thread_block Class Reference#
cooperative_groups::thread_block Class Reference
The workgroup (thread-block in CUDA terminology) cooperative group type. More...
#include <amd_hip_cooperative_groups.h>
Inheritance diagram for cooperative_groups::thread_block:
Collaboration diagram for cooperative_groups::thread_block:
Public Member Functions | |
| dim3 | group_dim () |
| Returns the group dimensions. | |
 Public Member Functions inherited from cooperative_groups::thread_group | |
| uint32_t | size () const |
| unsigned int | cg_type () const |
| Returns the type of the group. | |
| uint32_t | thread_rank () const |
| Rank of the calling thread within [0, size() ). | |
| bool | is_valid () const |
| Returns true if the group has not violated any API constraints. | |
| void | sync () const |
| Synchronizes the threads in the group. | |
Static Public Member Functions | |
| static dim3 | group_index () |
| Returns 3-dimensional block index within the grid. | |
| static dim3 | thread_index () |
| Returns 3-dimensional thread index within the block. | |
| static uint32_t | thread_rank () |
| Rank of the calling thread within [0, size() ). | |
| static uint32_t | size () |
| static bool | is_valid () |
| Returns true if the group has not violated any API constraints. | |
| static void | sync () |
| Synchronizes the threads in the group. | |
Protected Member Functions | |
| thread_block (uint32_t size) | |
| thread_group | new_tiled_group (unsigned int tile_size) const |
| Protected Member Functions inherited from cooperative_groups::thread_group | |
| thread_group (internal::group_type type, uint32_t size=static_cast< uint64_t >(0), uint64_t mask=static_cast< uint64_t >(0)) | |
Friends | |
| thread_block | this_thread_block () |
User-exposed API interface to construct workgroup cooperative group type object - thread_block. | |
| thread_group | tiled_partition (const thread_group &parent, unsigned int tile_size) |
| User-exposed API to partition groups. | |
| thread_group | tiled_partition (const thread_block &parent, unsigned int tile_size) |
Additional Inherited Members | |
| Protected Attributes inherited from cooperative_groups::thread_group | |
| uint32_t | _type |
| uint32_t | _size |
| Type of the thread_group. | |
| uint64_t | _mask |
| Total number of threads in the tread_group. | |
| struct cooperative_groups::thread_group::_coalesced_info | coalesced_info |
Detailed Description
The workgroup (thread-block in CUDA terminology) cooperative group type.
Represents an intra-workgroup cooperative group type, where the participating threads within the group are the same threads that participated in the currently executing workgroup.
- Note
- This function is implemented on Linux and is under development on Microsoft Windows.
The documentation for this class was generated from the following file:
- /home/docs/checkouts/readthedocs.org/user_builds/advanced-micro-devices-hip/checkouts/clr/hipamd/include/hip/amd_detail/amd_hip_cooperative_groups.h