Index to Lecture 8
- Lecture 8 Outline
- Branch Penalty
- Delay Slot Filling
- Conditional Operations
- Branch Prediction
- Two Types: Static and Dynamic
- Simple Static Prediction
- Complex Static Prediction
- Simple Dynamic Prediction
- Branch Target Buffer
- Extended Dynamic Prediction
- 2-bit Saturating Counter Predictor
- Aliasing
- Global History
- G-Share Predictor
- Verification of Prediction
- Parallel Branch Pipelines
- Run-Time Profiling
- Neural Net/Pattern Table Predictors
- Hybrid Predictors
- Tournament Predictors
- Statistical Hybrids
- Trace Cache
- Exceptions
- Precise Exceptions
- Branch Prediction Summary
- Instruction Level Parallelism
- Superpipelining
- Superscalar Origins
- Modern Superscalar
- Crossing the Threshold
- Implications of Multiple Issue
- Limits to Superscalar Issue
- Maximum ILP
- Limits to ILP
- Branch Prediction
- Limited Registers
- Imprecise Alias Analysis
- UltraSPARC
- PowerPC 601 part 1
- PowerPC 601 Part 2
- PowerPC 601 Part 3
- PowerPC 604 Part 1
- PowerPC 604 Part 2
- PowerPC 604 Part 3
- PPC 970 Part 1
- PPC 970 Part 2
- PPC 970 Part 3
- Pentium III Part 1
- Pentium III Part 2
- Pentium 4 Part 1
- Pentium 4 Part 2
- Enhancing ILP with Software
- Techniques
- Loop Unrolling
- Software Pipelining
- Global Code Motion
- Trace Scheduling
- Superblocks
- Loop Fusion and Splitting
- Loop Nest Reordering
- Software ILP Summary