atomics.cuh File Reference

Defines an atomic add function for double precision floating point types for older CUDA architectures. More...

Functions
__device__ __forceinline__ double	atomicAdd (double *addr, const double val)
	Atomically add the double precision val to the value denoted by addr. More...

Detailed Description

Defines an atomic add function for double precision floating point types for older CUDA architectures.

Author

Alexander Van Craen

Marcel Breyer

Copyright

2018-today The PLSSVM project - All Rights Reserved

License

This file is part of the PLSSVM project which is released under the MIT license. See the LICENSE.md file in the project root for full license information.

Function Documentation

atomicAdd()

__device__ __forceinline__ double atomicAdd	(	double *	addr,
		const double	val
	)

Atomically add the double precision val to the value denoted by addr.

Needed since CUDA versions before 6 don't nativelly support double-precisions floating point atomics.

Parameters

[in,out]	addr	the value to increment
[in]	val	the value to add

Returns

the old value before atomically adding val