📄 http:^^www.cs.cornell.edu^info^courses^fall-94^cs617^
字号:
MIME-Version: 1.0
Server: CERN/3.0
Date: Sunday, 24-Nov-96 22:04:08 GMT
Content-Type: text/html
Content-Length: 2989
Last-Modified: Sunday, 13-Nov-94 17:06:01 GMT
<html><head><title>CS617 Home Page</title></head><body><!WA0><!WA0><!WA0><!WA0><!WA0><!WA0><!WA0><!WA0><!WA0><!WA0><!WA0><!WA0><!WA0><!WA0><!WA0><!WA0><img aling=bottom src="http://www.cs.cornell.edu/Info/Courses/Fall-94/CS617/CS617_img.GIF"><h1>Frontiers of Parallel Systems</h1> <h2>Thorsten von Eicken</h2><h3>Fall, 1994</h3><hr><br><b>Location:</b> 111 Upson, MWF 11:05am-12:00pm<p><b>Office hours:</b> Mon 12:15pm-1:15pm, Thu 2pm-3pm<p><hr><h1>Course Description</h1>Parallel machines are here to stay. This is underscored by the fact that all system manufacturers offer multiprocessors at the top of their product line. However, the debate on how the parallel machines of the future will look like has heated-up considerably in the past few months: the federal spending cuts are on their way to eradicate the performance-at-any-price massively parallel processor (MPP) manufacturers. Competitors having bet on glorified workstation farms are smiling: while their machines cannot offer the same level of performance or ease of use, their business has not suffered as much because these machines are more price competitive and can leverage the latest microprocessor developments more quickly. At the core of the technological debate lie a large number of systems issues: how to integrate a large number of off-the-shelf processors into a cost-effective system which can be easily programmed in high-level parallel programming languages and which can host a varied application workload.<p>This course is not about how to program parallel machines (although some of that will be the topic of the first few weeks). Parallel algorithms, languages and architectures have matured considerably over the last few years to the point where parallel machines are almost usable, if only the system support were adequate to allow general-purpose use. This course will focus on the architecture and operating system aspects required to support features taken for granted in sequential computing such as portable parallel programs, powerful debuggers, multi-user machine access, virtual memory, and fast I/O.<p>The first part of the course will examine two complete parallel systems: the CM-5 with Split-C and an ATM network of workstations with CC++. Using a ``vertical'' approach we will study the interactions between algorithmic models, languages, associated execution models, operating systems, architectures and hardware implementations, focussing on the support required for each of these layers. The second part of the course will focus on specific topics and slice ``horizontally'' across systems, selecting a few key issues for an analysis of the design alternatives. An in-depth study of the KSR and the DASH will provide some shared-memory contrast to the course.<p><hr><h1>Course Materials</h1><UL><li><h2><!WA1><!WA1><!WA1><!WA1><!WA1><!WA1><!WA1><!WA1><!WA1><!WA1><!WA1><!WA1><!WA1><!WA1><!WA1><!WA1><a href="http://www.cs.cornell.edu/Info/Courses/Fall-94/CS617/CS617-format.html">Course format</a></h2><li><h2><!WA2><!WA2><!WA2><!WA2><!WA2><!WA2><!WA2><!WA2><!WA2><!WA2><!WA2><!WA2><!WA2><!WA2><!WA2><!WA2><a href="http://www.cs.cornell.edu/Info/Courses/Fall-94/CS617/CS617-notes.html">Lecture Notes</a></h2> <li><h2><!WA3><!WA3><!WA3><!WA3><!WA3><!WA3><!WA3><!WA3><!WA3><!WA3><!WA3><!WA3><!WA3><!WA3><!WA3><!WA3><a href="http://www.cs.cornell.edu/Info/Courses/Fall-94/CS617/CS617-problm.html">Problem sets</a></h2> <li><h2><!WA4><!WA4><!WA4><!WA4><!WA4><!WA4><!WA4><!WA4><!WA4><!WA4><!WA4><!WA4><!WA4><!WA4><!WA4><!WA4><a href="http://www.cs.cornell.edu/Info/Courses/Fall-94/CS617/CS617-proj.html">Term projects</a></h2></UL><hr> </html>⌨️ 快捷键说明
复制代码
Ctrl + C
搜索代码
Ctrl + F
全屏模式
F11
切换主题
Ctrl + Shift + D
显示快捷键
?
增大字号
Ctrl + =
减小字号
Ctrl + -