Getting swole: Generating access-aware code with predicate pullups

Andrew Crotty; Alex Galakatos; Tim Kraska

doi:10.1109/ICDE48307.2020.00114

Getting swole: Generating access-aware code with predicate pullups

Andrew Crotty, Alex Galakatos, Tim Kraska

Computer Science

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

7 Scopus citations

Abstract

Code generation for in-memory query processing is now commonplace. While existing approaches use a wide range of techniques (e.g., inline expansion, pipelining, SIMD vectorization, prefetching) to reduce processing effort, we argue that generating code with better data access patterns is often more important. Therefore, we propose SWOLE, the first access-aware code generation strategy. Contradictory to the conventional wisdom, SWOLE heavily leverages predicate pullups to produce code with better access patterns, which outweighs the overhead of performing wasted work. Our experiments show that SWOLE can outperform the state-of-the-art approach by over 2.6×.

Original language	English (US)
Title of host publication	Proceedings - 2020 IEEE 36th International Conference on Data Engineering, ICDE 2020
Publisher	IEEE Computer Society
Pages	1273-1284
Number of pages	12
ISBN (Electronic)	9781728129037
DOIs	https://doi.org/10.1109/ICDE48307.2020.00114
State	Published - Apr 2020
Event	36th IEEE International Conference on Data Engineering, ICDE 2020 - Dallas, United States Duration: Apr 20 2020 → Apr 24 2020

Publication series

Name	Proceedings - International Conference on Data Engineering
Volume	2020-April
ISSN (Print)	1084-4627

Conference

Conference	36th IEEE International Conference on Data Engineering, ICDE 2020
Country/Territory	United States
City	Dallas
Period	4/20/20 → 4/24/20

ASJC Scopus subject areas

Software
Signal Processing
Information Systems

Access to Document

10.1109/ICDE48307.2020.00114

Cite this

Crotty, A., Galakatos, A., & Kraska, T. (2020). Getting swole: Generating access-aware code with predicate pullups. In Proceedings - 2020 IEEE 36th International Conference on Data Engineering, ICDE 2020 (pp. 1273-1284). Article 9101186 (Proceedings - International Conference on Data Engineering; Vol. 2020-April). IEEE Computer Society. https://doi.org/10.1109/ICDE48307.2020.00114

@inproceedings{2ee8bca9d73c4b31b03709d22b2de9f1,

title = "Getting swole: Generating access-aware code with predicate pullups",

abstract = "Code generation for in-memory query processing is now commonplace. While existing approaches use a wide range of techniques (e.g., inline expansion, pipelining, SIMD vectorization, prefetching) to reduce processing effort, we argue that generating code with better data access patterns is often more important. Therefore, we propose SWOLE, the first access-aware code generation strategy. Contradictory to the conventional wisdom, SWOLE heavily leverages predicate pullups to produce code with better access patterns, which outweighs the overhead of performing wasted work. Our experiments show that SWOLE can outperform the state-of-the-art approach by over 2.6×.",

author = "Andrew Crotty and Alex Galakatos and Tim Kraska",

note = "Funding Information: ACKNOWLEDGMENTS This research was funded by NSF Career Award IIS-1453171 and supported by a Google PhD Fellowship. Publisher Copyright: {\textcopyright} 2020 IEEE.; 36th IEEE International Conference on Data Engineering, ICDE 2020 ; Conference date: 20-04-2020 Through 24-04-2020",

year = "2020",

month = apr,

doi = "10.1109/ICDE48307.2020.00114",

language = "English (US)",

series = "Proceedings - International Conference on Data Engineering",

publisher = "IEEE Computer Society",

pages = "1273--1284",

booktitle = "Proceedings - 2020 IEEE 36th International Conference on Data Engineering, ICDE 2020",

address = "United States",

}

Crotty, A, Galakatos, A & Kraska, T 2020, Getting swole: Generating access-aware code with predicate pullups. in Proceedings - 2020 IEEE 36th International Conference on Data Engineering, ICDE 2020., 9101186, Proceedings - International Conference on Data Engineering, vol. 2020-April, IEEE Computer Society, pp. 1273-1284, 36th IEEE International Conference on Data Engineering, ICDE 2020, Dallas, United States, 4/20/20. https://doi.org/10.1109/ICDE48307.2020.00114

Getting swole: Generating access-aware code with predicate pullups. / Crotty, Andrew; Galakatos, Alex; Kraska, Tim.
Proceedings - 2020 IEEE 36th International Conference on Data Engineering, ICDE 2020. IEEE Computer Society, 2020. p. 1273-1284 9101186 (Proceedings - International Conference on Data Engineering; Vol. 2020-April).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Getting swole

T2 - 36th IEEE International Conference on Data Engineering, ICDE 2020

AU - Crotty, Andrew

AU - Galakatos, Alex

AU - Kraska, Tim

PY - 2020/4

Y1 - 2020/4

N2 - Code generation for in-memory query processing is now commonplace. While existing approaches use a wide range of techniques (e.g., inline expansion, pipelining, SIMD vectorization, prefetching) to reduce processing effort, we argue that generating code with better data access patterns is often more important. Therefore, we propose SWOLE, the first access-aware code generation strategy. Contradictory to the conventional wisdom, SWOLE heavily leverages predicate pullups to produce code with better access patterns, which outweighs the overhead of performing wasted work. Our experiments show that SWOLE can outperform the state-of-the-art approach by over 2.6×.

AB - Code generation for in-memory query processing is now commonplace. While existing approaches use a wide range of techniques (e.g., inline expansion, pipelining, SIMD vectorization, prefetching) to reduce processing effort, we argue that generating code with better data access patterns is often more important. Therefore, we propose SWOLE, the first access-aware code generation strategy. Contradictory to the conventional wisdom, SWOLE heavily leverages predicate pullups to produce code with better access patterns, which outweighs the overhead of performing wasted work. Our experiments show that SWOLE can outperform the state-of-the-art approach by over 2.6×.

UR - http://www.scopus.com/inward/record.url?scp=85085865401&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85085865401&partnerID=8YFLogxK

U2 - 10.1109/ICDE48307.2020.00114

DO - 10.1109/ICDE48307.2020.00114

M3 - Conference contribution

AN - SCOPUS:85085865401

T3 - Proceedings - International Conference on Data Engineering

SP - 1273

EP - 1284

BT - Proceedings - 2020 IEEE 36th International Conference on Data Engineering, ICDE 2020

PB - IEEE Computer Society

Y2 - 20 April 2020 through 24 April 2020

ER -

Getting swole: Generating access-aware code with predicate pullups

Abstract

Publication series

Conference

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this