Mapping Parallelism in a Functional IR through Constraint Satisfaction: A Case Study on Convolution for Mobile GPUs

Naums Mogers, Lu Li, Valentin Radu, Christophe Dubach

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Graphics Processing Units (GPUs) are notoriously hard to optimize for manually. What is needed are good automatic code generators and optimizers. Accelerate, Futhark and Lift demonstrated that a functional approach is well suited for this challenge. Lift, for instance, uses a system of rewrite rules with a multi-stage approach. Algorithmic optimizations are first explored, followed by hardware-specific optimizations such as using shared memory and mapping parallelism.

While the algorithmic exploration leads to correct transformed programs by construction, it is not necessarily true for the latter phase. Exploiting shared memory and mapping parallelism while ensuring correct synchronization is a delicate balancing act, and is hard to encode in a rewrite system. Currently, Lift relies on heuristics with ad-hoc mechanisms to check for correctness. Although this practical approach eventually produces high-performance code, it is not an ideal state of affairs.

This paper proposes to extract parallelization constraints automatically from a functional IR and use a solver to identify valid rewriting. Using a convolutional neural network on a mobile GPU as a use case, this approach matches the performance of the ARM Compute Library GEMM convolution and the TVM-generated kernel consuming between 2.7x and 3.6x less memory on average. Furthermore, a speedup of 12x is achieved over the ARM Compute Library direct convolution implementation.
Original languageEnglish
Title of host publicationProceedings of the 31st ACM SIGPLAN International Conference on Compiler Construction
EditorsBernhard Egger, Aaron Smith
PublisherACM Association for Computing Machinery
Pages218-230
Number of pages13
ISBN (Print)9781450391832
DOIs
Publication statusPublished - 19 Mar 2022
EventACM SIGPLAN 2022 International Conference on Compiler Construction - Online Conference
Duration: 2 Apr 20223 Apr 2022
Conference number: 31

Conference

ConferenceACM SIGPLAN 2022 International Conference on Compiler Construction
Abbreviated titleCC 2022
Period2/04/223/04/22

Keywords / Materials (for Non-textual outputs)

  • code generation
  • convolution
  • mobile GPU
  • parallelism

Fingerprint

Dive into the research topics of 'Mapping Parallelism in a Functional IR through Constraint Satisfaction: A Case Study on Convolution for Mobile GPUs'. Together they form a unique fingerprint.
  • Eddie

    Mike Wallis (Other)

    IT Infrastructure

    Facility/equipment: Facility

Cite this