Lemire Breaks Down Where Ryzen's 50%+ Two-Year Gains Come From
On September 18, Daniel Lemire posted a five-tweet thread analyzing how AMD Ryzen 7 processors improved their overall Geekbench performance by more than 50% from 2022 to 2024, concluding that the gains come mainly from microarchitecture rather than frequency. This breakdown reveals what modern CPU performance progress is actually made of, and is worth a read.
Confirmed
- AMD Ryzen 7 improved overall Geekbench performance by more than 50% between 2022 and 2024.
- Frequency rose only about 15%, not enough to explain the performance gain.
- Transistor count increased roughly 50%, with most of that budget going to CPU cores rather than cache.
- Microarchitectural improvements: maximum instructions per cycle rose from 6 to 8, caches got larger, and the reorder buffer (used for better instruction scheduling) also improved.
- On the SIMD data-parallelism front, Zen 5 processors clearly lead other processors.
Unconfirmed
- Lemire noted that next-generation Zen 6 details are not yet public, but hinted it "could be wild."
Why it matters
- Lemire's analysis shows that recent CPU performance gains rest mainly on microarchitectural evolution (IPC, cache, scheduling, and SIMD) rather than frequency climbs, offering a clear framework for understanding how processors progress.
2026-09-18 ~ 2026-09-18 · 6 related posts
Primary sources
- [source] Zen 5's SIMD data parallelism leads the pack, and Zen 6 'could be wild' — lemire · 2026-09-18
- [source] AMD Ryzen 7 CPUs gained over 50% Geekbench performance from 2022 to 2024 — lemire · 2026-09-18
- Lemire: 15% frequency gain can't explain Ryzen's 50% performance jump — lemire · 2026-09-18
- Ryzen transistor count up ~50%, mostly in cores not cache — lemire · 2026-09-18
- [source] Zen's IPC jump: from 6 to 8 instructions per cycle, bigger caches and ROB — lemire · 2026-09-18
- Zen 5 leads the pack on SIMD data parallelism — lemire · 2026-09-18