Past Program, 2005

2005 P=ac2 Conference Program

__________________________________________________________________________________________

WEDNESDAY, SEPTEMBER 28
__________________________________________________________________________________________

WELCOMING REMARKS 8:15 - 8:45

Opening Address
Tilak Agerwala
IBM Research VP of Systems

KEYNOTE: 8:45 – 9:30

Alan Gara on Bluegene

SESSION 1: MULTIPROCESSOR ISSUES (Session Chair: P. Emma) 9:30 – 11:00

A Case for Increased Operating System Support in Chip Multi-Processors
David Nellans, Rajeev Balasubramonian, Erik Brunvand
University of Utah

Conditional Memory Ordering
Christoph von Praun and Harold W. Cain
IBM T.J. Watson Research Center

Architecture Cloning For PowerPC Processors
Edwin Chan, Raul Silvera, Roch Archambault
IBM Toronto Labs

SESSION 2: CACHE AND MEMORY (Session Chair: A. Hartstein) 11:20 – 12:50

A Near-Memory Processor for Vector, Streaming and Bit Manipulation Workloads
Mingliang Wei+, Marc Snir+, Josep Torrellas+, R. Brett Tremaine*
+ = Univ of Illinois at Urbana-Champaign * = IBM T.J. Watson Research Center

A Prediction Model for Alternative Cache Replacement Policies
Fei Guo and Yan Solihin,
North Carolina State University

Software-Hardware Cooperative Memory Disambiguation
Ruke Huang and Michael Huang
University of Rochester

SESSION 3: EFFICIENT THREAD EXTRACTION AND USE (Session Chair: D. Prener) 2:00 – 3:30

Scalable, Low-Complexity Instruction Schedulers for SMT Processors
Joseph J. Sharkey, Deniz Balkan, Dmitry V. Ponomarev
State University of New York at Binghamton

A New Approach to Thread Extraction for General-Purpose Programs
Guilherme Ottoni, Ram Rangan, Adam Stoler, David I. August
Princeton University

POSH: A Profiler-Enhanced TLS Compiler that Leverages Program Structure
Wei Liu+, James Tuck+, Luis Ceze+, Karin Strauss+, Jose Renau*, Josep Torrellas+
+ = Univ of Illinois at Urbana-Champaign, * = Univ of California at Santa Cruz

SESSION 4: COMPILATION POLICIES FOR ENERGY AND POWER (Session Chair: A. Buyuktosunoglu) 4:00 – 5:00

On the Energy Effectiveness of If-conversion in Superscalar Microprocessors
Eric Zimmerman, Craig Zilles
University of Illinois at Urbana-Champaign

Hybrid Leakage and Voltage Reduction under EDF Scheduling
Yifan Zhu Frank Mueller
North Carolina State University

__________________________________________________________________________________________

THURSDAY, SEPTEMBER 29 (IBM Day – OPEN TO THE PUBLIC)
__________________________________________________________________________________________

KEYNOTE: 8:30 - 9:30

Peter Hofstee on Cell

TUTORIAL 9:30 - 12:30

Compiling for the Heterogeneous Parallelism of the Cell Architecture
Kevin O’Brien, Kathryn O’Brien, Alexandre Eichenberger, and Michael Gschwind

KEYNOTE: 1:30 - 2:30

Bruce D'Amora on Cell Programming Models, Graphics and Visualization. Introduction to Demos.

DEMOS: 2:30 - 3:00

Cell

KEYNOTE: 3:00 - 3:45

Charles Webb on zSeries Mainframes

KEYNOTE: 3:45 - 4:30

Balaram Sinharoy on Power5 / PERCS

__________________________________________________________________________________________

FRIDAY, SEPTEMBER 30
__________________________________________________________________________________________

SESSION 5: BUFFERS, REPEATERS, CLOCKS, AND SRAM (Session Chair: A. Elfadel) 8:30 – 10:30

Highly Accurate Power Modeling Method for SRAM Structures with Simple Circuit Simulation
Xiaoyao Liang, David Brooks
Harvard University

Improving Adaptive Cache Leakage Reduction Techniques with Line Buffers
Gurhan Kucuk+, Kanad Ghose*
+ = Yeditepe University * = State University of New York at Binghamton

Using Rotary Clock For Low-Power Clock Distribution
Zhentao Yu, Xun Liu North Carolina State University

Simple Is Better: Keeping Elmore Delay In Repeater Insertion
Yuantao Peng, Xun Liu
North Carolina State University

SESSION 6: MICROARCHITECTURE OPTIMIZATIONS (Session Chair: B. Philhower) 11:00 – 12:30

DLL-Conscious Instruction Fetch Optimization for SMT Processors
Fayez Mohamood, Mrinmoy Ghosh, Hsien-Hsin S. Lee
Georgia Institute of Technology

Loop-based iTLB Resizing for Energy Reduction
Bramha Allu, Wei Zhang
Southern Illinois University, Carbondale

Improving the Energy and Execution Efficiency of a Small Instruction Cache by Using an Instruction Register File
Stephen Hines, David Whalley, Gary Tyson
Florida State University

SESSION 7: RELIABILITY (Session Chair: J. Rivers) 1:30 - 3:00

Compiler-guided Register Reliability Improvement Against Soft Errors
Yun Yan, Wei Zhang
Southern Illinois University, Carbondale

Instruction Error Rate Modeling for High Performance Microprocessors
Aneesh Aggarwal
State University of New York at Binghamton

Optimal Resource Allocation for Concurrent Error Detection Techniques in High Performance Microprocessors
Sumeet Kumar, Aneesh Aggarwal
State University of New York at Binghamton

PANEL: (Panel Chair: P. Bose) 3:30 - 5:00

Reliability Metrics for Processor and System Design: Insights and Pitfalls.
Panelists:
Prof. Jacob Abraham University of Texas
Prof. Sarita Adve University of Illinois at Urbana-Champaign
Dr. Phil Emma IBM T. J. Watson Research Center
Dr. Shubu Mukherjee Intel
Ms. Lisa Spainhower IBM Systems and Technology Group