WindFlow | A C17 Data Stream Processing Parallel Library | GPU library

WindFlow is a C++ library typically used in Hardware, GPU applications. WindFlow has no bugs, it has no vulnerabilities, it has a Weak Copyleft License and it has low support. You can download it from GitHub.

WindFlow is a C++17 library for parallel data stream processing targeting heterogeneous shared-memory architectures equipped with multi-core CPUs and an NVIDIA GPUs. The library provides traditional stream processing operators like map, flatmap, filter, fold/reduce as well as sliding-window operators designed with complex parallel processing modes. The API allows building streaming applications through the MultiPipe and the PipeGraph programming constructs. The first is used to create parallel pipelines, while the second one allows several MultiPipe instances to be interconnected through merge and split operations, thus creating complex directed acyclic graphs of interconnected operators. WindFlow is not thought to support streaming analytics applications only (e.g., the ones written with relational algebra query languages, as in traditional old-style DSMSs) but rather all general-purpose streaming applications can be easily supported through operators embedding user-defined custom logics. In terms of runtime system, WindFlow is particularly suitable for embedded architectures equipped with low-power multi-core CPUs and integrated NVIDIA GPUs (i.e., Jetson boards). However, it works well also on traditional multi-core servers equipped with discrete NVIDIA GPUs (e.g., Pascal/Volta models). Differently from existing research libraries for stream processing on multicores, WindFlow is thought to support real live-streaming applications, where inputs are continuously received from real-world sources, and not only offline streaming applications reading historical data already prepared in memory. The web site of the library is available at: