-
Notifications
You must be signed in to change notification settings - Fork 3
/
publications.html
162 lines (157 loc) · 16.3 KB
/
publications.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
<html >
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
<meta http-equiv="Content-Style-Type" content="text/css">
<meta name="generator" content="pandoc">
<meta name="viewport" content="width=device-width, initial-scale=1">
<link rel="stylesheet" href="https://maxcdn.bootstrapcdn.com/bootstrap/3.3.7/css/bootstrap.min.css" integrity="sha384-BVYiiSIFeK1dGmJRAkycuHAHRg32OmUcww7on3RYdg4Va+PmSTsz/K68vbdEjh4u" crossorigin="anonymous">
<link rel="stylesheet" type="text/css" href="/style.css">
<link rel="stylesheet" type="text/css" href="https://fonts.googleapis.com/css?family=Raleway:600,800|Open+Sans">
<script src="https://ajax.googleapis.com/ajax/libs/jquery/1.12.4/jquery.min.js"></script>
<script src="https://maxcdn.bootstrapcdn.com/bootstrap/3.3.7/js/bootstrap.min.js" integrity="sha384-Tc5IQib027qvyjSMfHjOMaLkfuWVxZxUPnCJA7l2mCWNIpG9mGCD8wGNIcPD7Txa" crossorigin="anonymous"></script>
<script src="https://use.fontawesome.com/c754f00c7e.js"></script>
<!-- <script src="https://cdn.polyfill.io/v2/polyfill.min.js"></script> -->
<title>Accelerate: Publications</title>
<style type="text/css">code{white-space: pre;}</style>
</head>
<nav class="navbar navbar-inverse navbar-fixed-top">
<div class="container-fluid">
<!-- Brand and toggle get grouped for better mobile display -->
<div class="navbar-header">
<button type="button" class="navbar-toggle collapsed" data-toggle="collapse" data-target="#bs-example-navbar-collapse-1" 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="/index.html" style="font-family: 'Raleway'; font-weight: 600;">AccelerateHS</a>
</div>
<!-- Collect the nav links, forms, and other content for toggling -->
<div class="collapse navbar-collapse" id="bs-example-navbar-collapse-1">
<ul class="nav navbar-nav">
<li><a href="/get-started.html">Get Started</a></li>
<li><a href="/documentation.html">Documentation</a></li>
<li><a href="/libraries.html">Libraries</a></li>
<li><a href="/examples.html">Examples</a></li>
<li><a href="/community.html">Community</a></li>
<li class="active"><a href="/publications.html">Publications</a></li>
</ul>
<ul class="nav navbar-nav navbar-right">
<li><a href="https://github.com/AccelerateHS/accelerate"><i class="fa fa-github fa-lg" aria-hidden="true"></i></a></li>
<li><a href="http://groups.google.com/group/accelerate-haskell"><i class="fa fa-envelope-o fa-lg" aria-hidden="true"></i></a></li>
</ul>
</div><!-- /.navbar-collapse -->
</div><!-- /.container-fluid -->
</nav>
<body>
<div class="container">
<div class="row">
<div id="TOC" class="well">
<!-- <ul> -->
<!-- <li class="nav-header" id="toctitle">Table of Contents</il> -->
<!-- </ul> -->
<ul>
<li><a href="#publications">Publications</a><ul>
<li><a href="#streaming-irregular-arrays">Streaming Irregular Arrays</a></li>
<li><a href="#type-safe-runtime-code-generation-accelerate-to-llvm">Type-safe Runtime Code Generation: Accelerate to LLVM</a></li>
<li><a href="#functional-array-streams">Functional Array Streams</a></li>
<li><a href="#converting-data-parallelism-to-tast-parallelism-by-rewrites">Converting Data-Parallelism to Tast-Parallelism by Rewrites</a></li>
<li><a href="#embedding-foreign-code">Embedding Foreign Code</a></li>
<li><a href="#optimising-purely-functional-gpu-programs">Optimising Purely Functional GPU Programs</a></li>
<li><a href="#accelerating-haskell-array-codes-with-multicore-gpus">Accelerating Haskell Array Codes with Multicore GPUs</a></li>
</ul></li>
</ul>
</div>
<div class="container">
<h1 id="publications">Publications</h1>
<p>If you use Accelerate for academic research, you are encouraged (though certainly not required) to cite the following papers, which explain various aspects of the system.</p>
<p>Accelerate is primarily developed by academics, so citations matter a lot to us. As an added benefit, you increase Accelerate's exposure and potential user (and developer!) base, which is a benefit to all users of Accelerate. Thanks in advance!</p>
<p>In reverse chronological order:</p>
<h3 id="streaming-irregular-arrays">Streaming Irregular Arrays</h3>
<p>Robert Clifton-Everest, Trevor L. McDonell, Manuel M. T. Chakravarty, and Gabriele Keller.</p>
<p>In <a href="https://www.haskell.org/haskell-symposium/2017/"><em>Haskell '17: The 10th ACM SIGPLAN Symposium on Haskell</em></a>, ACM, 2017.</p>
<p><strong>Abstract:</strong></p>
<p>Previous work has demonstrated that it is possible to generate efficient and highly parallel code for multicore CPUs and GPUs from combinator-based array languages for a range of applications. That work, however, has been limited to operating on flat, rectangular structures without any facilities for irregularity or nesting.</p>
<p>In this paper, we show that even a limited form of nesting provides substantial benefits both in terms of the expressiveness of the language (increasing modularity and providing support for simple irregular structures) and the portability of the code (increasing portability across resource-constrained devices, such as GPUs). Specifically, we generalise Blelloch's flattening transformation along two lines: (1) we explicitly distinguish between definitely regular and potentially irregular computations; and (2) we handle multidimensional arrays. We demonstrate the utility of this generalisation by an extension of the embedded array language Accelerate to include irregular streams of multidimensional arrays. We discuss code generation, optimisation, and irregular stream scheduling as well as a range of benchmarks on both multicore CPUs and GPUs.</p>
<ul>
<li><a href="https://github.com/AccelerateHS/acceleratehs.github.io/raw/master/papers/acc-seq2-haskell2017.pdf">PDF</a> (12 pages)</li>
<li><a href="https://github.com/AccelerateHS/acceleratehs.github.io/raw/master/papers/acc-seq2-haskell2017.bib">bibtex</a></li>
</ul>
<h3 id="type-safe-runtime-code-generation-accelerate-to-llvm">Type-safe Runtime Code Generation: Accelerate to LLVM</h3>
<p>Trevor L. McDonell, Manuel M. T. Chakravarty, Vinod Grover, and Ryan R. Newton.</p>
<p>In <a href="https://www.haskell.org/haskell-symposium/2015/"><em>Haskell '15: The 8th ACM SIGPLAN Symposium on Haskell</em></a>, ACM, 2015.</p>
<p><strong>Abstract:</strong></p>
<p>Embedded languages are often compiled at application runtime; thus, <em>embedded compile-time errors</em> become <em>application runtime errors</em>. We argue that advanced type system features, such as GADTs and type families, play a crucial role in minimising such runtime errors. Specifically, a rigorous type discipline reduces run- time errors due to bugs in both embedded language applications and the implementation of the embedded language compiler itself.</p>
<p>In this paper, we focus on the safety guarantees achieved by type preserving approach by creating a new type-safe interface to the industrial-strength LLVM are able to preserve types from the source language down to a low-level register compilation. We discuss the compilation pipeline of <em>Accelerate</em>, a compiler infrastructure, which we used to build two new Accelerate backends that high-performance array language targeting both multicore CPUs and GPUs, where we language in SSA form. Specifically, we demonstrate the practicability of our show competitive runtimes on a set of benchmarks across both CPUs and GPUs.</p>
<ul>
<li><a href="https://github.com/AccelerateHS/acceleratehs.github.io/raw/master/papers/acc-llvm-haskell2015.pdf">PDF</a> (12 pages)</li>
<li><a href="https://www.youtube.com/watch?v=snXhXA5noVc">video</a> (22 mins)</li>
<li><a href="https://speakerdeck.com/tmcdonell/type-safe-runtime-code-generation-accelerate-to-llvm">slides</a></li>
<li><a href="https://github.com/AccelerateHS/acceleratehs.github.io/raw/master/papers/acc-llvm-haskell2015.bib">bibtex</a></li>
</ul>
<h3 id="functional-array-streams">Functional Array Streams</h3>
<p>Frederik M. Madsen, Robert Clifton-Everest, Manuel M. T. Chakravarty, and Gabriele Keller</p>
<p>In <a href="https://sites.google.com/site/fhpcworkshops/fhpc-2015"><em>FHPC '15: The 4th ACM SIGPLAN Workshop on Functional High-Performance Computing</em></a>, ACM, 2015.</p>
<p><strong>Abstract:</strong></p>
<p>Regular array languages for high performance computing based on aggregate operations provide a convenient parallel programming model, which enables the generation of efficient code for SIMD architectures, such as GPUs. However, the data sets that can be processed with current implementations are severely constrained by the limited amount of main memory available in these architectures.</p>
<p>In this paper, we propose an extension of the embedded array language Accelerate with a notion of sequences, resulting in a two level hierarchy which allows the programmer to specify a partitioning strategy which facilitates automatic resource allocation. Depending on the available memory, the runtime system processes the overall data set in streams of chunks appropriate to the hardware parameters.</p>
<p>In this paper, we present the language design for the sequence operations, as well as the compilation and runtime support, and demonstrate with a set of benchmarks the feasibility of this approach.</p>
<ul>
<li><a href="https://github.com/AccelerateHS/acceleratehs.github.io/raw/master/papers/acc-seq-fhpc2015.pdf">PDF</a> (12 pages)</li>
<li><a href="https://github.com/AccelerateHS/acceleratehs.github.io/raw/master/papers/acc-seq-fhpc2015.bib">bibtex</a></li>
</ul>
<h3 id="converting-data-parallelism-to-tast-parallelism-by-rewrites">Converting Data-Parallelism to Tast-Parallelism by Rewrites</h3>
<p>Bo Joel Svensson, Michael Vollmer, Eric Holk, Trevor L. McDonell, and Ryan R. Newton</p>
<p>In <a href="https://sites.google.com/site/fhpcworkshops/fhpc-2015"><em>FHPC '15: The 4th ACM SIGPLAN Workshop on Functional High-Performance Computing</em></a>, ACM, 2015.</p>
<p><strong>Abstract:</strong></p>
<p>High-level domain-specific languages for array processing on the GPU are increasingly common, but they typically only run on a single GPU. As computational power is distributed across more devices, languages must target <em>multiple</em> devices simultaneously. To this end, we present a compositional translation that fissions data- parallel programs in the <em>Accelerate</em> language, allowing subsequent compiler and runtime stages to map computations onto multiple devices for improved performance—even programs that begin as a single data-parallel kernel.</p>
<ul>
<li><a href="https://github.com/AccelerateHS/acceleratehs.github.io/raw/master/papers/acc-multidev-fhpc2015.pdf">PDF</a> (12 pages)</li>
<li><a href="https://github.com/AccelerateHS/acceleratehs.github.io/raw/master/papers/acc-multidev-fhpc2015.bib">bibtex</a></li>
</ul>
<h3 id="embedding-foreign-code">Embedding Foreign Code</h3>
<p>Robert Clifton-Everest, Trevor L. McDonell, Manuel M. T. Chakravarty, and Gabriele Keller.</p>
<p>In <a href="http://www.ist.unomaha.edu/padl2014/"><em>PADL '14: The 16th International Symposium on Practical Aspects of Declarative Languages</em></a>, Springer-Verlag, LNCS, 2014.</p>
<p><strong>Abstract:</strong></p>
<p>Special purpose embedded languages facilitate generating high-performance code from purely functional high-level code; for example, we want to program highly parallel GPUs without the usual high barrier to entry and the time-consuming development process. We previously demonstrated the feasibility of a skeleton-based, generative approach to compiling such embedded languages.</p>
<p>In this paper, we (a) describe our solution to some of the practical problems with skeleton-based code generation and (b) introduce our approach to enabling interoperability with native code. In particular, we show, in the context of a functional embedded language for GPU programming, how template meta programming simplifies code generation and optimisation. Furthermore, we present our design for a foreign function interface for an embedded language.</p>
<ul>
<li><a href="https://github.com/AccelerateHS/acceleratehs.github.io/raw/master/papers/acc-ffi-padl2014.pdf">PDF</a> (16 pages)</li>
<li><a href="https://github.com/AccelerateHS/acceleratehs.github.io/raw/master/papers/acc-ffi-padl2014.bib">bibtex</a></li>
</ul>
<h3 id="optimising-purely-functional-gpu-programs">Optimising Purely Functional GPU Programs</h3>
<p>Trevor L. McDonell, Manuel M. T. Chakravarty, Gabriele Keller, and Ben Lippmeier.</p>
<p>In <a href="http://www.icfpconference.org/icfp2013/"><em>ICFP '13: The 18th ACM SIGPLAN International Conference on Functional Programming</em></a>, ACM, 2013.</p>
<p><strong>Abstract:</strong></p>
<p>Purely functional, embedded array programs are a good match for SIMD hardware, such as GPUs. However, the naive compilation of such programs quickly leads to both code explosion and an excessive use of intermediate data structures. The resulting slow-down is not acceptable on target hardware that is usually chosen to achieve high performance.</p>
<p>In this paper, we discuss two optimisation techniques, <em>sharing recovery</em> and <em>array fusion</em>, that tackle code explosion and eliminate superfluous intermediate structures. Both techniques are well known from other contexts, but they present unique challenges for an embedded language compiled for execution on a GPU. We present novel methods for implementing sharing recovery and array fusion, and demonstrate their effectiveness on a set of benchmarks.</p>
<ul>
<li><a href="https://github.com/AccelerateHS/acceleratehs.github.io/raw/master/papers/acc-optim-icfp2013.pdf">PDF</a> (12 pages)</li>
<li><a href="https://speakerdeck.com/tmcdonell/optimising-purely-functional-gpu-programs">slides</a></li>
<li><a href="https://github.com/AccelerateHS/acceleratehs.github.io/raw/master/papers/acc-optim-icfp2013.bib">bibtex</a></li>
</ul>
<h3 id="accelerating-haskell-array-codes-with-multicore-gpus">Accelerating Haskell Array Codes with Multicore GPUs</h3>
<p>Manuel M. T. Chakravarty, Gabriele Keller, Sean Lee, Trevor L. McDonell, and Vinod Grover.</p>
<p>In <a href="http://damp2011.cs.uchicago.edu/"><em>DAMP '11: Declarative Aspects of Multicore Programming</em></a>, ACM, 2011.</p>
<p><strong>Abstract:</strong></p>
<p>Current GPUs are massively parallel multicore processors optimised for workloads with a large degree of SIMD parallelism. Good performance requires highly idiomatic programs, whose development is work intensive and requires expert knowledge.</p>
<p>To raise the level of abstraction, we propose a domain-specific high-level language of array computations that captures appropriate idioms in the form of collective array operations. We embed this purely functional array language in Haskell with an online code generator for NVIDIA's CUDA GPGPU programming environment. We regard the embedded language's collective array operations as algorithmic skeletons; our code generator instantiates CUDA implementations of those skeletons to execute embedded array programs.</p>
<p>This paper outlines our embedding in Haskell, details the design and implementation of the dynamic code generator, and reports on initial benchmark results. These results suggest that we can compete with moderately optimised native CUDA code, while enabling much simpler source programs.</p>
<ul>
<li><a href="https://github.com/AccelerateHS/acceleratehs.github.io/raw/master/papers/acc-cuda-damp2011.pdf">PDF</a> (12 pages)</li>
<li><a href="https://github.com/AccelerateHS/acceleratehs.github.io/raw/master/papers/acc-cuda-damp2011.bib">bibtex</a></li>
</ul>
</div>
</div>
</div>
<footer class="footer">
<div class="container text-center">
<p class="text-muted">
<a href="https://github.com/AccelerateHS/accelerate">GitHub Project</a> ·
<a href="https://github.com/AccelerateHS/accelerate/issues">Issue Tracker</a> ·
<a href="http://groups.google.com/group/accelerate-haskell">Mailing List</a>
</p>
</div>
</footer>
</body>
</html>