Load#

group thread_load

Enums

enum cache_load_modifier#

These enum values are used to specify caching behaviour on load.

Values:

enumerator load_default#

Default (no modifier)

enumerator load_ca#

Cache at all levels.

enumerator load_cg#

Cache at global level.

enumerator load_cs#

Cache streaming (likely to be accessed once)

enumerator load_cv#

Cache as volatile (including cached system lines)

enumerator load_ldg#

Cache as texture.

enumerator load_volatile#

Volatile (any memory space)

Functions

template<cache_load_modifier MODIFIER = load_default, typename InputIteratorT>
__device__ inline std::iterator_traits<InputIteratorT>::value_type thread_load(InputIteratorT itr)#

Store data using the default load instruction. No support for cache modified stores yet.

Template Parameters:
  • MODIFIER – Value in enum for determine which type of cache store modifier to be used

  • InputIteratorT – Type of Output Iterator

Parameters:

itr – [in] Iterator to location where data is to be stored

Returns:

Data that is loaded from memory

template<cache_load_modifier MODIFIER = load_default, typename T>
__device__ inline T thread_load(T *ptr)#

Load data using the default load instruction. No support for cache modified loads yet.

Template Parameters:
  • MODIFIER – Value in enum for determine which type of cache store modifier to be used

  • T – Type of Data to be loaded

Parameters:

ptr – [in] - Pointer to data to be loaded

Returns:

Data that is loaded from memory