Applying deep learning to the cache replacement problem

Z Shi, X Huang, A Jain, C Lin - Proceedings of the 52nd Annual IEEE …, 2019 - dl.acm.org
Despite its success in many areas, deep learning is a poor fit for use in hardware predictors
because these models are impractically large and slow, but this paper shows how we can …

New attacks and defense for encrypted-address cache

MK Qureshi - Proceedings of the 46th International Symposium on …, 2019 - dl.acm.org
Conflict-based cache attacks can allow an adversary to infer the access pattern of a co-
running application by orchestrating evictions via cache conflicts. Such attacks can be …

Pythia: A customizable hardware prefetching framework using online reinforcement learning

R Bera, K Kanellopoulos, A Nori, T Shahroodi… - MICRO-54: 54th Annual …, 2021 - dl.acm.org
Past research has proposed numerous hardware prefetching techniques, most of which rely
on exploiting one specific type of program context information (eg, program counter …

An imitation learning approach for cache replacement

E Liu, M Hashemi, K Swersky… - International …, 2020 - proceedings.mlr.press
Program execution speed critically depends on increasing cache hits, as cache hits are
orders of magnitude faster than misses. To increase cache hits, we focus on the problem of …

Back to the future: Leveraging Belady's algorithm for improved cache replacement

A Jain, C Lin - ACM SIGARCH Computer Architecture News, 2016 - dl.acm.org
Belady's algorithm is optimal but infeasible because it requires knowledge of the future. This
paper explains how a cache replacement algorithm can nonetheless learn from Belady's …

Fundamental latency trade-off in architecting dram caches: Outperforming impractical sram-tags with a simple and practical design

MK Qureshi, GH Loh - 2012 45th Annual IEEE/ACM …, 2012 - ieeexplore.ieee.org
This paper analyzes the design trade-offs in architecting large-scale DRAM caches. Prior
research, including the recent work from Loh and Hill, have organized DRAM caches similar …

A hierarchical neural model of data prefetching

Z Shi, A Jain, K Swersky, M Hashemi… - Proceedings of the 26th …, 2021 - dl.acm.org
This paper presents Voyager, a novel neural network for data prefetching. Unlike previous
neural models for prefetching, which are limited to learning delta correlations, our model can …

Cameo: A two-level memory organization with capacity of main memory and flexibility of hardware-managed cache

CC Chou, A Jaleel, MK Qureshi - 2014 47th Annual IEEE/ACM …, 2014 - ieeexplore.ieee.org
This paper analyzes the trade-offs in architecting stacked DRAM either as part of main
memory or as a hardware-managed cache. Using stacked DRAM as part of main memory …

A survey of machine learning-based system performance optimization techniques

H Choi, S Park - Applied Sciences, 2021 - mdpi.com
Recently, the machine learning research trend expands to the system performance
optimization field, where it has still been proposed by researchers based on their intuitions …

Perceptron learning for reuse prediction

E Teran, Z Wang, DA Jiménez - 2016 49th Annual IEEE/ACM …, 2016 - ieeexplore.ieee.org
The disparity between last-level cache and memory latencies motivates the search for
efficient cache management policies. Recent work in predicting reuse of cache blocks …