Computing and Networks
Microarchitecture and SoC Design
Pipelines, caches, interconnect and real RISC-V cores
AI generated by Nikhil Kumar Rajput
In the reading list Computer architecture to silicon (book 2 of 3)
11 chapters
Contents¶
Front matter
Part I · The processor core
Chapter 3: Out-of-Order and Superscalar Execution
- 3.1 The cost of a single stall
- 3.2 True and false dependences
- 3.3 Register renaming
- 3.4 Dynamic scheduling: the scoreboard and Tomasulo's algorithm
- 3.5 Putting results back in order
- 3.6 Speculation: running ahead of branches
- 3.7 Wider: several instructions per cycle
- 3.8 The whole family
- 3.9 What caps a real core's IPC?
- 3.10 Summary
- 3.11 Exercises
Part II · Memory
Part III · Connecting the chip
Part IV · Real cores and measurement
Chapter 8: Real RISC-V Cores
- 8.1 Why RISC-V is the perfect study subject
- 8.2 Single-cycle, multicycle and pipelined
- 8.3 Tiny cores: SERV and PicoRV32
- 8.4 In-order pipelined cores: Rocket and its relatives
- 8.5 Out-of-order and superscalar cores: BOOM, CVA6, XiangShan
- 8.6 Comparing the cores
- 8.7 The ecosystem: Chisel, rocket-chip and Chipyard
- 8.8 From core to SoC
- 8.9 Summary
- 8.10 Exercises
Part V · Security and practice
Chapter 10: Side Channels and Hardware Security
Chapter 11: Labs and a Design Study
- 11.1 The ladder of tools
- 11.2 Lab 1: Ripes, a pipeline that moves one clock at a time
- 11.3 Lab 2: gem5, measuring a real out-of-order core
- 11.4 Lab 3: Chipyard, a real SoC from Rocket and BOOM
- 11.5 Lab 4: Verilator and GTKWave, down to the wires
- 11.6 The open-source toolbox
- 11.7 Capstone: measure one design decision
- 11.8 Summary
- 11.9 Exercises
Back matter
About this book¶
How a processor is made fast and connected to the rest of a chip: pipelining, out-of-order execution, caches and coherence, on-chip interconnect, peripheral interfaces, real RISC-V cores, benchmarking and hardware side channels, with worked examples, exercises and labs.