EE382V -- Computer Architecture: User System Interplay

Spring 2007
Working Page

Vital Information

Class will meet on Mondays and Wednesdays 12:30 - 2:00pm in ENS 116.

The class descriptor and syllabus can be found at http://www.ece.utexas.edu/~merez/EE382V/descriptor.html. Additional material may occasionally be made available on Blackboard.

Instructors

Professor:
Mattan Erez
ENS 538
mattan.erez@mail.utexas.edu
471-7846
Office Hours: TW 2PM - 3PM or by appointment

TA:
Min Kyu Jeong
Office TBA
mjeong@ece.utexas.edu

Office Hours: ThF 2PM - 3PM

Announcements

Please don't email me or Min any of your homework assignments. We will only accept submissions through blackboard (04/12/2007).
Sixteenth reading is out and due on 4/16/2007. There are two papers. The first is long but has no results, and the second has results and should go quickly aftedr reading the first. (4/12/2007).
Sixteenth reading is out and due on 4/11/2007 (4/09/2007).
Fifteenth reading is out and due on 4/4/2007 (3/30/2007).
No reading for Monday 4/2/2007! I'll do an actual lecture (3/30/2007).
Thirteenth reading is out (3/20/2007).
Added some background for the twelfth reading that should help explain Id and set the tone a bit (TAM was during the era of dataflow architectures and suggested a different way of dealing with extremely fine-grained parallelism (3/17/2007).
Twelfth reading is out (3/13/2007).
Eleventh reading is out (3/08/2007).
Tenth reading is out and this time please do the prep writeup (2/28/2007).
Ninth reading is out, and remember, no writeup due. Instead we'll discuss the paper briefly and will also take a look back at what we covered so far in class (2/23/2007).
Eighth reading is due on 2/21/2007 (2/17/2007).
When answering the indented users question, please try to list users who might actually use the described system and not just readers that the paper targets (e.g., architects). well (not just users of the idea).
Seventh reading is due on 2/19/2007 (2/14/2007).
Fifth and sixth reading assignments are out (2/8/2007).
New and improved description of the class preparation assignments is now available (2/2/2007).
Fourth reading assignment (two slightly shorter papers) is due on Wednesday 02/07/2007 (2/2/2007).
First scribe notes for the second lecture are out (01/31/2007).
Third reading assignment is due on Monday 02/05/2007 (01/30/2007).
EE Times has a related article you may be interested in reading (01/26/2007).
Second reading assignment is due on Wednesday 01/31/2007 (01/26/2007).
If you did not get email with these announcements (and would like to be on the list) please let me know (01/25/2007).
On-going anonymous survey is up at https://web.austin.utexas.edu/diia/oca/ (01/25/2007).
Office hours changed (see above) (01/25/2007).
First reading assignment is due on Monday 01/29/2007 (01/25/2007).

Surveys

On-going anonymous survey is up at https://web.austin.utexas.edu/diia/oca/. Please use it as often as you feel like to let me know how the class is going. Please feel free to ignore any of the (too many) questions I placed there. The survey will be refreshed once a week roughly.

Active Schedule (with Reading)

This active class schedule will show topics and reading assignments for the various classes as well as important comments about the reading or activities.

Reading material, whenever possible, will be publicly available links or links to digital libraries that can be accessed from the UT Austin network (or from off-campus using Internet Explorer and the provided link).

Please check below for what class the reading is due and make sure to have prepared and submitted your preparation writeup before class starts.

Class	Date	Topic	Reading Due	Comments	Scribe Notes
1	01/22/2007	Description and procedures			descriptor.html
2	01/24/2007	User Requirements and System Evaluation			lect02.pdf
3	01/29/2007	GPP binary compatibility	J. Denhert et al., "The Transmeta Code Morphing(TM) Software: using speculation, recovery, and adaptive retranslation to address real-life challenges" (off-campus)		lect03.pdf
4	01/31/2007	GPP source-level compatibility	S Swanson et al., "WaveScalar"		lect04.pdf
5	02/05/2007	GPP virtualization	BACKGROUND (not required): M. Rosenblum, "The Reincarnation of Virtual Machines", ACM Queue vol. 2, no. 5. Required: K. Adams and O. Agesen, "A comparison of software and hardware techniques for x86 virtualization", ASPLOS-XII (off-campus).		lect05.pdf
6	02/07/2007	Design for Manufacturability and Defect Tolerance	SHORT INTRO (required)Y. Zorian and Z. Gizopoulos, "Guest Editor's Introduction: Design for Yield and Reliability", IEEE Design and Test, Vol. 21, No. 3, May 2004 (off-campus). W. Culbertson et al., "Defect Tolerance on the Teramac Custom Computer", IEEE Symposium on FPGAs for Custom Computing Machines, 1997 (off-campus).		lect06.pdf
7	02/12/2007	OS Reliability	E. Witchel et al., "Mondrix: Memory Isolation for Linux using Mondriaan Memory Protection ", Proceedings of the twentieth ACM symposium on Operating systems principles, 2005 (off-campus).		lect07.pdf
8	02/14/2007	Soft-Error Fault Tolerance	M. Erez et al., "Fault Tolerance Techniques for the Merrimac Streaming Supercomputer ", Proceedings of SC'05, 2005.	Recommended reading: S. Mitra et al., "Robust system design with built-in soft-error resilience"(off campus).	lect08.pdf
9	02/19/2007	Low power and Circuit Variability	D. Ernst et al., "Razor: A Low-Power Pipeline Based on Circuit-Level Timing Speculation", MICRO-36, 2003 (off-campus).		lect09.pdf
10	02/21/2007	Virtual Memory and Evaluation	D. W. Clark and J. S. Emer., "Performance of the VAX-11/780 translation buffer: simulation and measurement", ACM TOCS, 1985 (off-campus).		lect10.pdf
11	02/26/2007	Power and Quiz Prep.	T. Mudge, "Power: A First-Class Architectural Design Constraint", IEEE Computer, Vol 34., No. 4, 2001.
13	03/05/2007	Compilation for Parallelism	A. Lim and M. Lam, "Maximizing Parallelism and Minimizing Synchronization with Affine Transforms", Symposium on Principles of Programming Languages, January 1997.
14	03/07/2007	Parallel Architectures	D. Patterson et al., "The Berkeley View of Parallelism", 2007. REQUIRED: video, slides, (not required) web page.
15	03/19/2007	J-Machine Message-Driven Parallel Systen	M. Noakes et al., "The J-Machine Multicomputer: An Architecural Evaluation", ISCA 20, 1993. PLEASE ALSO READ THIS OVERVIEW WITH PICTURES W. J. Dally et al., "The J-Machine: A Retrospective", 1998.
16	03/21/2007	Lightweight Parallel Threads	D. E. Culler et al., "Fine-Grain Parralelism with Minimal Hardware Support: A Compiler-Controlled Threaded Abstract Machine",ASPLOS IV, 1991 (off-campus). BACKGROUND ON ID: K. R. Traub, "A Compiler for the MIT Tagged-Token Dataflow Architecture (pages 13--21 only)", MIT Masters Thesis, 1986.
17	03/26/2007	Cache Coherence	Lenoski et al., "The Stanford DASH Multiprocessor", IEEE Computer, Volume 25 , Issue 3 (March 1992).		lect17.pdf
18	03/28/2007	Multicore Processors	L. A. Barroso et al., "Piranha: A Scalable Architecture Based on Single-Chip Multiprocessing", ISCA 27, 2000. B. Sander, "Barcelona: AMD's Next-Generation Quad-core Microprocessor", ACES AVAYA Auditorium, 3/28/2007 1 - 2pm.		lect18.pdf
19	04/02/2007	More Parallel Architectures (including GPUs)	NO READING
21	04/09/2007	Parallel Programming with X10	P. Charles et al., "X10: An Object-Oriented Approach to Non-Uniform Cluster Computing", OOPSLA 20, 2005.
22	04/11/2007	Another PGAS example -- Titanium	K. Datta et al., "Titanium Performance and Potential: an NPB Experimental Study", LCPC 2005.
23	04/16/2007	IBM BlueGene/L Supercomputer	Nicer figures and description but no results: A. Gara et al., "An Overview of the BlueGene/L System Arcyitecture, IBM Journal of R&D, Vol. 49 No. 2/3, 2005. K. Davis et al., "A Performance and Scalability Analysis of the BlueGene/L Architecture", SC'04.
25	04/23/2007	OS for multi-processors (transactional memory focus)	Overview of TM: A-R Adl-Tabatabai et al., "Unlocking Concurrency: Multicore programming with transactional memory", ACM Queue, Volume 4 , Issue 10, 2006. H. Ramadan et al., "MetaTM/TxLinux: Transactional Memory For An Operating System", ISCA 2007.	Papers that we decided not to read this time: H. Chapin et al., "Hive: fault containment for shared-memory multiprocessors", SOSP 1995. K. Govil et al., "Disco: Running Commodity Operating Systems on Scalable Multiprocessors", ACM TOCS, Vol. 15, No. 4, 1997.

Useful Links and Notes

Google Scholar for finding papers and forward citations.
CiteSeer is an excellent citation tool for backward and forward citations for topics in computer science.

EE382V -- Computer Architecture: User System Interplay

Spring 2007 Working Page

Spring 2007
Working Page