DMCA
Parallelizing irregular C codes assisted by interprocedural shape analysis (2008)
| Venue: | In IPDPS |
| Citations: | 1 - 0 self |
BibTeX
@INPROCEEDINGS{Asenjo08parallelizingirregular,
author = {R. Asenjo and R. Castillo and F. Corbera and A. Navarro and A. Tineo and E. L. Zapata},
title = {Parallelizing irregular C codes assisted by interprocedural shape analysis},
booktitle = {In IPDPS},
year = {2008},
pages = {1--12},
publisher = {IEEE Computer Society}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
In the new multicore architecture arena, the problem of improving the performance of a code is more in the soft-ware side than in the hardware one. However, optimizing irregular dynamic data structure based codes for such ar-chitectures is not easy, either by hand or compiler assisted. Regarding this last approach, shape analysis is a static tech-nique that achieves abstraction of dynamic memory and can help to disambiguate, quite accurately, memory references in programs that create and traverse recursive data struc-tures. This kind of analysis has promising applicability for accurate data dependence tests in loops or recursive func-tions that traverse dynamic data structures. However, sup-port for interprocedural programs in shape analysis is still a challenge, especially in the presence of recursive func-tions. In this work we present a novel fully context-sensitive interprocedural shape analysis algorithm that supports re-cursion and can be used to uncover parallelism. Our ap-proach is based on three key ideas: i) intraprocedural sup-port based on “Coexistent Links Sets ” to precisely describe the memory configurations during the abstract interpreta-tion of the C code; ii) interprocedural support based on “Recursive Flow Links ” to trace the state of pointers in previous calls; and iii) annotations of the read/written heap locations during the program analysis. We present prelim-inary experiments that reveal that our technique compares favorably with related work, and obtains precise memory abstractions in a variety of recursive programs that create and manipulate dynamic data structures. We have also im-plemented a data dependence test over our interprocedural shape analysis. With this test we have obtained promis-ing results, automatically detecting parallelism in three C codes, which have been successfully parallelized.
Keyphrases
interprocedural shape analysis irregular code shape analysis dynamic data structure recursive func-tions interprocedural program memory configuration precise memory abstraction recursive data struc-tures accurate data dependence test context-sensitive interprocedural shape analysis algorithm memory reference related work recursive program static tech-nique irregular dynamic data structure promis-ing result last approach interprocedural support new multicore architecture arena abstract interpreta-tion previous call program analysis key idea intraprocedural sup-port data dependence test heap location dynamic memory present prelim-inary experiment soft-ware side promising applicability recursive flow link coexistent link set
