For the complete documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /get-started.md).

Mojo function

cp_async_bulk_tensor_global_shared_cta_elect

def cp_async_bulk_tensor_global_shared_cta_elect[src_type: AnyType, rank: Int, /, eviction_policy: CacheEviction = CacheEviction.EVICT_NORMAL](src_mem: Pointer[src_type, address_space=AddressSpace.SHARED], tma_descriptor: Pointer[NoneType], coords: IndexList[rank], elect: Int32)

Elect-predicated variant of cp_async_bulk_tensor_global_shared_cta.

Behaves exactly like cp_async_bulk_tensor_global_shared_cta except that the TMA store is guarded by a PTX predicate derived from elect: the instruction is issued only when elect != 0. All lanes follow the same PTX control flow, so there is no Mojo-level branch and no warp-divergent if elect != 0: wrapper at the call site.

Parameters:

• ​src_type (AnyType): The data type of the source tensor elements.
• ​rank (Int): The dimensionality of the tensor (1, 2, 3, 4, or 5).
• ​eviction_policy (CacheEviction): Optional cache eviction policy. Defaults to EVICT_NORMAL.

Args:

• ​src_mem (Pointer[src_type, address_space=AddressSpace.SHARED]): Pointer to the source data in shared memory.
• ​tma_descriptor (Pointer[NoneType]): Pointer to the TMA descriptor.
• ​coords (IndexList[rank]): Coordinates specifying which tile to copy.
• ​elect (Int32): 0 on non-elected lanes (skip the TMA), non-zero on the single elected lane (issue the TMA). Typically the Int32 returned by elect() from nn.attention.gpu.nvidia.common.