Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

Instruction Selection and Scheduling for DSP Kernels

Arslan, Mehmet Ali LU and Kuchcinski, Krzysztof LU orcid (2014) In Microprocessors and Microsystems 38(8). p.803-813
Abstract
As custom multicore architectures become more and more common for DSP applications, instruction selection and scheduling for such applications and architectures become important topics. In this paper, we explore the effects of defining the problem of finding an optimal instruction selection and scheduling as a constraint satisfaction problem (CSP). We incorporate methods based on sub-graph isomorphism and global constraints designed for scheduling. We experiment using several media applications on a custom architecture, a generic VLIW architecture and a RISC architecture, all three with several cores. Our results show that defining the problem with constraints gives flexibility in modelling, while state-of-the-art constraint solvers enable... (More)
As custom multicore architectures become more and more common for DSP applications, instruction selection and scheduling for such applications and architectures become important topics. In this paper, we explore the effects of defining the problem of finding an optimal instruction selection and scheduling as a constraint satisfaction problem (CSP). We incorporate methods based on sub-graph isomorphism and global constraints designed for scheduling. We experiment using several media applications on a custom architecture, a generic VLIW architecture and a RISC architecture, all three with several cores. Our results show that defining the problem with constraints gives flexibility in modelling, while state-of-the-art constraint solvers enable optimal solutions for large problems, hinting a new method for code generation. (Less)
Please use this url to cite or link to this publication:
author
and
organization
publishing date
type
Contribution to journal
publication status
published
subject
keywords
Instruction selection, Scheduling, Custom architecture, VLIW, Constraint programming
in
Microprocessors and Microsystems
volume
38
issue
8
pages
803 - 813
publisher
Elsevier
external identifiers
  • wos:000347755200007
  • scopus:84910608322
ISSN
0141-9331
DOI
10.1016/j.micpro.2014.04.004
project
High Performance Embedded Computing
language
English
LU publication?
yes
id
facbb018-7866-4c40-b9d4-585805292431 (old id 4586660)
alternative location
http://dx.doi.org/10.1016/j.micpro.2014.04.004
date added to LUP
2016-04-01 10:35:51
date last changed
2022-01-26 00:45:11
@article{facbb018-7866-4c40-b9d4-585805292431,
  abstract     = {{As custom multicore architectures become more and more common for DSP applications, instruction selection and scheduling for such applications and architectures become important topics. In this paper, we explore the effects of defining the problem of finding an optimal instruction selection and scheduling as a constraint satisfaction problem (CSP). We incorporate methods based on sub-graph isomorphism and global constraints designed for scheduling. We experiment using several media applications on a custom architecture, a generic VLIW architecture and a RISC architecture, all three with several cores. Our results show that defining the problem with constraints gives flexibility in modelling, while state-of-the-art constraint solvers enable optimal solutions for large problems, hinting a new method for code generation.}},
  author       = {{Arslan, Mehmet Ali and Kuchcinski, Krzysztof}},
  issn         = {{0141-9331}},
  keywords     = {{Instruction selection; Scheduling; Custom architecture; VLIW; Constraint programming}},
  language     = {{eng}},
  number       = {{8}},
  pages        = {{803--813}},
  publisher    = {{Elsevier}},
  series       = {{Microprocessors and Microsystems}},
  title        = {{Instruction Selection and Scheduling for DSP Kernels}},
  url          = {{http://dx.doi.org/10.1016/j.micpro.2014.04.004}},
  doi          = {{10.1016/j.micpro.2014.04.004}},
  volume       = {{38}},
  year         = {{2014}},
}