Profile-guided meta-programming

William J. Bowman; Swaha Miller; Vincent St Amour; R. Kent Dybvig

doi:10.1145/2737924.2737990

Profile-guided meta-programming

William J. Bowman, Swaha Miller, Vincent St Amour, R. Kent Dybvig

Computer Science

Research output: Contribution to journal › Article › peer-review

6 Scopus citations

Abstract

Contemporary compiler systems such as GCC,.NET, and LLVM incorporate profile-guided optimizations (PGOs) on low-level intermediate code and basic blocks, with impressive results over purely static heuristics. Recent work shows that profile information is also useful for performing source-to-source optimizations via meta-programming. For example, using profiling information to inform decisions about data structures and algorithms can potentially lead to asymptotic improvements in performance. We present a design for profile-guided meta-programming in a general-purpose meta-programming system. Our design is parametric over the particular profiler and meta-programming system. We implement this design in two different meta-programming systems-the syntactic extensions systems of Chez Scheme and Racket-and provide several profile-guided meta-programs as usability case studies.

Original language	English (US)
Pages (from-to)	403-412
Number of pages	10
Journal	ACM SIGPLAN Notices
Volume	50
Issue number	6
DOIs	https://doi.org/10.1145/2737924.2737990
State	Published - Jun 2015

Keywords

Meta-programming
Optimization
PGO
Profile-guided optimization
Profiling

ASJC Scopus subject areas

General Computer Science

Access to Document

10.1145/2737924.2737990

Cite this

@article{581442ecd2b44f8bbb6dc08000fe3fd3,

title = "Profile-guided meta-programming",

abstract = "Contemporary compiler systems such as GCC,.NET, and LLVM incorporate profile-guided optimizations (PGOs) on low-level intermediate code and basic blocks, with impressive results over purely static heuristics. Recent work shows that profile information is also useful for performing source-to-source optimizations via meta-programming. For example, using profiling information to inform decisions about data structures and algorithms can potentially lead to asymptotic improvements in performance. We present a design for profile-guided meta-programming in a general-purpose meta-programming system. Our design is parametric over the particular profiler and meta-programming system. We implement this design in two different meta-programming systems-the syntactic extensions systems of Chez Scheme and Racket-and provide several profile-guided meta-programs as usability case studies.",

keywords = "Meta-programming, Optimization, PGO, Profile-guided optimization, Profiling",

author = "Bowman, {William J.} and Swaha Miller and {St Amour}, Vincent and {Kent Dybvig}, R.",

note = "Publisher Copyright: {\textcopyright} 2015 ACM.",

year = "2015",

month = jun,

doi = "10.1145/2737924.2737990",

language = "English (US)",

volume = "50",

pages = "403--412",

journal = "ACM SIGPLAN Notices",

issn = "1523-2867",

publisher = "Association for Computing Machinery (ACM)",

number = "6",

}

TY - JOUR

T1 - Profile-guided meta-programming

AU - Bowman, William J.

AU - Miller, Swaha

AU - St Amour, Vincent

AU - Kent Dybvig, R.

PY - 2015/6

Y1 - 2015/6

N2 - Contemporary compiler systems such as GCC,.NET, and LLVM incorporate profile-guided optimizations (PGOs) on low-level intermediate code and basic blocks, with impressive results over purely static heuristics. Recent work shows that profile information is also useful for performing source-to-source optimizations via meta-programming. For example, using profiling information to inform decisions about data structures and algorithms can potentially lead to asymptotic improvements in performance. We present a design for profile-guided meta-programming in a general-purpose meta-programming system. Our design is parametric over the particular profiler and meta-programming system. We implement this design in two different meta-programming systems-the syntactic extensions systems of Chez Scheme and Racket-and provide several profile-guided meta-programs as usability case studies.

AB - Contemporary compiler systems such as GCC,.NET, and LLVM incorporate profile-guided optimizations (PGOs) on low-level intermediate code and basic blocks, with impressive results over purely static heuristics. Recent work shows that profile information is also useful for performing source-to-source optimizations via meta-programming. For example, using profiling information to inform decisions about data structures and algorithms can potentially lead to asymptotic improvements in performance. We present a design for profile-guided meta-programming in a general-purpose meta-programming system. Our design is parametric over the particular profiler and meta-programming system. We implement this design in two different meta-programming systems-the syntactic extensions systems of Chez Scheme and Racket-and provide several profile-guided meta-programs as usability case studies.

KW - Meta-programming

KW - Optimization

KW - PGO

KW - Profile-guided optimization

KW - Profiling

UR - http://www.scopus.com/inward/record.url?scp=84951006448&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84951006448&partnerID=8YFLogxK

U2 - 10.1145/2737924.2737990

DO - 10.1145/2737924.2737990

M3 - Article

AN - SCOPUS:84951006448

SN - 1523-2867

VL - 50

SP - 403

EP - 412

JO - ACM SIGPLAN Notices

JF - ACM SIGPLAN Notices

IS - 6

ER -

Profile-guided meta-programming

Abstract

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this