1.2.2. Half2 Arithmetic Functions

[Half Precision Intrinsics]

To use these functions, include the header file cuda_fp16.h in your program.

Functions

__device__ ​ __half2 __h2div ( const __half2 a, const __half2 b )
Performs half2 vector division in round-to-nearest-even mode.
__device__ ​ __half2 __habs2 ( const __half2 a )
Calculates the absolute value of both halves of the input half2 number and returns the result.
__device__ ​ __half2 __hadd2 ( const __half2 a, const __half2 b )
Performs half2 vector addition in round-to-nearest-even mode.
__device__ ​ __half2 __hadd2_rn ( const __half2 a, const __half2 b )
Performs half2 vector addition in round-to-nearest-even mode.
__device__ ​ __half2 __hadd2_sat ( const __half2 a, const __half2 b )
Performs half2 vector addition in round-to-nearest-even mode, with saturation to [0.0, 1.0].
__device__ ​ __half2 __hcmadd ( const __half2 a, const __half2 b, const __half2 c )
Performs fast complex multiply-accumulate.
__device__ ​ __half2 __hfma2 ( const __half2 a, const __half2 b, const __half2 c )
Performs half2 vector fused multiply-add in round-to-nearest-even mode.
__device__ ​ __half2 __hfma2_relu ( const __half2 a, const __half2 b, const __half2 c )
Performs half2 vector fused multiply-add in round-to-nearest-even mode with relu saturation.
__device__ ​ __half2 __hfma2_sat ( const __half2 a, const __half2 b, const __half2 c )
Performs half2 vector fused multiply-add in round-to-nearest-even mode, with saturation to [0.0, 1.0].
__device__ ​ __half2 __hmul2 ( const __half2 a, const __half2 b )
Performs half2 vector multiplication in round-to-nearest-even mode.
__device__ ​ __half2 __hmul2_rn ( const __half2 a, const __half2 b )
Performs half2 vector multiplication in round-to-nearest-even mode.
__device__ ​ __half2 __hmul2_sat ( const __half2 a, const __half2 b )
Performs half2 vector multiplication in round-to-nearest-even mode, with saturation to [0.0, 1.0].
__device__ ​ __half2 __hneg2 ( const __half2 a )
Negates both halves of the input half2 number and returns the result.
__device__ ​ __half2 __hsub2 ( const __half2 a, const __half2 b )
Performs half2 vector subtraction in round-to-nearest-even mode.
__device__ ​ __half2 __hsub2_rn ( const __half2 a, const __half2 b )
Performs half2 vector subtraction in round-to-nearest-even mode.
__device__ ​ __half2 __hsub2_sat ( const __half2 a, const __half2 b )
Performs half2 vector subtraction in round-to-nearest-even mode, with saturation to [0.0, 1.0].
__device__ ​ __half2 atomicAdd ( const __half2* address, const __half2 val )
Vector add val to the value stored at address in global or shared memory, and writes this value back to address. The atomicity of the add operation is guaranteed separately for each of the two __half elements; the entire __half2 is not guaranteed to be atomic as a single 32-bit access.

Functions

__device__ ​ __half2 __h2div ( const __half2 a, const __half2 b )
Performs half2 vector division in round-to-nearest-even mode.
Description

Divides half2 input vector a by input vector b in round-to-nearest mode.

__device__ ​ __half2 __habs2 ( const __half2 a )
Calculates the absolute value of both halves of the input half2 number and returns the result.
Parameters
a
- half2. Is only being read.
Returns

half2

  • Returns a with the absolute value of both halves.

Description

Calculates the absolute value of both halves of the input half2 number and returns the result.

__device__ ​ __half2 __hadd2 ( const __half2 a, const __half2 b )
Performs half2 vector addition in round-to-nearest-even mode.
Description

Performs half2 vector add of inputs a and b, in round-to-nearest mode.

__device__ ​ __half2 __hadd2_rn ( const __half2 a, const __half2 b )
Performs half2 vector addition in round-to-nearest-even mode.
Description

Performs half2 vector add of inputs a and b, in round-to-nearest mode. Prevents floating-point contractions of mul+add into fma.

__device__ ​ __half2 __hadd2_sat ( const __half2 a, const __half2 b )
Performs half2 vector addition in round-to-nearest-even mode, with saturation to [0.0, 1.0].
Parameters
a
- half2. Is only being read.
b
- half2. Is only being read.
Returns

half2

  • The sum of a and b, with respect to saturation.

Description

Performs half2 vector add of inputs a and b, in round-to-nearest mode, and clamps the results to range [0.0, 1.0]. NaN results are flushed to +0.0.

__device__ ​ __half2 __hcmadd ( const __half2 a, const __half2 b, const __half2 c )
Performs fast complex multiply-accumulate.
Parameters
a
- half2. Is only being read.
b
- half2. Is only being read.
c
- half2. Is only being read.
Returns

half2

  • The result of complex multiply-accumulate operation on complex numbers a, b, and c

Description

Interprets vector half2 input pairs a, b, and c as complex numbers in half precision and performs complex multiply-accumulate operation: a*b + c

__device__ ​ __half2 __hfma2 ( const __half2 a, const __half2 b, const __half2 c )
Performs half2 vector fused multiply-add in round-to-nearest-even mode.
Description

Performs half2 vector multiply on inputs a and b, then performs a half2 vector add of the result with c, rounding the result once in round-to-nearest-even mode.

__device__ ​ __half2 __hfma2_relu ( const __half2 a, const __half2 b, const __half2 c )
Performs half2 vector fused multiply-add in round-to-nearest-even mode with relu saturation.
Parameters
a
- half2. Is only being read.
b
- half2. Is only being read.
c
- half2. Is only being read.
Returns

half2

  • The result of elementwise fused multiply-add operation on vectors a, b, and c with relu saturation.

Description

Performs half2 vector multiply on inputs a and b, then performs a half2 vector add of the result with c, rounding the result once in round-to-nearest-even mode. Then negative result is clamped to 0. NaN result is converted to canonical NaN.

__device__ ​ __half2 __hfma2_sat ( const __half2 a, const __half2 b, const __half2 c )
Performs half2 vector fused multiply-add in round-to-nearest-even mode, with saturation to [0.0, 1.0].
Parameters
a
- half2. Is only being read.
b
- half2. Is only being read.
c
- half2. Is only being read.
Returns

half2

  • The result of elementwise fused multiply-add operation on vectors a, b, and c, with respect to saturation.

Description

Performs half2 vector multiply on inputs a and b, then performs a half2 vector add of the result with c, rounding the result once in round-to-nearest-even mode, and clamps the results to range [0.0, 1.0]. NaN results are flushed to +0.0.

__device__ ​ __half2 __hmul2 ( const __half2 a, const __half2 b )
Performs half2 vector multiplication in round-to-nearest-even mode.
Description

Performs half2 vector multiplication of inputs a and b, in round-to-nearest-even mode.

__device__ ​ __half2 __hmul2_rn ( const __half2 a, const __half2 b )
Performs half2 vector multiplication in round-to-nearest-even mode.
Description

Performs half2 vector multiplication of inputs a and b, in round-to-nearest-even mode. Prevents floating-point contractions of mul+add or sub into fma.

__device__ ​ __half2 __hmul2_sat ( const __half2 a, const __half2 b )
Performs half2 vector multiplication in round-to-nearest-even mode, with saturation to [0.0, 1.0].
Parameters
a
- half2. Is only being read.
b
- half2. Is only being read.
Returns

half2

  • The result of elementwise multiplication of vectors a and b, with respect to saturation.

Description

Performs half2 vector multiplication of inputs a and b, in round-to-nearest-even mode, and clamps the results to range [0.0, 1.0]. NaN results are flushed to +0.0.

__device__ ​ __half2 __hneg2 ( const __half2 a )
Negates both halves of the input half2 number and returns the result.
Description

Negates both halves of the input half2 number a and returns the result.

__device__ ​ __half2 __hsub2 ( const __half2 a, const __half2 b )
Performs half2 vector subtraction in round-to-nearest-even mode.
Description

Subtracts half2 input vector b from input vector a in round-to-nearest-even mode.

__device__ ​ __half2 __hsub2_rn ( const __half2 a, const __half2 b )
Performs half2 vector subtraction in round-to-nearest-even mode.
Description

Subtracts half2 input vector b from input vector a in round-to-nearest-even mode. Prevents floating-point contractions of mul+sub into fma.

__device__ ​ __half2 __hsub2_sat ( const __half2 a, const __half2 b )
Performs half2 vector subtraction in round-to-nearest-even mode, with saturation to [0.0, 1.0].
Parameters
a
- half2. Is only being read.
b
- half2. Is only being read.
Returns

half2

  • The subtraction of vector b from a, with respect to saturation.

Description

Subtracts half2 input vector b from input vector a in round-to-nearest-even mode, and clamps the results to range [0.0, 1.0]. NaN results are flushed to +0.0.

__device__ ​ __half2 atomicAdd ( const __half2* address, const __half2 val )
Vector add val to the value stored at address in global or shared memory, and writes this value back to address. The atomicity of the add operation is guaranteed separately for each of the two __half elements; the entire __half2 is not guaranteed to be atomic as a single 32-bit access.
Parameters
address
- half2*. An address in global or shared memory.
val
- half2. The value to be added.
Returns

half2

  • The old value read from address.

Description

The location of address must be in global or shared memory. This operation has undefined behavior otherwise. This operation is only supported by devices of compute capability 6.x and higher.

Note:

For more details for this function see the Atomic Functions section in the CUDA C++ Programming Guide.