Hacker News Reader

They're using PassMark instead of Geekbench.

Apple M1 - CPU Mark: 14,872 / Single thread: 3,630

AMD Ryzen 7 Pro 5850U - CPU Mark: 20,659 / Single thread: 3,237

The AMD is 38.9% faster on the multi-core benchmark, and 10.8% slower on the single-core benchmark.

That's not really surprising, is it? The M1's eight cores are in a big.LITTLE configuration, half of them are not as powerful as the other half.

Also the Ryzen has hyperthreading where M1 doesn’t no?

Yes, M1 doesn’t have hyper threading

OK, so is PassMark more accurate than Geekbench?

(And, still, that's with M1 having 4 cores tied behind its back for better battery life).

Neither bench reflects real world performance. Run cinebench, 7zip, h265 encode, blender render, etc.

Unless your primary workload is media encoding or rendering in Blender, those benchmarks are just as artificial as the truly artificial ones.

Most people’s performance workload is dominated by media consumption—not media creation. But the latter sure does make for good benchmarks.