7.5. Notes on Numba Runtime¶
The Numba Runtime (NRT) provides the language runtime to the nopython mode Python subset. NRT is a standalone C library with a Python binding. This allows NPM runtime feature to be used without the GIL. Currently, the only language feature implemented in NRT is memory management.
7.5.1. Memory Management¶
NRT implements memory management for NPM code. It uses atomic reference count for threadsafe, deterministic memory management. NRT maintains a separate MemInfo structure for storing information about each allocation.
18.104.22.168. Cooperating with CPython¶
For NRT to cooperate with CPython, the NRT python binding provides adaptors for converting python objects that export a memory region. When such an object is used as an argument to a NPM function, a new MemInfo is created and it acquires a reference to the Python object. When a NPM value is returned to the Python interpreter, the associated MemInfo (if any) is checked. If the MemInfo references a Python object, the underlying Python object is released and returned instead. Otherwise, the MemInfo is wrapped in a Python object and returned. Additional process maybe required depending on the type.
The current implementation supports Numpy array and any buffer-exporting types.
22.214.171.124. Compiler-side Cooperation¶
NRT reference counting requires the compiler to emit incref/decref operations according to the usage. When the reference count drops to zero, the compiler must call the destructor routine in NRT.
The compiler is allowed to emit incref/decref operations naively. It relies on an optimization pass that to remove the redundant reference count operations.
The optimization pass runs on block level to avoid control flow analysis. It depends on LLVM function optimization pass to simplify the control flow, stack-to-register, and simplify instructions. It works by matching and removing incref and decref pairs within each block.
All NRT routines currently initializes refcount to 0, because the compiler relies on variable binding for doing incref/decref. Every value is bound to a variable in the Numba IR.
Since the refcount optimization pass requires LLVM function optimization pass, the pass works on the LLVM IR as text. The optimized IR is then materialized again as a new LLVM in-memory bitcode object.
126.96.36.199. Debugging Leaks¶
To debug reference leaks in NRT MemInfo, each MemInfo python object has a .refcount attribute for inspection. To get the MemInfo from a ndarray allocated by NRT, use the .base attribute.
To debug memory leaks in NRT, the numba.runtime.rtsys defines .get_allocation_stats(). It returns a namedtuple containing the number of allocation and deallocation since the start of the program. Checking that the allocation and deallocation counters are matching is the simplest way to know if the NRT is leaking.
7.5.2. Future Plan¶
The plan for NRT is to make a standalone shared library that can be linked to Numba compiled code, including use within the Python interpreter and without the Python interpreter. To make that work, we will be doing some refactoring:
- numba NPM code references statically compiled code in “helperlib.c”. Those functions should be moved to NRT.