Parallelism & Performance
A 12-module path through parallelism & performance, in study order from foundations to the frontier.
Modules
- Performance Metrics and the CPU Time Equation
- Amdahl's Law and Speedup
- Gustafson's Law and Scaled Speedup
- Instruction-Level Parallelism and Dependences
- Superscalar Execution
- Out-of-Order Execution
- Multicore Processors and Thread-Level Parallelism
- Hardware Multithreading and Simultaneous Multithreading
- Cache Coherence
- Memory Consistency Models
- GPUs and SIMD Execution
- The Roofline Model