Copyrights to these papers may be held by the publishers. The download files are preprints. It is understood that all persons copying this information will adhere to the terms and constraints invoked by each author's copyright. These works may not be reposted without the explicit permission of the copyright holder.
Franz Franchetti, Yevgen Voronenko, Peter A. Milder, Srinivas Chellappa, Marek Telgarsky, Hao Shen, Paolo D'Alberto, Frédéric de Mesmay, James C. Hoe, José M. F. Moura and Markus Püschel (Proc. NSF Next Generation Software Program Workshop (NSFNGS) colocated with IPDPS, 2008)
Domain-Specific Library Generation for Parallel Software and Hardware Platforms
Preprint (180 KB)
Published paper (link to publisher)
We overview a library generation framework called Spiral. For the domain of linear transforms, Spiral automatically generates implementations for parallel platforms including SIMD vector extensions, multicore processors, field-programmable gate arrays (FPGAs) and FPGA accelerated processors. The performance of the generated code is competitive with the best available hand-written libraries.Keywords: Learn the current Spiral system, Multithreading, SIMD vectorization, Walsh-Hadamard transform, SPIRAL program generation system for transforms, IP cores for FPGA/ASIC, Spiral overview paper, Discrete/fast Fourier transform, Cell BE Processor