Dependence compensation for sparse computations
作者： Anthony Roide , Antonio Valles , Dattatraya Kulkarni , Gautam Doshi
DOI:
关键词:
摘要: An embodiment of a compiler technique for decreasing sparse matrix computation runtime parallelizes loads from adjacent iterations unrolled loop code. A dependence check code is statically inserted to identify between store and load dynamically, information passed scheduler scheduling independent parallel potentially dependent computations at suitable latencies.
lens.org 参考文章(8)
Francky Catthoor, Koen Danckaert, Loop optimization with mapping code on an architecture ,(2000)
William C. Moyer, Lea Hwang Lee, Method and apparatus for instruction execution in a data processing system ,(2000)
Uma Srinivasan, Carol L. Thompson, Richard E. Hank, Dale Morris, Method and apparatus for efficient code generation for modulo scheduled uncounted loops ,(2002)
Robert Y. Geva, Method of constructing and unrolling speculatively counted loops ,(1999)
David Sehr, Daniel Lavery, Rakesh Ghiya, Method for collection of memory reference information and memory disambiguation ,(2001)
Dattatraya Kulkarni, Kalyan Muthukumar, Gautam Doshi, System and method for software-pipelining of loops with sparse matrix routines ,(2003)
Uma Srinivasan, Carol Thompson, System and method for improved register allocation in an optimizing compiler ,(2002)
Michael S. Schlansker, Darren C. Cronquist, System and method for creating systolic solvers ,(2003)