Intel Core i9-12900K QS allegedly outperforms Ryzen 9 5950X in Cinebench R20 test
According to OneRaichu, a qualification sample of Intel Core i9-12900K processor is already faster than AMD Ryzen 9 5950X in both single and multi-core benchmarks in the popular Cinebench R20 software. However, one should be aware that the CPU will consume a lot of power.
The processor allegedly scores around 810 points in the single-thread Cinebnech R20 test and as many as 11600 points in the multi-threaded test. Both values are higher than any processor currently offered on LGA1200/AM4 sockets. One thing to note, Raichu originally wrote ‘KS’, although he clarified that this was a typo, the SKU is actually “12900K”.
To quickly compare these results we can use, for instance, the latest data from Guru3D. According to those results, Intel Core i9-12900K would outperform AMD Ryzen 9 5950X by 26% and 11% in single and multi-thread benchmarks respectively:
Intel and AMD mainstream desktop processors in Cinebench R20, Source: Guru3D
Raichu does not want anyone to jump to conclusions just yet. While Intel Core i9-12900K might indeed be faster than Ryzen 9 5950X in some benchmarks, the performance crown will heavily depend on workloads. However, both CPUs should be ‘well-matched’, the leaker claims. Furthermore, he claims that that the processor might reach as high as 200W in full turbo mode, which is to be expected as the processor allegedly features a PL2 value of 228W:
Intel Core i9-12900K features a hybrid architecture featuring 8 high-performance Golden Cove cores as well as 8 high-efficiency Gracemont cores. This configuration gives 16 cores and 24 threads, as the smaller cores do not support hyperthreading. In terms of clocks, the 12900K should boost up to 5.3 GHz on Golden Cove and up to 3.9 GHz with Gracemont. This processor is based on a 10nm Enhanced SuperFin fabrication process and will be the first mainstream desktop CPU to support both DDR5 memory technology as well as the latest PCIe Gen5 interface.
Unfortunately, Cinebench results are not stored in a public database, thus the values reported by Raichu cannot be verified. However, this leaker has a good track record so they are definitely worth reporting. It is probably only a matter of time before we start seeing those qualification samples appearing on public benchmark sites, such as Geekbench or SiSoftware Sandra.