines the value 1.0 of type bfloat16. |
asc_bf16.h |
|
ASCRT_ZERO_BF16 |
Defines the positive zero value of type bfloat16. |
asc_bf16.h |
|
ASCRT_INF_FP16 |
Defines the positive infinity value of type half. |
asc_fp16.h |
|
ASCRT_MAX_NORMAL_FP16 |
Defines the maximum representable value of type half. |
asc_fp16.h |
|
ASCRT_MIN_DENORM_FP16 |
Defines the minimum representable value of type half. |
asc_fp16.h |
|
ASCRT_NAN_FP16 |
Defines the NaN value of type half. |
asc_fp16.h |
|
ASCRT_NEG_ZERO_FP16 |
Defines the negative zero value of type half. |
asc_fp16.h |
|
ASCRT_ONE_FP16 |
Defines the value 1.0 of type half. |
asc_fp16.h |
|
ASCRT_ZERO_FP16 |
Defines the positive zero value of type half. |
asc_fp16.h |
|
ASCRT_INF_F |
Defines the positive infinity value of type float. |
math_constants.h |
|
ASCRT_NAN_F |
Defines the NaN value of type float. |
math_constants.h |
|
ASCRT_MIN_DENORM_F |
Defines the minimum representable value of type float. |
math_constants.h |
|
ASCRT_MAX_NORMAL_F |
Defines the maximum representable value of type float. |
math_constants.h |
|
ASCRT_NEG_ZERO_F |
Defines the negative zero value of type float. |
math_constants.h |
|
ASCRT_ZERO_F |
Defines the positive zero value of type float. |
math_constants.h |
|
ASCRT_ONE_F |
Defines the value 1.0 of type float. |
math_constants.h |
|
ASCRT_SQRT_HALF_F |
Defines the floating-point value of |
math_constants.h |
|
ASCRT_SQRT_HALF_HI_F |
Defines the high-order part of the floating-point value of |
math_constants.h |
|
ASCRT_SQRT_HALF_LO_F |
Defines the low-order part of the floating-point value of |
math_constants.h |
|
ASCRT_SQRT_TWO_F |
Defines the floating-point value of |
math_constants.h |
|
ASCRT_THIRD_F |
Defines the floating-point value of |
math_constants.h |
|
ASCRT_PIO4_F |
Defines the floating-point value of |
math_constants.h |
|
ASCRT_PIO2_F |
Defines the floating-point value of |
math_constants.h |
|
ASCRT_3PIO4_F |
Defines the floating-point value of |
math_constants.h |
|
ASCRT_2_OVER_PI_F |
Defines the floating-point value of |
math_constants.h |
|
ASCRT_SQRT_2_OVER_PI_F |
Defines the floating-point value of |
math_constants.h |
|
ASCRT_PI_F |
Defines the floating-point value of π. |
math_constants.h |
|
ASCRT_L2E_F |
Defines the floating-point value of log2(e). |
math_constants.h |
|
ASCRT_L2T_F |
Defines the floating-point value of log2(10). |
math_constants.h |
|
ASCRT_LG2_F |
Defines the floating-point value of log10(2). |
math_constants.h |
|
ASCRT_LGE_F |
Defines the floating-point value of log10(e). |
math_constants.h |
|
ASCRT_LN2_F |
Defines the floating-point value of ln(2). |
math_constants.h |
|
ASCRT_LNT_F |
Defines the floating-point value of ln(10). |
math_constants.h |
|
ASCRT_LNPI_F |
Defines the floating-point value of ln(π). |
math_constants.h |
|
ASCRT_TWO_TO_M126_F |
Defines the floating-point value of 2-126. |
math_constants.h |
|
ASCRT_TWO_TO_126_F |
Defines the floating-point value of 2126. |
math_constants.h |
|
ASCRT_NORM_HUGE_F |
Defines the maximum finite value of type float. Any value larger than this overflows to inf. |
math_constants.h |
|
ASCRT_TWO_TO_23_F |
Defines the floating-point value of 223. |
math_constants.h |
|
ASCRT_TWO_TO_24_F |
Defines the floating-point value of 224. |
math_constants.h |
|
ASCRT_TWO_TO_31_F |
Defines the floating-point value of 231. |
math_constants.h |
|
ASCRT_TWO_TO_32_F |
Defines the floating-point value of 232. |
math_constants.h |
|
ASCRT_REMQUO_BITS_F |
Extracts the three least significant bits of the quotient from the remquof function. The value is 3. |
math_constants.h |
|
ASCRT_REMQUO_MASK_F |
Defines a bitmask for extracting the least significant 3 bits of the result returned by remquof. |
math_constants.h |
|
ASCRT_TRIG_PLOSS_F |
105615.0F, which indicates the critical threshold of precision loss in trigonometric function calculation. |
math_constants.h |
