<?xml version="1.0" encoding="UTF-8"?>
<feed xmlns="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <title>DRUM Collection: Electrical &amp; Computer Engineering Research Works</title>
  <link rel="alternate" href="http://hdl.handle.net/1903/1658" />
  <subtitle />
  <id>http://hdl.handle.net/1903/1658</id>
  <updated>2013-05-19T20:50:49Z</updated>
  <dc:date>2013-05-19T20:50:49Z</dc:date>
  <entry>
    <title>Empirical Speedup Study of Truly Parallel Data Compression</title>
    <link rel="alternate" href="http://hdl.handle.net/1903/13890" />
    <author>
      <name>Edwards, James A.</name>
    </author>
    <author>
      <name>Vishkin, Uzi</name>
    </author>
    <id>http://hdl.handle.net/1903/13890</id>
    <updated>2013-05-04T02:32:43Z</updated>
    <published>2013-04-20T00:00:00Z</published>
    <summary type="text">Title: Empirical Speedup Study of Truly Parallel Data Compression
Authors: Edwards, James A.; Vishkin, Uzi
Abstract: We present an empirical study of novel work-optimal parallel&#xD;
algorithms for Burrows-Wheeler compression and decompression&#xD;
of strings over a constant alphabet. To validate&#xD;
these theoretical algorithms, we implement them on the experimental&#xD;
XMT computing platform developed especially&#xD;
for supporting parallel algorithms at the University of Maryland.&#xD;
We show speedups of up to 25x for compression, and&#xD;
13x for decompression, versus bzip2, the de facto standard&#xD;
implementation of Burrows-Wheeler compression. Unlike&#xD;
existing approaches, which assign an entire (e.g., 900KB)&#xD;
block to a processor that processes the block serially, our&#xD;
approach is “truly parallel” as it processes in parallel the&#xD;
entire input. Besides the theoretical interest in solving the&#xD;
“right” problem, the importance of data compression speed&#xD;
for small inputs even at great expense of quality (compressed&#xD;
size of data) is demonstrated by the introduction of Google’s&#xD;
Snappy for MapReduce. Perhaps surprisingly, we show feasibility&#xD;
of holding on to quality, while even beating Snappy&#xD;
on speed.&#xD;
In turn, this work adds new evidence in support of the&#xD;
XMT/PRAM thesis: that an XMT-like many-core hardware/&#xD;
software platform may be necessary for enabling general-purpose&#xD;
parallel computing. Comparison of our results to recently&#xD;
published work suggests 70x improvement over what&#xD;
current commercial parallel hardware can achieve.</summary>
    <dc:date>2013-04-20T00:00:00Z</dc:date>
  </entry>
  <entry>
    <title>Complex life forms may arise from electrical  processes</title>
    <link rel="alternate" href="http://hdl.handle.net/1903/13369" />
    <author>
      <name>Elson, Edward C</name>
    </author>
    <id>http://hdl.handle.net/1903/13369</id>
    <updated>2013-01-11T03:51:35Z</updated>
    <published>2010-06-24T00:00:00Z</published>
    <summary type="text">Title: Complex life forms may arise from electrical  processes
Authors: Elson, Edward C
Abstract: There is still not an appealing and testable model to explain how single-celled organisms, usually following fusion of male and female gametes, proceed to grow and evolve into multi-cellular, complexly differentiated systems, a particular species following virtually an invariant and unique growth pattern. An intrinsic electrical oscillator, resembling the cardiac pacemaker, may explain the process. Highly auto-correlated, it could live independently of ordinary thermodynamic processes which mandate increasing disorder, and could coordinate growth and differentiation of organ anlage.</summary>
    <dc:date>2010-06-24T00:00:00Z</dc:date>
  </entry>
  <entry>
    <title>A review of communication-oriented optical wireless systems</title>
    <link rel="alternate" href="http://hdl.handle.net/1903/13347" />
    <author>
      <name>Borah, Deva K</name>
    </author>
    <author>
      <name>Boucouvalas, Anthony C</name>
    </author>
    <author>
      <name>Davis, Christopher C</name>
    </author>
    <author>
      <name>Hranilovic, Steve</name>
    </author>
    <author>
      <name>Yiannopoulos, Konstantinos</name>
    </author>
    <id>http://hdl.handle.net/1903/13347</id>
    <updated>2013-01-11T03:53:38Z</updated>
    <published>2012-03-07T00:00:00Z</published>
    <summary type="text">Title: A review of communication-oriented optical wireless systems
Authors: Borah, Deva K; Boucouvalas, Anthony C; Davis, Christopher C; Hranilovic, Steve; Yiannopoulos, Konstantinos
Abstract: This article presents an overview of optical wireless (OW) communication systems that operate both in the short-&#xD;
(personal and indoor systems) and the long-range (outdoor and hybrid) regimes. Each of these areas is discussed&#xD;
in terms of (a) key requirements, (b) their application framework, (c) major impairments and applicable mitigation&#xD;
techniques, and (d) current and/or future trends. Personal communication systems are discussed within the context&#xD;
of point-to-point ultra-high speed data transfer. The most relevant application framework and related standards are&#xD;
presented, including the next generation Giga-IR standard that extends personal communication speeds to over 1&#xD;
Gb/s. As far as indoor systems are concerned, emphasis is given on modeling the dispersive nature of indoor OW&#xD;
channels, on the limitations that dispersion imposes on user mobility and dispersion mitigation techniques. Visible&#xD;
light communication systems, which provide both illumination and communication over visible or hybrid visible/&#xD;
infrared LEDs, are presented as the most important representative of future indoor OW systems. The discussion on&#xD;
outdoor systems focuses on the impact of atmospheric effects on the optical channel and associated mitigation&#xD;
techniques that extend the realizable link lengths and transfer rates. Currently, outdoor OW is commercially&#xD;
available at 10 Gb/s Ethernet speeds for Metro networks and Local-Area-Network interconnections and speeds are&#xD;
expected to increase as faster and more reliable optical components become available. This article concludes with&#xD;
hybrid optical wireless/radio-frequency (OW/RF) systems that employ an additional RF link to improve the overall&#xD;
system reliability. Emphasis is given on cooperation techniques between the reliable RF subsystem and the&#xD;
broadband OW system.</summary>
    <dc:date>2012-03-07T00:00:00Z</dc:date>
  </entry>
  <entry>
    <title>Parallel Algorithms for Burrows-Wheeler Compression and Decompression</title>
    <link rel="alternate" href="http://hdl.handle.net/1903/13299" />
    <author>
      <name>Edwards, James A.</name>
    </author>
    <author>
      <name>Vishkin, Uzi</name>
    </author>
    <id>http://hdl.handle.net/1903/13299</id>
    <updated>2012-11-14T03:31:24Z</updated>
    <published>2012-11-12T00:00:00Z</published>
    <summary type="text">Title: Parallel Algorithms for Burrows-Wheeler Compression and Decompression
Authors: Edwards, James A.; Vishkin, Uzi
Abstract: We present work-optimal PRAM algorithms for Burrows-Wheeler compression&#xD;
and decompression of strings over a constant alphabet. For a string of&#xD;
length n, the depth of the compression algorithm is O(log2 n), and the depth&#xD;
of the the corresponding decompression algorithm is O(log n). These appear&#xD;
to be the first polylogarithmic-time work-optimal parallel algorithms for any&#xD;
standard lossless compression scheme.&#xD;
The algorithms for the individual stages of compression and decompression&#xD;
may also be of independent interest: 1. a novel O(log n)-time, O(n)-work&#xD;
PRAM algorithm for Huffman decoding; 2. original insights into the stages of&#xD;
the BW compression and decompression problems, bringing out parallelism&#xD;
that was not readily apparent, allowing them to be mapped to elementary&#xD;
parallel routines that have O(log n)-time, O(n)-work solutions, such as: (i)&#xD;
prefix-sums problems with an appropriately-defined associative binary operator&#xD;
for several stages, and (ii) list ranking for the final stage of decompression.</summary>
    <dc:date>2012-11-12T00:00:00Z</dc:date>
  </entry>
</feed>

