Building TritonX: How I Got 1200x Speedups over Python Matrix Loops using Rust & C-ABI

I built TritonX, an open-source Rust-powered matrix compute engine designed to offload heavy calculations from Python using Rayon parallel worker pools and raw C-ABI bindings.

Benchmarks & Performance

Running local benchmarks via .\run_tritonx.ps1 shows up to ~1200x speedups over baseline Python matrix loops by bypassing GIL bottlenecks and utilizing lock-free worker pools.

Source Code

Check out the full repository, benchmarks, and build setup:
https://github.com/pulisherij-rgb/TritonX

I would love feedback on FFI boundary overhead, raw pointers, and memory layout optimization when bridging multithreaded Rust pools back into Python!