Presentation is loading. Please wait.

Presentation is loading. Please wait.

EECE 571e (Fall 2015) (Massively) Parallel Computing Platforms Matei Ripeanu ece.ubc.ca.

Similar presentations


Presentation on theme: "EECE 571e (Fall 2015) (Massively) Parallel Computing Platforms Matei Ripeanu ece.ubc.ca."— Presentation transcript:

1 EECE 571e (Fall 2015) (Massively) Parallel Computing Platforms Matei Ripeanu matei @ ece.ubc.ca

2 Contact Info Course page: http://www.ece.ubc.ca/~matei/EECE571/ Office hours: after each class OR by appointment (email me) OR when my office door is open (just stop by) Email: matei @ ece.ubc.ca Office: KAIS 4033

3 EECE 571e: Course Goals  Primary –Gain an understanding of fundamental issues that affect the design of: Parallel applications Massively multicore processors –Survey the main current research themes, hardware and software trends.  Secondary –By studying a set of outstanding papers, build knowledge of how to do & present research –Learn how to read research papers & evaluate ideas

4 What I’ll Assume You Know You are familiar with system-level C programming, –Basic notions of parallelism, synchronizations, Computer Architecture Basics If there are things that don’t make sense, ask!

5 Schedule Weekly schedule (tentative) Note: the dates are tentative –But always on Mondays 5-8pm, KAIS 4018 –Possible travel: Nov 16 th – we’ll use some other day for a make- up class(Dec 7 th ?). Department ‘distinguished lecture’ series: Mondays 4pm –When related we’ll join these

6 Course Organization/Syllabus/etc.

7 Administravia: Course structure Lectures Paper reviews and discussion ( most classes ) Student projects –Aim high! Have fun! It’s a class project, not your PhD! –Teams of up to 3 students –Project presentations at the end of the term

8 Administravia: Grading Paper reviews:25% Class participation 15% Discussion leading: 5% Project: 50%

9 Administravia: Paper Reviewing (1) Goals: –Think of what you read before the class –Expand your knowledge beyond the papers that are assigned (read related material) –Get used to writing paper reviews / prepare to write your own papers Reviews due by 12am the day of the class (this is Monday at noon) –This gives you a chance to read everyone else’s reviews before the class –Gives discussion leader a chance to prepare. Have an eye on the writing style / Be professional in your writing –Clarity & brevity –Logic arguments –Detect (and stay away from) trivial claims. E.g., 1 st sentence in the Introduction: “The tremendous/unprecedented/phenomenal growth/scale/ubiquity of the Internet…”

10 Administravia: Paper Reviewing (2) Follow the form provided when relevant. Summarize the main contribution of the paper Critique the main contribution: –Significance: Discuss the significance of the paper and explain your rating in a couple of sentences. [Rating: 5 ( breakthrough ), 4 ( significant contribution ), 3 ( modest contribution ), 2 ( incremental ), 1 ( no contribution or negative contribution ). ] The discussion and your argument are more important than the rating itself –Discuss how convincing the methodology is. Do the claims and conclusions follow from the experiments? Are the assumptions realistic? Are the experiments well designed? Are there different experiments that would be more convincing? Are there alternatives the authors should have considered? (And, of course, is the paper free of methodological errors?) –What is the most important limitation of the approach?

11 Administravia: Paper Reviewing (3) What are two/three strongest and/or most interesting ideas in the paper? What’s there to remember? What are two/three most striking weaknesses in the paper? Name two questions that you would like to ask the authors. Detail an interesting extension to the work (not mentioned in the future work section). Optional comments on the paper that you’d like to see discussed in class.

12 Administravia: Discussions Leading Come prepared! –Background and related work –Prepare a 5-10 minute background (position the paper vs. related work) and summarize the paper With slides if you want Keep in mind who is your audience: they have already read the paper! –Prepare a discussion outline –Prepare questions: “What if”s; e.g., What if you drop this assumption? “What else”: What else could the authors have done to prove/disprove the same hypothesis Unclear aspects of the solution proposed … –Similar ideas in different contexts –Initiate short brainstorming sessions Leaders do NOT need to submit paper reviews Your goals: (1) Keep discussion relevant, (2) Keep discussion flowing (3) Engage everybody

13 Administravia: Projects It’s just a class, not your PhD. Aim high! –Goal: With one/two extra months of work you project should be ready to be submitted to a decent workshop / short conference paper –It is doable! Combine with your own research if relevant to this course –You can get more than twice as many results with less than twice as much work –Get informal approval from all instructors if you overlap final projects: Don’t sell the same piece of work twice Past projects available form course webpage

14 Administravia: Project timeline (tentative) 3 rd week – 3-slide idea presentation: –S1: “What is the (research) question you aim to answer?”, “What is the hypothesis you try to prove” –S2: “Why is this important / relevant?” & –S3: “How do you plan to go ahead?” 5 th week: 1-2 pages (ACM format) project proposal –Address above questions, have a clear image of what’s doable, define success. –Think of this as a ‘contract’ 8 th week: 3-4 page Midterm report –Report preliminary results, Include related work Final week [see schedule] in-class project presentation –PresentationDemo, if appropriate –5-6-page preliminary write-up. Final project report at the end of the term ACM formatting guidelines:

15 Project example ideas Past project reports: http://www.ece.ubc.ca/~matei/EECE571/PastReports/ http://www.ece.ubc.ca/~matei/EECE571/PastReports/ Some directions: –Port a ‘challenging’ application to a ‘new’ architecture. challenging application: one that does not have good parallel structure – e.g., graph processing, genomic read alignment; one that’s hard to load-balance – e.g., MRI e.g., a graph processing (BFS, or Totem engine) on AMD Fusion –Characterize a new platform (e.g., nVidia Jetson) –Model a hardware extension e.g., add NVRAM to a discrete GPU. Estimate performance, energy impact –Take an existing code and optimize it. Implement existing algorithm e.g., graph clustering on discrete GPU 15 5/27/2016

16 Next Class (Monday – September 21 st ) Note room change: KAIS 4018 Discussion of –Project ideas –Papers To do: Subscribe to Piazza (see instructions on class webpage) Come prepared with 2+project ideas to discuss Two papers to review

17 Questions?


Download ppt "EECE 571e (Fall 2015) (Massively) Parallel Computing Platforms Matei Ripeanu ece.ubc.ca."

Similar presentations


Ads by Google