TU Darmstadt / ULB / TUbiblio

Learning to Make Compiler Optimizations More Effective

Mammadli, Rahim ; Selakovic, Marija ; Pradel, Michael ; Wolf, Felix (2021)
Learning to Make Compiler Optimizations More Effective.
PLDI '21: 42nd ACM SIGPLAN International Conference on Programming Language Design and Implementation. virtual Conference (21.06.2021)
doi: 10.1145/3460945.3464952
Conference or Workshop Item, Bibliographie

Abstract

Because loops execute their body many times, compiler developers place much emphasis on their optimization. Nevertheless, in view of highly diverse source code and hardware, compilers still struggle to produce optimal target code. The sheer number of possible loop optimizations, including their combinations, exacerbates the problem further. Today's compilers use hard-coded heuristics to decide when, whether, and which of a limited set of optimizations to apply. Often, this leads to highly unstable behavior, making the success of compiler optimizations dependent on the precise way a loop has been written. This paper presents LoopLearner, which addresses the problem of compiler instability by predicting which way of writing a loop will lead to efficient compiled code. To this end, we train a neural network to find semantically invariant source-level transformations for loops that help the compiler generate more efficient code. Our model learns to extract useful features from the raw source code and predicts the speedup that a given transformation is likely to yield. We evaluate LoopLearner with 1,895 loops from various performance-relevant benchmarks. Applying the transformations that our model deems most favorable prior to compilation yields an average speedup of 1.14x. When trying the top-3 suggested transformations, the average speedup even increases to 1.29x. Comparing the approach with an exhaustive search through all available code transformations shows that LoopLearner helps to identify the most beneficial transformations in several orders of magnitude less time.

Item Type: Conference or Workshop Item
Erschienen: 2021
Creators: Mammadli, Rahim ; Selakovic, Marija ; Pradel, Michael ; Wolf, Felix
Type of entry: Bibliographie
Title: Learning to Make Compiler Optimizations More Effective
Language: English
Date: 20 June 2021
Place of Publication: New York, NY, USA
Publisher: ACM
Book Title: MAPS 2021: Proceedings of the 5th ACM SIGPLAN International Symposium on Machine Programming
Event Title: PLDI '21: 42nd ACM SIGPLAN International Conference on Programming Language Design and Implementation
Event Location: virtual Conference
Event Dates: 21.06.2021
DOI: 10.1145/3460945.3464952
Abstract:

Because loops execute their body many times, compiler developers place much emphasis on their optimization. Nevertheless, in view of highly diverse source code and hardware, compilers still struggle to produce optimal target code. The sheer number of possible loop optimizations, including their combinations, exacerbates the problem further. Today's compilers use hard-coded heuristics to decide when, whether, and which of a limited set of optimizations to apply. Often, this leads to highly unstable behavior, making the success of compiler optimizations dependent on the precise way a loop has been written. This paper presents LoopLearner, which addresses the problem of compiler instability by predicting which way of writing a loop will lead to efficient compiled code. To this end, we train a neural network to find semantically invariant source-level transformations for loops that help the compiler generate more efficient code. Our model learns to extract useful features from the raw source code and predicts the speedup that a given transformation is likely to yield. We evaluate LoopLearner with 1,895 loops from various performance-relevant benchmarks. Applying the transformations that our model deems most favorable prior to compilation yields an average speedup of 1.14x. When trying the top-3 suggested transformations, the average speedup even increases to 1.29x. Comparing the approach with an exhaustive search through all available code transformations shows that LoopLearner helps to identify the most beneficial transformations in several orders of magnitude less time.

Divisions: Study Areas
20 Department of Computer Science
20 Department of Computer Science > Parallel Programming
Study Areas > Study area Computational Engineering
Date Deposited: 13 Feb 2024 15:05
Last Modified: 21 May 2024 08:14
PPN: 518445658
Export:
Suche nach Titel in: TUfind oder in Google
Send an inquiry Send an inquiry

Options (only for editors)
Show editorial Details Show editorial Details