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_prefetch
def cp_async_bulk_prefetch[src_type: AnyType, /, *, eviction_policy: CacheEviction = CacheEviction.EVICT_NORMAL](src_mem: Pointer[src_type, address_space=src_mem.address_space], size: Int32)
Initiates an asynchronous prefetch from global memory to L2 cache.
Performs a non-blocking prefetch of size bytes from global memory into
the L2 cache. This is a hint to the memory subsystem and does not
guarantee the data will be in cache when accessed.
src_mem must be 16-byte aligned and size must be a multiple of 16.
Requires sm_100 or higher.
Parameters:
- src_type (
AnyType): The element type of the source global memory. - eviction_policy (
CacheEviction): Cache eviction policy for the L2 cache. Defaults toEVICT_NORMAL.
Args:
- src_mem (
Pointer[src_type, address_space=src_mem.address_space]): Source pointer in global or generic memory (16-byte aligned). - size (
Int32): Number of bytes to prefetch (must be a multiple of 16).