numba-cache-lock 0.1.0a1

A caching system for Numba with file locking support

numba-lock-cache

Intro

numba-lock-cache monkey-patch the Numba caching mechanism smartjit @jit decorator adds extra customization of when code execution should fall back to the interpreter. It works as follow:

1. For jitted functions with cache (overloads), use the jitted function if available, and interpreted code otherwise
2. Add a dispatching logic, an optional function to pass to the jit decorator, which will decide wether to use jit or not.

Install

numba-smartjit is available on PyPI and can be installed with the command below:

pip install numba-smartjit

numba-locking-cache

A Python package that monkey-patches Numba's caching mechanism to safely coordinate concurrent cache access using file locks.

Why?

Numba’s function-level caching (@jit(cache=True)) is not concurrency-safe by default. This can lead to:

• Crashes when multiple processes load/write to the same cache
• Especially problematic on shared filesystems (e.g., NFS)

Locking Mechanism

• Uses flufl.lock for file-based locking.
• Lock file is created next to the cache index file: /path/to/cache/func.nbi.lock
• Lock behavior:
  • Timeout to acquire: 60 seconds (configurable)
  • Lifetime: None (lock persists until released)
  • NFS-safe: relies on atomic file creation

Installation

pip install numba-locking-cache

How to use it

Just import the patch Numba in your application:

import numba_locking_cache

numba_locking_cache.patch_numba_cache()

from numba import jit

@jit(cache=True)
def my_func(x):
    return x * 2

By default, patch_numba_cache will hold a lock for 1 hour in the worst case. One can change this value by assigning the lifetime= keyword argument to a timedelta object.