Introduction

Intel Core i9-13900K “Raptor Lake” has been out for a few weeks now, and as we concluded in its main review, Intel’s bet with the Hybrid architecture for client PC platforms appears to be paying off. Intel isn’t chasing AMD down in raw CPU core-counts, but is adapting a different strategy—of building multi-core processors with two different kinds of cores, the larger and faster “Performance” cores, and the smaller and numerous “Efficiency” cores. The idea behind that is that compute-heavy client workloads such as games don’t need more than an 8-core/16-thread setup made of high IPC cores; whereas multi-threaded general-purpose or creator workloads can take advantage of both the Performance cores, and a large number of Efficiency cores. The i9-13900K enjoys 8 P-cores and 16 E-cores; and our testing puts it at an advantage versus the 16-core/32-thread Ryzen 9 7950X “Zen 4” despite the latter having 16 “full” cores.

Over the past several weeks, we’ve been doing massive 53-game mega-benches of various combinations of hardware, involving AMD Ryzen 7 5800X, the very interesting 5800X3D, and the swanky i9-13900K, with the fastest graphics card you can buy, the GeForce RTX 4090 “Ada.” Most of these articles seek to identify and quantify system/CPU-level bottlenecks with a next-generation GPU. Our article last week pitted the Ryzen 7 5800X3D against the i9-13900K, where the previous-generation “Zen 3” processor put up a valiant fight against the “Raptor Lake,” ending up within a few percent of it.

In today’s article we’re investigating whether there’s some gaming performance to be gained when disabling the E-Cores on the Core i9-13900K. Why would we want to disable E-cores? Some theory: with its 16 E-cores disabled, the 8-core/16-thread Core i9-13900K becomes a P-core-only beast, with its “Raptor Cove” P-cores enjoying the entire power budget of the processor. Since the 36 MB L3 cache remains untouched, there’s more of it accessible per core. The added power budget could also improve boost frequency residency of the P-cores. With E-cores out of the way, you also eliminate any possible failures by the OS scheduler, or Intel thread Director, ensuring the games stick to the P-cores (because there’s no other cores available). Of course all of this is theoretical, and any performance delta between this, and a stock i9-13900K processor should indicate that the games are beginning to send some workload to the E-cores, or rather, the E-cores are able to soak up much of the non-game workload of the OS, such as background tasks, services, network or audio host-signal processing stacks, etc.

In this article, we are comparing the i9-13900K against itself—with its E-cores disabled, versus when left untouched (stock). The idea is to determine if a pure P-core configuration has any benefits to gaming, and if not, just how much performance is lost with the 16 E-cores being disabled. All our testing is done with the GeForce RTX 4090. We specifically picked Windows 10 for this testing, because we keep hearing that you must use Windows 11 with Raptor Lake, or you will see significant performance issues.

Benchmarks

Among our games, you’ll find titles that have been included in TechPowerUp graphics card reviews over the past years, as well as some of the newer ones joining our bench soon. Going forward, we will of course make changes to the game selection for TPU50. Our goal is to have a rich diversity of game genres, engines, and 3D graphics APIs. Equal settings were used for both platforms.

All games are tested in custom bench scenes, as the integrated benchmarks often paint a completely inaccurate picture compared to actual gameplay. Also, the GPU vendors actively optimize their drivers to achieve good results in integrated benchmarks.

Test System