Zen's IPC jump: from 6 to 8 instructions per cycle, bigger caches and ROB
lemire · x · 2026-09-18
Part 4 of Lemire's thread: the key changes are max instructions per cycle going from 6 to 8, larger caches, and an improved reorder buffer for better instruction scheduling — together delivering the performance gains.
Related event: Lemire Breaks Down Where Ryzen's 50%+ Two-Year Gains Come From(6 posts)→
More from Infra
- Kirin 9030S teardown confirms N+2 node; Kirin 9050 Pro pairs N+3 with N+2 — teortaxesTex · 2026-09-18
- Local model Bonsai 2 reads medication lists on Mac Studio in 5.7 seconds — MaziyarPanahi · 2026-09-18
- Hugging Face Kernels: Swap Faster Kernels Into Models Without Rewriting — RisingSayak · 2026-09-18
- From-scratch speculative decoding engine: 3k lines of PyTorch, 1.94x on H100 — hongyangzh · 2026-09-18
- SSAD 2026 closing talk: Dirk Hornung on OpenXLA as a ready-to-use end-to-end compiler stack — abursuc · 2026-09-18
- AMD Ryzen 7 CPUs gained over 50% Geekbench performance from 2022 to 2024 — lemire · 2026-09-18