Files
hugo-cwpearson/public/publication/2014chen/index.html
2017-03-27 14:36:32 -05:00

401 lines
11 KiB
HTML

<!DOCTYPE html>
<html lang="en-us">
<head>
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1">
<meta http-equiv="X-UA-Compatible" content="IE=edge">
<meta name="theme" content="hugo-academic">
<meta name="generator" content="Hugo 0.19" />
<meta name="author" content="Carl Pearson">
<meta name="description" content="PhD Candidate">
<link rel="stylesheet" href="/css/highlight.min.css">
<link rel="stylesheet" href="/css/bootstrap.min.css">
<link rel="stylesheet" href="/css/font-awesome.min.css">
<link rel="stylesheet" href="/css/academicons.min.css">
<link rel="stylesheet" href="//fonts.googleapis.com/css?family=Lato:400,700|Merriweather|Roboto+Mono">
<link rel="stylesheet" href="/css/hugo-academic.css">
<link rel="alternate" href="https://cwpearson.github.io/index.xml" type="application/rss+xml" title="Carl Pearson">
<link rel="feed" href="https://cwpearson.github.io/index.xml" type="application/rss+xml" title="Carl Pearson">
<link rel="icon" type="image/png" href="/img/icon.png">
<link rel="apple-touch-icon" type="image/png" href="/img/apple-touch-icon.png">
<link rel="canonical" href="https://cwpearson.github.io/publication/2014chen/">
<title>Adaptive Cache Bypass and Insertion for Many-Core Accelerators | Carl Pearson</title>
</head>
<body id="top" data-spy="scroll" data-target="#navbar-main" data-offset="71">
<nav class="navbar navbar-default navbar-fixed-top" id="navbar-main">
<div class="container">
<div class="navbar-header">
<button type="button" class="navbar-toggle collapsed" data-toggle="collapse"
data-target=".navbar-collapse" aria-expanded="false">
<span class="sr-only">Toggle navigation</span>
<span class="icon-bar"></span>
<span class="icon-bar"></span>
<span class="icon-bar"></span>
</button>
<a class="navbar-brand" href="/">Carl Pearson</a>
</div>
<div class="collapse navbar-collapse">
<ul class="nav navbar-nav navbar-right">
<li class="nav-item">
<a href="/#about">
<span>Home</span>
</a>
</li>
<li class="nav-item">
<a href="/#publications">
<span>Publications</span>
</a>
</li>
<li class="nav-item">
<a href="/#projects">
<span>Projects</span>
</a>
</li>
<li class="nav-item">
<a href="/#posts">
<span>Posts</span>
</a>
</li>
<li class="nav-item">
<a href="/#talks">
<span>Talks</span>
</a>
</li>
<li class="nav-item">
<a href="/#teaching">
<span>Teaching</span>
</a>
</li>
<li class="nav-item">
<a href="/#patents">
<span>Patents</span>
</a>
</li>
<li class="nav-item">
<a href="/#awards">
<span>Recognition</span>
</a>
</li>
<li class="nav-item">
<a href="/#contact">
<span>Contact</span>
</a>
</li>
</ul>
</div>
</div>
</nav>
<div class="pub" itemscope itemtype="http://schema.org/CreativeWork">
<div class="container pub-title">
<h1 itemprop="name">Adaptive Cache Bypass and Insertion for Many-Core Accelerators</h1>
<span class="pub-authors" itemprop="author">
Xuhao Chen, Shengzhao Wu, Li-Wen Chang, Wei-Sheng Huang, Carl Pearson, Wen-mei Hwu
</span>
<span class="pull-right">
<div class="share-box" aria-hidden="true">
<ul class="share">
<li>
<a class="facebook"
href="https://www.facebook.com/sharer.php?u=https%3a%2f%2fcwpearson.github.io%2fpublication%2f2014chen%2f"
target="_blank">
<i class="fa fa-facebook"></i>
</a>
</li>
<li>
<a class="twitter"
href="https://twitter.com/intent/tweet?text=Adaptive%20Cache%20Bypass%20and%20Insertion%20for%20Many-Core%20Accelerators&amp;url=https%3a%2f%2fcwpearson.github.io%2fpublication%2f2014chen%2f"
target="_blank">
<i class="fa fa-twitter"></i>
</a>
</li>
<li>
<a class="linkedin"
href="https://www.linkedin.com/shareArticle?mini=true&amp;url=https%3a%2f%2fcwpearson.github.io%2fpublication%2f2014chen%2f&amp;title=Adaptive%20Cache%20Bypass%20and%20Insertion%20for%20Many-Core%20Accelerators"
target="_blank">
<i class="fa fa-linkedin"></i>
</a>
</li>
<li>
<a class="weibo"
href="http://service.weibo.com/share/share.php?url=https%3a%2f%2fcwpearson.github.io%2fpublication%2f2014chen%2f&amp;title=Adaptive%20Cache%20Bypass%20and%20Insertion%20for%20Many-Core%20Accelerators"
target="_blank">
<i class="fa fa-weibo"></i>
</a>
</li>
<li>
<a class="email"
href="mailto:?subject=Adaptive%20Cache%20Bypass%20and%20Insertion%20for%20Many-Core%20Accelerators&amp;body=https%3a%2f%2fcwpearson.github.io%2fpublication%2f2014chen%2f">
<i class="fa fa-envelope"></i>
</a>
</li>
</ul>
</div>
</span>
</div>
</div>
<div class="article-container">
<h3>Abstract</h3>
<p class="pub-abstract" itemprop="text">Many-core accelerators, e.g. GPUs, are widely used for accelerating general-purpose compute kernels.
With the SIMT execution model, GPUs can hide memory latency through massive multithreading for many regular applications.
To support more applications with irregular memory access pattern, cache hierarchy is introduced to GPU architecture to capture input data sharing and mitigate the effect of irregular accesses.
However, GPU caches suffer from poor efficiency due to severe contention, which makes it difficult to adopt heuristic management policies, and also limits system performance and energy-efficiency.
We propose an adaptive cache management policy specifically for many-core accelerators.
The tag array of L2 cache is enhanced with extra bits to track memory access history, an thus the locality information is captured and provided to L1 cache as heuristics to guide its run-time bypass and insertion decisions.
By preventing un-reused data from polluting the cache and alleviating contention, cache efficiency is significantly improved.
As a result, the system performance is improved by 31% on average for cache sensitive benchmarks, compared to the baseline GPU architecture.</p>
<div class="row">
<div class="col-sm-1"></div>
<div class="col-sm-10">
<div class="row">
<div class="col-xs-12 col-sm-3 pub-row-heading">Publication</div>
<div class="col-xs-12 col-sm-9"><em>Proceedings of International Workshop on Manycore Embedded Systems.</em> ACM, 2014.</div>
</div>
</div>
<div class="col-sm-1"></div>
</div>
<div class="visible-xs space-below"></div>
<div class="row">
<div class="col-sm-1"></div>
<div class="col-sm-10">
<div class="row">
<div class="col-xs-12 col-sm-3 pub-row-heading">Date</div>
<div class="col-xs-12 col-sm-9" itemprop="datePublished">
January, 2014
</div>
</div>
</div>
<div class="col-sm-1"></div>
</div>
<div class="visible-xs space-below"></div>
<div class="row" style="padding-top: 10px">
<div class="col-sm-1"></div>
<div class="col-sm-10">
<div class="row">
<div class="col-xs-12 col-sm-3 pub-row-heading" style="line-height:34px;">Links</div>
<div class="col-xs-12 col-sm-9">
<a class="btn btn-primary btn-outline" href="https://cwpearson.github.io/pdf/2014chen.pdf">
PDF
</a>
</div>
</div>
</div>
<div class="col-sm-1"></div>
</div>
<div class="visible-xs space-below"></div>
<div class="space-below"></div>
<div class="article-style"></div>
</div>
<div class="container">
<nav>
<ul class="pager">
<li class="next"><a href="https://cwpearson.github.io/publication/2016dakkak/">WebGPU: A Scalable Online Development Platform for GPU Programming Courses <span
aria-hidden="true">&rarr;</span></a></li>
</ul>
</nav>
</div>
<footer class="site-footer">
<div class="container">
<p class="powered-by">
&copy; 2017 Carl Pearson &middot;
Powered by the <a href="https://github.com/gcushen/hugo-academic" target="_blank">Academic
theme</a> for <a href="http://gohugo.io" target="_blank">Hugo</a>.
<span class="pull-right" aria-hidden="true">
<a href="#" id="back_to_top">
<span class="button_icon">
<i class="fa fa-chevron-up fa-2x"></i>
</span>
</a>
</span>
</p>
</div>
</footer>
<script src="//cdnjs.cloudflare.com/ajax/libs/gsap/1.18.4/TweenMax.min.js"></script>
<script src="//cdnjs.cloudflare.com/ajax/libs/gsap/latest/plugins/ScrollToPlugin.min.js"></script>
<script src="/js/jquery-1.12.3.min.js"></script>
<script src="/js/bootstrap.min.js"></script>
<script src="/js/isotope.pkgd.min.js"></script>
<script src="//cdnjs.cloudflare.com/ajax/libs/jquery.imagesloaded/4.1.1/imagesloaded.pkgd.min.js"></script>
<script src="/js/hugo-academic.js"></script>
<script>
(function(i,s,o,g,r,a,m){i['GoogleAnalyticsObject']=r;i[r]=i[r]||function(){
(i[r].q=i[r].q||[]).push(arguments)},i[r].l=1*new Date();a=s.createElement(o),
m=s.getElementsByTagName(o)[0];a.async=1;a.src=g;m.parentNode.insertBefore(a,m)
})(window,document,'script','https://www.google-analytics.com/analytics.js','ga');
ga('create', 'UA-54975774-2', 'auto');
ga('send', 'pageview');
var links = document.querySelectorAll('a');
Array.prototype.map.call(links, function(item) {
if (item.host != document.location.host) {
item.addEventListener('click', function() {
var action = item.getAttribute('data-action') || 'follow';
ga('send', 'event', 'outbound', action, item.href);
});
}
});
</script>
<script src="//cdnjs.cloudflare.com/ajax/libs/highlight.js/9.9.0/highlight.min.js"></script>
<script>hljs.initHighlightingOnLoad();</script>
</body>
</html>