Table of Contents
Workload-Driven Evaluation
Workload-Driven Evaluation
Working Set Perspective
Example Application Set
Working Sets (P=16, assoc, 8 byte)
Working Sets Change with P (NPB)
Where the Time Goes: NPB LU-a
False Sharing Misses: Artifactual Comm.
Questions in Scaling
Under What Constraints to Scale?
Problem Constrained Scaling
Time Constrained Scaling
Memory Constrained Scaling
Scaling Summary
Types of Workloads
NOW Ultra 170 vs Enterprise 5000
Microbenchmarks
Coverage: Stressing Features
Coverage: Levels of Optimization
Concurrency
Workload/Benchmark Suites
Evaluating a Fixed-size Machine
Our problem today
Steps in Choosing Problem Sizes
Steps in Choosing Problem Sizes
Choosing Problem Sizes (contd.)
Multiprocessor Simulation
Execution-driven Simulation
Difficulties in Simulation-based Evaluation
Choosing Parameters
Our Cache Sizes (16x1MB, 16x64KB)
Focus on protocol tradeoffs
Bandwidth per transition
Bandwidth Trade-off
Smaller (64KB) Caches
Cache Block Size
Miss Classification
Breakdown of Miss Rates with Block Size
Breakdown (cont)
Breakdown with 64KB Caches
Traffic
Traffic with 64 KB caches
Traffic SimOS 1 MB
Making Large Blocks More Effective
Update versus Invalidate
Update vs Invalidate: Miss Rates
Upgrade and Update Rates (Traffic)
Summary
|
Author: David E. Culler
Home Page: http://www.cs.berkeley.edu/~culler/cs258-s99/
Other information:
David E. Culler UC Berkeley
CS258 Parallel Computer Architecture
Lecture 5
|