StreamIt Compiler Release - Saman Amarasinghe [saman@mit.edu]

We are pleased to announce the second public release of the StreamIt compilation infrastructure. StreamIt is a programming language and compiler for streaming systems, encompassing audio, video, DSP, networking, and cryptographic processing kernels. By naturally exposing the parallelism in these applications, StreamIt provides an efficient mapping to the emerging class of multicore architectures. To download StreamIt, or for more information, please visit: http://cag.csail.mit.edu/streamit/ Highlights of StreamIt 2.1 include: - An optimizing compiler for uniprocessors, multicore architectures and clusters of workstations. The compiler relies on a new C++ runtime library with improved uniprocessor performance (up to 2X). - Language constructs that improve programmer productivity for streaming, including hierarchical structured streams, graph parameterization, and circular buffer management. StreamIt 2.1 further includes dynamic I/O rates and "teleport messaging", a language construct for control messages between parallel components. - A set of benchmarks for the evaluation of streaming optimizations and architectures, including beamforming, bitonic sort, DCT, DES, FFT, filterbanks, FM software radio, JPEG encoding/decoding, MP3 decoding, MPEG-2 encoding/decoding, radar array front end, synthetic aperture radar, and Serpent. Some applications (JPEG, MPEG-2) have limited compiler support, but all run in our Java library. - A snapshot of our latest research optimizations, including cache optimizations for embedded processors, and domain-specific optimizations targeting linear and linear statespace filters. - A high-level stream IR (in Java) with a host of graph transformations, including graph canonicalization, synchronization removal, refactoring of parallel components, fusion of adjacent filters (running in parallel or in sequence), and fission of individual filters (into data-parallel or task-parallel components.) We are releasing StreamIt for both the user community and the developer community. We look forward to interacting with you as you might start to use or build upon the infrastructure. The StreamIt Team