Xiaomi's new mobile CPU matches Apple in single-core speed

Xiaomi’s new custom processor reportedly matches Apple in single-core performance and heavily outperforms it on multithreaded tasks.
The Chinese tech giant built a new CPU called the Xring O3. Analysis shared by computer scientist Daniel Lemire shows the chip uses powerful C1-Ultra cores packed with 44 MB of cache—more than you'll find in most Intel laptop processors. It features 21 execution ports (including six 128-bit SIMD ports), along with SME2 for matrix and AI acceleration. While finding a phone with the new processor might be difficult and Apple could announce its next chip soon, the design signals a major hardware shift.
Why it matters: Chip designers are spending their transistor budgets on massive execution parallelism and huge caches. Adding dozens of execution units and wider SIMD support lets mobile hardware process vastly more arithmetic operations per cycle.
Know this: AMD's Zen 5 still leads in vector width with 4x512-bit support, but Xiaomi's 6x128-bit SIMD setup is currently top-tier for ARM architecture.
Expect mobile chips to keep making standard laptop processors look surprisingly small.
Sources
- Daniel Lemire on X — https://twitter.com/lemire/status/2091894299289874926
- Hacker News Discussion — https://news.ycombinator.com/item?id=49420873

