weekly report ph.d. student: leo lee date: oct. 9, 2009
Post on 15-Jan-2016
224 views
TRANSCRIPT
![Page 1: Weekly Report Ph.D. Student: Leo Lee date: Oct. 9, 2009](https://reader030.vdocuments.net/reader030/viewer/2022033023/56649d265503460f949fd1f3/html5/thumbnails/1.jpg)
Weekly Report
Ph.D. Student: Leo Leedate: Oct. 9, 2009
![Page 2: Weekly Report Ph.D. Student: Leo Lee date: Oct. 9, 2009](https://reader030.vdocuments.net/reader030/viewer/2022033023/56649d265503460f949fd1f3/html5/thumbnails/2.jpg)
Outline
• Courses
• Research
• Work plan
![Page 3: Weekly Report Ph.D. Student: Leo Lee date: Oct. 9, 2009](https://reader030.vdocuments.net/reader030/viewer/2022033023/56649d265503460f949fd1f3/html5/thumbnails/3.jpg)
Outline
• Courses
• Research
• Work plan
![Page 4: Weekly Report Ph.D. Student: Leo Lee date: Oct. 9, 2009](https://reader030.vdocuments.net/reader030/viewer/2022033023/56649d265503460f949fd1f3/html5/thumbnails/4.jpg)
Courses
• Data mining– Homework;
– Hidden Markov Model
– Read the most classical tutorial;• Forward-backward procedure;• Viterbi algorithm;
![Page 5: Weekly Report Ph.D. Student: Leo Lee date: Oct. 9, 2009](https://reader030.vdocuments.net/reader030/viewer/2022033023/56649d265503460f949fd1f3/html5/thumbnails/5.jpg)
Courses
• Network security– Check the homework;
– Modify the tutorial for next week;• Learn C# ;• Dev. an easy chat application.
![Page 6: Weekly Report Ph.D. Student: Leo Lee date: Oct. 9, 2009](https://reader030.vdocuments.net/reader030/viewer/2022033023/56649d265503460f949fd1f3/html5/thumbnails/6.jpg)
Outline
• Courses
• Research
• Work plan
![Page 7: Weekly Report Ph.D. Student: Leo Lee date: Oct. 9, 2009](https://reader030.vdocuments.net/reader030/viewer/2022033023/56649d265503460f949fd1f3/html5/thumbnails/7.jpg)
Research
![Page 8: Weekly Report Ph.D. Student: Leo Lee date: Oct. 9, 2009](https://reader030.vdocuments.net/reader030/viewer/2022033023/56649d265503460f949fd1f3/html5/thumbnails/8.jpg)
Mars: A MapReduce Framework on Graphics Processors
• Introduction – For search engines and other web server
applications, high performance is essential.
– The MapReduce framework is a successful paradigm to support such data processing applications, which reduces the complexity of parallel programming.
– Encouraged by the success of the CPU-based MapReduce frameworks, we develop Mars, a MapReduce framework on graphics processors, or GPUs.
![Page 9: Weekly Report Ph.D. Student: Leo Lee date: Oct. 9, 2009](https://reader030.vdocuments.net/reader030/viewer/2022033023/56649d265503460f949fd1f3/html5/thumbnails/9.jpg)
Mars: A MapReduce Framework on Graphics Processors
• Introduction
– Since GPUs are traditionally designed as special-purpose co-processors for gaming applications, their languages lack support for some basic programming constructs.
• variable-length data types;• more complex functions such as recursion.
– GPU architectural details are highly vendor-specific and programmers have limited access to these details.
– All these factors make the GPU programming a difficult task in general and more so for complex tasks such as web data analysis. Therefore, we propose to develop a MapReduce framework on the GPU so that programmers can easily harness the GPU computation power for their data processing tasks.
![Page 10: Weekly Report Ph.D. Student: Leo Lee date: Oct. 9, 2009](https://reader030.vdocuments.net/reader030/viewer/2022033023/56649d265503460f949fd1f3/html5/thumbnails/10.jpg)
Mars: A MapReduce Framework on Graphics Processors
• Introduction– First, the synchronization overhead must be low so
that the system can scale to hundreds of processors.
– Second, due to the lack of dynamic thread scheduling on current GPUs, it is essential to allocate work evenly across threads on the GPU to exploit its massive thread parallelism.
– Third, the core tasks of MapReduce programs, including string processing, file manipulation and concurrent reads and writes, are unconventional to GPUs and must be handled efficiently.
![Page 11: Weekly Report Ph.D. Student: Leo Lee date: Oct. 9, 2009](https://reader030.vdocuments.net/reader030/viewer/2022033023/56649d265503460f949fd1f3/html5/thumbnails/11.jpg)
Mars: A MapReduce Framework on Graphics Processors
• Preliminaries and overview– GPUs
– GPGPU
– MapReduce
![Page 12: Weekly Report Ph.D. Student: Leo Lee date: Oct. 9, 2009](https://reader030.vdocuments.net/reader030/viewer/2022033023/56649d265503460f949fd1f3/html5/thumbnails/12.jpg)
Mars: A MapReduce Framework on Graphics Processors
• Design and implementation
– Ease of programming. Ease of programming encourages developers to use the GPU for their tasks.
– Performance. The overall performance of our GPU-based MapReduce should be comparable to or better than that of the state-of-the-art CPU counterparts.
![Page 13: Weekly Report Ph.D. Student: Leo Lee date: Oct. 9, 2009](https://reader030.vdocuments.net/reader030/viewer/2022033023/56649d265503460f949fd1f3/html5/thumbnails/13.jpg)
Mars: A MapReduce Framework on Graphics Processors
• Design and implementation-APIs– User-implemented
![Page 14: Weekly Report Ph.D. Student: Leo Lee date: Oct. 9, 2009](https://reader030.vdocuments.net/reader030/viewer/2022033023/56649d265503460f949fd1f3/html5/thumbnails/14.jpg)
Mars: A MapReduce Framework on Graphics Processors
• Design and implementation-APIs– System-provided
![Page 15: Weekly Report Ph.D. Student: Leo Lee date: Oct. 9, 2009](https://reader030.vdocuments.net/reader030/viewer/2022033023/56649d265503460f949fd1f3/html5/thumbnails/15.jpg)
Mars: A MapReduce Framework on Graphics Processors
• System Workflow and Configuration
![Page 16: Weekly Report Ph.D. Student: Leo Lee date: Oct. 9, 2009](https://reader030.vdocuments.net/reader030/viewer/2022033023/56649d265503460f949fd1f3/html5/thumbnails/16.jpg)
Mars: A MapReduce Framework on Graphics Processors
• Optimization Techniques– Coalesced accesses
– Accesses using built-in vector types: char4 and int4?
![Page 17: Weekly Report Ph.D. Student: Leo Lee date: Oct. 9, 2009](https://reader030.vdocuments.net/reader030/viewer/2022033023/56649d265503460f949fd1f3/html5/thumbnails/17.jpg)
Mars: A MapReduce Framework on Graphics Processors
• Experimental evaluation
![Page 18: Weekly Report Ph.D. Student: Leo Lee date: Oct. 9, 2009](https://reader030.vdocuments.net/reader030/viewer/2022033023/56649d265503460f949fd1f3/html5/thumbnails/18.jpg)
Mars: A MapReduce Framework on Graphics Processors
• Experimental evaluation
![Page 19: Weekly Report Ph.D. Student: Leo Lee date: Oct. 9, 2009](https://reader030.vdocuments.net/reader030/viewer/2022033023/56649d265503460f949fd1f3/html5/thumbnails/19.jpg)
Mars: A MapReduce Framework on Graphics Processors
• Experimental evaluation
![Page 20: Weekly Report Ph.D. Student: Leo Lee date: Oct. 9, 2009](https://reader030.vdocuments.net/reader030/viewer/2022033023/56649d265503460f949fd1f3/html5/thumbnails/20.jpg)
Mars: A MapReduce Framework on Graphics Processors
• Experimental evaluation
![Page 21: Weekly Report Ph.D. Student: Leo Lee date: Oct. 9, 2009](https://reader030.vdocuments.net/reader030/viewer/2022033023/56649d265503460f949fd1f3/html5/thumbnails/21.jpg)
Mars: A MapReduce Framework on Graphics Processors
• Experimental evaluation
![Page 22: Weekly Report Ph.D. Student: Leo Lee date: Oct. 9, 2009](https://reader030.vdocuments.net/reader030/viewer/2022033023/56649d265503460f949fd1f3/html5/thumbnails/22.jpg)
Mars: A MapReduce Framework on Graphics Processors
• Experimental evaluation
![Page 23: Weekly Report Ph.D. Student: Leo Lee date: Oct. 9, 2009](https://reader030.vdocuments.net/reader030/viewer/2022033023/56649d265503460f949fd1f3/html5/thumbnails/23.jpg)
Mars: A MapReduce Framework on Graphics Processors
• Experimental evaluation
![Page 24: Weekly Report Ph.D. Student: Leo Lee date: Oct. 9, 2009](https://reader030.vdocuments.net/reader030/viewer/2022033023/56649d265503460f949fd1f3/html5/thumbnails/24.jpg)
Mars: A MapReduce Framework on Graphics Processors
• Experimental evaluation
![Page 25: Weekly Report Ph.D. Student: Leo Lee date: Oct. 9, 2009](https://reader030.vdocuments.net/reader030/viewer/2022033023/56649d265503460f949fd1f3/html5/thumbnails/25.jpg)
Mars: A MapReduce Framework on Graphics Processors
![Page 26: Weekly Report Ph.D. Student: Leo Lee date: Oct. 9, 2009](https://reader030.vdocuments.net/reader030/viewer/2022033023/56649d265503460f949fd1f3/html5/thumbnails/26.jpg)
Outline
• Courses
• Research
• Work plan
![Page 27: Weekly Report Ph.D. Student: Leo Lee date: Oct. 9, 2009](https://reader030.vdocuments.net/reader030/viewer/2022033023/56649d265503460f949fd1f3/html5/thumbnails/27.jpg)
Work Plan
• Go on paper reading
• Learn more CUDA applications
• Work hard on data mining, try to implement some classical algorithm
• Learn C#
![Page 28: Weekly Report Ph.D. Student: Leo Lee date: Oct. 9, 2009](https://reader030.vdocuments.net/reader030/viewer/2022033023/56649d265503460f949fd1f3/html5/thumbnails/28.jpg)
• Thanks for your listening