Chapter 19 Further Reading: x86-64 vs. ARM64 Comparison
1. "Why the Apple M1 Chip is So Fast" — Erik Engheim Medium / erik-engheim.medium.com, 2020 — free
A detailed technical blog post explaining M1's performance advantages: the instruction window size, cache hierarchy, the Apple Neural Engine, and the unified memory architecture. Uses accessible language while being technically accurate. Read this if you want the "why" behind the M1 benchmarks.
2. "A Look at Apple M1's Instruction Set Architecture" — Asahi Linux Blog https://asahilinux.org/2021/03/towards-perfect-emulation
The Asahi Linux team (who ported Linux to Apple Silicon) documented Apple's custom ARM64 extensions, undocumented instructions, and hardware behavior. Essential reading for understanding what makes Apple Silicon not just "generic ARM64" — Apple uses ARMv8.5-A features and some proprietary extensions.
3. Chips and Cheese — Microarchitecture Analysis https://chipsandcheese.com
A technical blog that does deep microarchitecture comparisons using performance counters and custom microbenchmarks. Their Apple Silicon analysis (M1 through M4), AMD Zen 4 analysis, and Intel Golden Cove analysis are the best publicly available detailed microarchitecture comparisons. Not affiliated with any vendor.
4. "The Case for RISC-V" — Patterson, Waterman, and Lee Communications of the ACM, 2014 — available online
The original academic paper making the case for a clean, open RISC ISA. Includes quantitative arguments about ISA complexity overhead and the historical forces that led to x86-64's accumulated complexity. Foundational reading for understanding the CISC vs. RISC debate from a computer architecture perspective.
5. "Inside the Apple M1: Processor Analysis" — Anandtech Anandtech.com, 2020 — free
Anandtech's deep dive into M1 performance, power consumption, memory bandwidth, and GPU benchmarks. Includes the "M1 vs. Intel Core i9" comparisons that shocked the industry. Scientifically rigorous with repeatable benchmarks.
6. AWS Graviton Performance Benchmarking — AWS Documentation https://aws.amazon.com/ec2/graviton
AWS publishes price/performance comparisons, supported languages, and migration guides for Graviton (ARM64) instances. Includes real-world benchmarks for web servers, databases, and scientific computing. Essential if you're deciding whether to target ARM64 for cloud workloads.
7. "Rosetta 2: How Apple Translated x86 to ARM" — Low Level Learning YouTube, 2021 — free
Video walkthrough of Rosetta 2's binary translation mechanism, the TSO emulation problem, and how the translation cache works. More accessible than reading Apple's developer documentation directly.
8. RISC-V Foundation Technical Specifications https://riscv.org/technical/specifications
The RV64GC specification: the base integer ISA (RV64I), standard extensions (M=multiply, A=atomics, F=float, D=double, C=compressed). Read Chapter 1 (intro) and Chapter 2 (base ISA) to understand how RISC-V differs from both ARM64 and x86-64 at the ISA level. The C extension (16-bit compressed instructions) is what keeps RISC-V code size competitive.
9. "Computer Architecture: A Quantitative Approach" — Patterson and Hennessy, 6th Ed. Morgan Kaufmann — textbook
Chapter 1 covers performance analysis and the CPU performance equation (IPC × Clock × Instructions). Chapter A (appendix) covers the historical RISC vs. CISC debate with original benchmarks. Appendix K covers memory hierarchies. The canonical academic treatment of everything discussed in this chapter.
10. "Hot Chips 2023: Apple M3 Architecture Deep Dive" IEEE Hot Chips symposium proceedings — available via IEEE Xplore
Apple engineers presented details of the M3 microarchitecture at Hot Chips 2023: fetch width, ROB size, the GPU architecture, and memory subsystem. This is the most authoritative technical source for M-series internals short of actually decapping the chip.