Efficient HPC libraries often expose multiple tunable parameters, algorithmic implementations, or a combination of them, to provide optimized routines. The optimal parameters and algorithmic choices may depend on input properties such as the shapes of the matrices involved in the operation. Traditionally, these parameters are manually tuned or set by auto-tuners. In emerging applications such as deep learning, this approach is not effective across the wide range of inputs and architectures used in practice. In this work, we analyze different machine learning techniques and predictive models to accelerate the convolution operator and GEMM. Moreover, we address the problem of dataset generation, and we study the performance, accuracy, and generalization ability of the models. Our insights allow us to improve the performance of computationally expensive deep learning primitives on high-end GPUs as well as low-power embedded GPU architectures on three different libraries. Experimental results show significant improvement in the target applications from 50% up to 300% compared to auto-tuned and high-optimized vendor-based heuristics by using simple decision tree- and MLP-based models.

On the Anatomy of Predictive Models for Accelerating GPU Convolution Kernels and Beyond / Paolo Sylos Labini; Marco Cianfriglia; Damiano Perri; Osvaldo Gervasi; Grigori Fursin; Anton Lokhmotov; Cedric Nugteren; Bruno Carpentieri; Fabiana Zollo; Flavio Vella. - In: ACM TRANSACTIONS ON ARCHITECTURE AND CODE OPTIMIZATION. - ISSN 1544-3566. - ELETTRONICO. - 18:(2021), pp. 1-24. [10.1145/3434402]

On the Anatomy of Predictive Models for Accelerating GPU Convolution Kernels and Beyond

Damiano Perri;
2021

Abstract

Efficient HPC libraries often expose multiple tunable parameters, algorithmic implementations, or a combination of them, to provide optimized routines. The optimal parameters and algorithmic choices may depend on input properties such as the shapes of the matrices involved in the operation. Traditionally, these parameters are manually tuned or set by auto-tuners. In emerging applications such as deep learning, this approach is not effective across the wide range of inputs and architectures used in practice. In this work, we analyze different machine learning techniques and predictive models to accelerate the convolution operator and GEMM. Moreover, we address the problem of dataset generation, and we study the performance, accuracy, and generalization ability of the models. Our insights allow us to improve the performance of computationally expensive deep learning primitives on high-end GPUs as well as low-power embedded GPU architectures on three different libraries. Experimental results show significant improvement in the target applications from 50% up to 300% compared to auto-tuned and high-optimized vendor-based heuristics by using simple decision tree- and MLP-based models.
2021
18
1
24
Paolo Sylos Labini; Marco Cianfriglia; Damiano Perri; Osvaldo Gervasi; Grigori Fursin; Anton Lokhmotov; Cedric Nugteren; Bruno Carpentieri; Fabiana Zollo; Flavio Vella
File in questo prodotto:
File Dimensione Formato  
3434402.pdf

accesso aperto

Tipologia: Pdf editoriale (Version of record)
Licenza: Creative commons
Dimensione 2.95 MB
Formato Adobe PDF
2.95 MB Adobe PDF

I documenti in FLORE sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificatore per citare o creare un link a questa risorsa: https://hdl.handle.net/2158/1293233
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 6
  • ???jsp.display-item.citation.isi??? 6
social impact