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 package

memory

GPU memory operations package.

This package provides GPU memory operations including:

• async_copy: Asynchronous memory copy operations
• cp_async_bulk: Bulk tensor memory copy operations (TMA)
• load/store: Memory access operations with cache control
• fence operations: Memory ordering and synchronization
• mbarrier: Memory barrier operations
• edge-masked loads: Apple M5 (AGX3) hardware bounds-checked vector loads

These operations enable efficient data movement between different GPU memory spaces and coordination of memory operations across threads.

Modules

• ​masked_load_apple: Apple M5 (AGX3) hardware edge-masked vector loads.
• ​memory: This module provides GPU memory operations and utilities.