sigtekx 0.0.2

GPU-accelerated signal processing pipeline builder for real-time FFT analysis

SigTekX

Package name reservation - Full release coming soon

About

SigTekX (Signal + Tekton "builder" + X "acceleration") is a high-performance GPU-accelerated signal processing library for building real-time FFT analysis pipelines.

This placeholder reserves the sigtekx package name on PyPI. The library is currently in development (v0.9.4) and will be published under this name in an upcoming release.

Current Status

The library is in active development (Beta) with a working implementation featuring:

• Sub-200μs latency CUDA FFT engine
• Python API with C++ backend via pybind11
• Multi-stream async execution for throughput optimization
• Research infrastructure (Hydra, MLflow, DVC, Snakemake)
• Professional benchmarking suite with statistical analysis
• NVTX profiling support for NVIDIA tools

Coming Soon

The rebranded sigtekx package will include:

• Complete CUDA-accelerated FFT processing engine
• Type-safe Python API with Pydantic configuration
• Batch and streaming execution modes
• Comprehensive benchmark and profiling tools
• Pre-configured presets for common use cases
• Full documentation and examples

Repository

Development is happening at: https://github.com/SEAL-Embedded/ionosense-hpc-lib

Note: Repository will be renamed to sigtekx upon package release.

Requirements (Future Release)

• Python 3.11+
• NVIDIA GPU with CUDA compute capability 6.0+ (Pascal or newer)
• CUDA Toolkit 13.0+
• Windows 11 or Linux

Contact

Author: Kevin Rahsaz Email: [e-mail]

License

MIT License