Benchmarks of Intel’s next-gen Core i9-13900K processor have hit the web, showing the gains conferred by the additional Gracemont cores. The test was run using CPU-Z on an early engineering sample so do keep in mind that the single-threaded score will be notably lower than the final retail SKU. In the single-thread benchmark, the Core i9-13900K is 5-6% slower than the Ryzen 9 5950X but as already stated this benchmark is rather redundant.
In the multi-threaded benchmark, the Raptor Lake flagship is 10% faster than the 16-core AMD processor. Considering that we’re looking at an early sample running at 4.5GHz (P-core) and 3GHz+/- (E-core), this is a fairly decent result. The chip was tested on the ASUS ROG Maximus Z690 Apex and paired with super-fast DDR5-7400 memory. As noted in earlier leaks, the L3 cache has been increased to 36MB, and each performance core and efficiency core cluster will pack 2MB and 12MB of L2 cache, respectively. AVX-512 instructions appear to be disabled, and the BIOS version is from late March 2022.
An engineering sample with a boost clock of 4.5GHz is rather promising. Intel might very well be able to push the final retail version over 5.5GHz by the time of launch.
Source: HXL (Twitter)