For Full-Text PDF, please login, if you are a member of IEICE,|
or go to Pay Per View on menu list, if you are a nonmember of IEICE.
Relaxing Constraints due to Data and Control Dependences
Katsuhiko METSUGI Kazuaki MURAKAMI
IEICE TRANSACTIONS on Information and Systems
Publication Date: 2003/05/01
Print ISSN: 0916-8532
Type of Manuscript: PAPER
Category: Computer Systems
TLSP, ILP, data dependence, control dependence, value prediction,
Full Text: PDF(391.1KB)>>
TLSP (Thread-Level Speculative Parallel processing) architecture is a growing processor architecture. Parallelism of a program executed on this architecture is ruled by the combination of techniques which relax data dependences. In this paper, we evaluate the limits of parallelism of the TLSP architecture by using abstract machine models. We have three major results. First, if we use solely each technique which relaxes data dependences, "renaming" has a large effect on the TLSP architecture. Second, combinatorial use of "memory disambiguation" and "renaming" leads to huge parallelism. Third, constant effects are obtained by concurrent use of "value prediction" and other techniques.