Machine learning-based prefetch optimization for data center applications

Authors:
Shih-wei Liao;Tzu-Han Hung;Donald Nguyen;Chinyen Chou;Chiaheng Tu;Hucheng Zhou
Affiliations:
Google Inc., Mountain View, California and National Taiwan University, Taipei, Taiwan;Princeton University, Princeton, New Jersey;University of Texas at Austin, Austin, Texas;National Taiwan University, Taipei, Taiwan;National Taiwan University, Taipei, Taiwan;Tsinghua University, Beijing, China
Venue:
Proceedings of the Conference on High Performance Computing Networking, Storage and Analysis
Year:
2009

Citing 18
Cited 3

Instance-Based Learning Algorithms

Machine Learning
C4.5: programs for machine learning

C4.5: programs for machine learning
Prefetching using Markov predictors

Proceedings of the 24th annual international symposium on Computer architecture
Fast training of support vector machines using sequential minimal optimization

Advances in kernel methods
Neural Networks: A Comprehensive Foundation

Neural Networks: A Comprehensive Foundation
SMARTS: accelerating microarchitecture simulation via rigorous statistical sampling

Proceedings of the 30th annual international symposium on Computer architecture
Radial Basis Functions

Radial Basis Functions
Scaling and Charact rizing Database Workloads: Bridging the Gap between Research and Practice

Proceedings of the 36th annual IEEE/ACM International Symposium on Microarchitecture
Temporal Streaming of Shared Memory

Proceedings of the 32nd annual international symposium on Computer Architecture
Spatial Memory Streaming

Proceedings of the 33rd annual international symposium on Computer Architecture
Accurate and efficient regression modeling for microarchitectural performance and power prediction

Proceedings of the 12th international conference on Architectural support for programming languages and operating systems
Efficiently exploring architectural design spaces via predictive modeling

Proceedings of the 12th international conference on Architectural support for programming languages and operating systems
Memory Prefetching Using Adaptive Stream Detection

Proceedings of the 39th Annual IEEE/ACM International Symposium on Microarchitecture
Using Machine Learning to Guide Architecture Simulation

The Journal of Machine Learning Research
Low-Cost Epoch-Based Correlation Prefetching for Commercial Applications

Proceedings of the 40th Annual IEEE/ACM International Symposium on Microarchitecture
Efficiency trends and limits from comprehensive microarchitectural adaptivity

Proceedings of the 13th international conference on Architectural support for programming languages and operating systems
Spatio-temporal memory streaming

Proceedings of the 36th annual international symposium on Computer architecture
Estimating continuous distributions in Bayesian classifiers

UAI'95 Proceedings of the Eleventh conference on Uncertainty in artificial intelligence

Evaluation of streaming aggregation on parallel hardware architectures

Proceedings of the Fourth ACM International Conference on Distributed Event-Based Systems
Enabling dynamic data centers with a smart bare-metal server platform

Cluster Computing
Making data prefetch smarter: adaptive prefetching on POWER7

Proceedings of the 21st international conference on Parallel architectures and compilation techniques

Quantified Score

Hi-index	0.00

Visualization

Abstract

Performance tuning for data centers is essential and complicated. It is important since a data center comprises thousands of machines and thus a single-digit performance improvement can significantly reduce cost and power consumption. Unfortunately, it is extremely difficult as data centers are dynamic environments where applications are frequently released and servers are continually upgraded. In this paper, we study the effectiveness of different processor prefetch configurations, which can greatly influence the performance of memory system and the overall data center. We observe a wide performance gap when comparing the worst and best configurations, from 1.4% to 75.1%, for 11 important data center applications. We then develop a tuning framework which attempts to predict the optimal configuration based on hardware performance counters. The framework achieves performance within 1% of the best performance of any single configuration for the same set of applications.