VectorTA

Technical analysis library with 340 indicators, built in Rust + CUDA, with Python (PyO3) and WebAssembly bindings (SIMD128; AVX2/AVX-512 where beneficial)

Browse Indicators Try Demo

Implemented

ALMA benchmark snapshot

cargo bench results for VectorTA’s ALMA indicator, comparing scalar, AVX2, and AVX-512 performance on 10k candles, plus AVX-512 versus CUDA on a roughly 250 million-operation batch workload.

ALMA on CPU (10k candles)

VectorTA benchmarks

cargo bench

▊

alma_cpu_10k

ready

ALMA batch: AVX-512 vs CUDA (~250M ops)

VectorTA benchmarks

cargo bench

▊

alma_batch_250m

ready

View Benchmarks Try WASM Optimizer Browse Indicators

Benchmarks shown were run on an AMD 9950X CPU and NVIDIA RTX 4090 GPU.