research-article

Vectorization-aware loop unrolling with seed forwarding

Authors:

Hugh LeatherAuthors Info & Claims

CC 2020: Proceedings of the 29th International Conference on Compiler Construction

Pages 1 - 13

https://doi.org/10.1145/3377555.3377890

Published: 24 February 2020 Publication History

Get Access

Abstract

Loop unrolling is a widely adopted loop transformation, commonly used for enabling subsequent optimizations. Straight-line-code vectorization (SLP) is an optimization that benefits from unrolling. SLP converts isomorphic instruction sequences into vector code. Since unrolling generates repeatead isomorphic instruction sequences, it enables SLP to vectorize more code. However, most production compilers apply these optimizations independently and uncoordinated. Unrolling is commonly tuned to avoid code bloat, not maximizing the potential for vectorization, leading to missed vectorization opportunities.

We are proposing VALU, a novel loop unrolling heuristic that takes vectorization into account when making unrolling decisions. Our heuristic is powered by an analysis that estimates the potential benefit of SLP vectorization for the unrolled version of the loop. Our heuristic then selects the unrolling factor that maximizes the utilization of the vector units. VALU also forwards the vectorizable code to SLP, allowing it to bypass its greedy search for vectorizable seed instructions, exposing more vectorization opportunities.

Our evaluation on a production compiler shows that VALU uncovers many vectorization opportunities that were missed by the default loop unroller and vectorizers. This results in more vectorized code and significant performance speedups for 17 of the kernels of the TSVC benchmarks suite, reaching up to 2× speedup over the already highly optimized -O3. Our evaluation on full benchmarks from FreeBench and MiBench shows that VALU results in a geo-mean speedup of 1.06×.

Cited By

View all

Nuzman DZaks ABen-Zion ZRodríguez GSadayappan PSukumaran-Rajam A(2024)If-Convert as Early as You MustProceedings of the 33rd ACM SIGPLAN International Conference on Compiler Construction10.1145/3640537.3641562(26-38)Online publication date: 17-Feb-2024
https://dl.acm.org/doi/10.1145/3640537.3641562
Zhou HHan QShi HZhang YYao JTsafrir DMusuvathi MGupta RAbu-Ghazaleh N(2024)Boost Linear Algebra Computation Performance via Efficient VNNI UtilizationProceedings of the 29th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 310.1145/3620666.3651333(149-163)Online publication date: 27-Apr-2024
https://dl.acm.org/doi/10.1145/3620666.3651333
Murtovi AGeorgakoudis GParasyris KLiao CLaguna ISteffen BGrosser TDubach CSteuwer MXue JOttoni GQuintão Pereira F(2024)Enhancing Performance through Control-Flow Unmerging and Loop Unrolling on GPUsProceedings of the 2024 IEEE/ACM International Symposium on Code Generation and Optimization10.1109/CGO57630.2024.10444819(106-118)Online publication date: 2-Mar-2024
https://dl.acm.org/doi/10.1109/CGO57630.2024.10444819
Show More Cited By

Index Terms

Vectorization-aware loop unrolling with seed forwarding
1. Software and its engineering
  1. Software notations and tools
    1. Compilers

Recommendations

All you need is superword-level parallelism: systematic control-flow vectorization with SLP
PLDI 2022: Proceedings of the 43rd ACM SIGPLAN International Conference on Programming Language Design and Implementation

Superword-level parallelism (SLP) vectorization is a proven technique for vectorizing straight-line code. It works by replacing independent, isomorphic instructions with equivalent vector instructions. Larsen and Amarasinghe originally proposed using ...
Outer-loop vectorization: revisited for short SIMD architectures
PACT '08: Proceedings of the 17th international conference on Parallel architectures and compilation techniques

Vectorization has been an important method of using data-level parallelism to accelerate scientific workloads on vector machines such as Cray for the past three decades. In the last decade it has also proven useful for accelerating multi-media and ...
Fast-path loop unrolling of non-counted loops to enable subsequent compiler optimizations
ManLang '18: Proceedings of the 15th International Conference on Managed Languages & Runtimes

Java programs can contain non-counted loops, that is, loops for which the iteration count can neither be determined at compile time nor at run time. State-of-the-art compilers do not aggressively optimize them, since unrolling non-counted loops often ...

Comments

Information & Contributors

Information

Published In

CC 2020: Proceedings of the 29th International Conference on Compiler Construction

February 2020

222 pages

ISBN:9781450371209

DOI:10.1145/3377555

General Chairs:
Louis-Noël Pouchet
Colorado State University, USA
,
Alexandra Jimborean
Uppsala University, Sweden

Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than ACM must be honored. Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Request permissions from [email protected]

Publisher

Association for Computing Machinery

New York, NY, United States

Publication History

Published: 24 February 2020

Permissions

Request permissions for this article.

Request Permissions

Check for updates

Author Tags

Qualifiers

Research-article

Funding Sources

EPSRC

Conference

CC '20

Sponsor:

SIGPLAN

CC '20: 29th International Conference on Compiler Construction

February 22 - 23, 2020

CA, San Diego, USA

Contributors

Other Metrics

View Article Metrics

Bibliometrics & Citations

Bibliometrics

Article Metrics

11
Total Citations
View Citations
543
Total Downloads

Downloads (Last 12 months)85
Downloads (Last 6 weeks)10

Reflects downloads up to 16 Oct 2024

Other Metrics

View Author Metrics

Citations

Cited By

View all

Nuzman DZaks ABen-Zion ZRodríguez GSadayappan PSukumaran-Rajam A(2024)If-Convert as Early as You MustProceedings of the 33rd ACM SIGPLAN International Conference on Compiler Construction10.1145/3640537.3641562(26-38)Online publication date: 17-Feb-2024
https://dl.acm.org/doi/10.1145/3640537.3641562
Zhou HHan QShi HZhang YYao JTsafrir DMusuvathi MGupta RAbu-Ghazaleh N(2024)Boost Linear Algebra Computation Performance via Efficient VNNI UtilizationProceedings of the 29th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 310.1145/3620666.3651333(149-163)Online publication date: 27-Apr-2024
https://dl.acm.org/doi/10.1145/3620666.3651333
Murtovi AGeorgakoudis GParasyris KLiao CLaguna ISteffen BGrosser TDubach CSteuwer MXue JOttoni GQuintão Pereira F(2024)Enhancing Performance through Control-Flow Unmerging and Loop Unrolling on GPUsProceedings of the 2024 IEEE/ACM International Symposium on Code Generation and Optimization10.1109/CGO57630.2024.10444819(106-118)Online publication date: 2-Mar-2024
https://dl.acm.org/doi/10.1109/CGO57630.2024.10444819
Tayeb HPaillat LBramas B(2023)Autovesk: Automatic Vectorized Code Generation from Unstructured Static Kernels Using Graph TransformationsACM Transactions on Architecture and Code Optimization10.1145/363170921:1(1-25)Online publication date: 9-Nov-2023
https://dl.acm.org/doi/10.1145/3631709
Mosaner RBarany GLeopoldseder DMössenböck HKotselidis CProkopec A(2022)Improving Vectorization Heuristics in a Dynamic Compiler with Machine Learning ModelsProceedings of the 14th ACM SIGPLAN International Workshop on Virtual Machines and Intermediate Languages10.1145/3563838.3567679(36-47)Online publication date: 29-Nov-2022
https://dl.acm.org/doi/10.1145/3563838.3567679
Niu WGuan JShen XWang YAgrawal GRen B(2022) GCD 2 : A Globally Optimizing Compiler for Mapping DNNs to Mobile DSPs 2022 55th IEEE/ACM International Symposium on Microarchitecture (MICRO)10.1109/MICRO56248.2022.00044(512-529)Online publication date: Oct-2022
https://doi.org/10.1109/MICRO56248.2022.00044
Singh ISingh SSingh RKumar S(2022)Efficient Loop Unrolling Factor Prediction Algorithm using Machine Learning Models2022 3rd International Conference for Emerging Technology (INCET)10.1109/INCET54531.2022.9825092(1-8)Online publication date: 27-May-2022
https://doi.org/10.1109/INCET54531.2022.9825092
Rocha RPetoumenos PFranke BBhatotia PO'Boyle M(2022)Loop Rolling for Code Size Reduction2022 IEEE/ACM International Symposium on Code Generation and Optimization (CGO)10.1109/CGO53902.2022.9741256(217-229)Online publication date: 2-Apr-2022
https://doi.org/10.1109/CGO53902.2022.9741256
Damásio TPacheco VGoes FPereira FRocha R(2021)Inlining for Code Size ReductionProceedings of the 25th Brazilian Symposium on Programming Languages10.1145/3475061.3475081(17-24)Online publication date: 27-Sep-2021
https://dl.acm.org/doi/10.1145/3475061.3475081
Porpodas VRatnalikar P(2021)PostSLP: Cross-Region Vectorization of Fully or Partially Vectorized CodeLanguages and Compilers for Parallel Computing10.1007/978-3-030-72789-5_2(15-31)Online publication date: 26-Mar-2021
https://doi.org/10.1007/978-3-030-72789-5_2
Show More Cited By

View Options

Get Access

Login options

Check if you have access through your login credentials or your institution to get full access on this article.

Cited By

Index Terms

Recommendations

All you need is superword-level parallelism: systematic control-flow vectorization with SLP

Outer-loop vectorization: revisited for short SIMD architectures

Fast-path loop unrolling of non-counted loops to enable subsequent compiler optimizations

Comments

Published In

Sponsors

Publisher

Publication History

Permissions

Check for updates

Author Tags

Qualifiers

Funding Sources

Conference

Other Metrics

Article Metrics

Other Metrics

Cited By

Login options

Full Access

PDF

eReader

Abstract

Cited By

Index Terms

Recommendations

All you need is superword-level parallelism: systematic control-flow vectorization with SLP

Outer-loop vectorization: revisited for short SIMD architectures

Fast-path loop unrolling of non-counted loops to enable subsequent compiler optimizations

Comments

Information

Published In

Sponsors

Publisher

Publication History

Permissions

Check for updates

Author Tags

Qualifiers

Funding Sources

Conference

Contributors

Other Metrics

Bibliometrics

Article Metrics

Other Metrics

Citations

Cited By

Get Access

Login options

Full Access

View options

PDF

eReader

Figures

Other

Share

Share this Publication link

Share on social media

Affiliations