default_memmgr

cubie.memory.default_memmgr = MemoryManager(totalmem=8589934592, registry={}, stream_groups=StreamGroups(groups={}, streams={}), _mode='passive', _allocator=<class 'cubie.cuda_simsafe.FakeNumbaCUDAMemoryManager'>, _auto_pool=[], _manual_pool=[], _queued_allocations={})

Singleton interface coordinating GPU memory allocation and stream usage.

Parameters:

  • totalmem – Total GPU memory in bytes. Determined automatically when omitted.

  • registry – Registry mapping instance identifiers to their memory settings.

  • stream_groups – Manager for organizing instances into stream groups.

  • _mode – Memory management mode, either "passive" or "active".

  • _allocator – Memory allocator class registered with Numba.

  • _auto_pool – List of instance identifiers using automatic memory allocation.

  • _manual_pool – List of instance identifiers using manual memory allocation.

  • _queued_allocations – Queued allocation requests organized by stream group.

Notes

The manager accepts ArrayRequest objects and returns ArrayResponse instances that reference allocated arrays and chunking information. Active mode enforces per-instance VRAM proportions while passive mode mirrors standard allocation behaviour using chunking only when necessary.