Apple M4 sustains 10 instructions per cycle, beating most rivals; M5 speedup explained
lemire · x · 2026-09-19
Part 5/8 of Lemire's Apple Silicon series: the M4 can theoretically sustain 10 instructions per cycle — higher than most competitors — yielding more compute per cycle. The base M4 also adds two efficiency cores that help multicore rather than single-core performance. With little visible difference between M4 and M5 on these metrics yet M5 being significantly faster, he turns to memory bandwidth next.
More from Infra
- Leaving DigitalOcean, one site at a time — carnevalem · 2026-09-20
- Google engineers: LLM benchmark harnesses silently drop requests — 200 QPS in, 38 out — AI Engineer · 2026-09-20
- The rig built to run Emacs and doomscroll X is now worth more than its owner's car — tetsuoai · 2026-09-19
- Apple M6 bumps cores to 12 with two super cores; CPUs keep improving fast — lemire · 2026-09-19
- Apple M-series chips gained ~50% Geekbench 6 performance over three years — lemire · 2026-09-19
- Inside OpenAI's inference routing: why the proportional controller had to go — AI Engineer · 2026-09-19