Two Mid-Range CPUs, One Demanding Test
Ray tracing at 1080p sits in an awkward spot for CPU testing. The resolution is low enough that the GPU finishes its rasterization work fast, putting pressure back on the processor to keep the pipeline fed. Add ray traced reflections, shadows, and ambient occlusion into that pipeline, and the BVH traversal workloads start to expose real differences between CPU architectures. That is exactly the scenario where the AMD Ryzen 5 9600X and Intel Core Ultra 5 245K end up telling very different stories.
Both chips target the same buyer: someone building a capable gaming PC without spending flagship money on the processor. The 9600X runs six Zen 5 cores at up to 5.4 GHz, with AMD’s improved branch predictor and wider front-end feeding into a cleaner IPC jump over Zen 4. The Core Ultra 5 245K takes a different approach entirely, pairing six Performance cores with eight Efficient cores under Intel’s Arrow Lake architecture, trading single-thread ceiling for a broader workload spread. On paper, neither chip looks like a clear winner at ray traced workloads specifically. In practice, the gap is more decisive than the spec sheet suggests.
Test Setup and Methodology
Testing was conducted with both CPUs paired to the same RTX 4080 Super to ensure the GPU ceiling remained consistent across runs. Memory was set to DDR5-6000 on both platforms using XMP/EXPO profiles, which represents a realistic enthusiast configuration rather than a stock-speed handicap. Games were tested at 1080p with ray tracing enabled at medium-to-high quality presets – not path tracing maxed out, but the settings a real player would actually use. Titles included Cyberpunk 2077 with Ray Tracing Ultra, Alan Wake 2 with ray traced lighting, Control with full RT effects, and Dying Light 2 with ray traced global illumination enabled.
Frame time consistency was tracked alongside raw average frame rates, because a chip that averages 120 fps but stutters every few seconds is worse for playability than one averaging 110 fps smoothly. The 1% low figures carry serious weight here. Ray tracing pipelines generate irregular workload spikes tied to on-screen geometry complexity, and CPUs handle those spikes differently depending on how aggressively they can boost a single core versus distributing load across many cores.
Game-by-Game Performance
In Cyberpunk 2077 at Ray Tracing Ultra, the 9600X pulled ahead by roughly 8-11 percent in average frame rates across the tested scenes. More telling was the 1% low gap, where the Ryzen chip maintained steadier frame delivery through the game’s dense Night City districts. The Core Ultra 5 245K’s efficient cores do not engage meaningfully in this scenario because ray tracing work does not distribute well to lower-frequency E-cores – the BVH traversal tasks essentially queue up on the P-cores anyway.
Alan Wake 2 told a similar story. Remedy’s ray traced lighting system is one of the most CPU-sensitive RT implementations available, and the 9600X’s Zen 5 IPC advantage translated directly into better minimum frame rates in the darker interior sections where shadow rays multiply fast. The 245K was not dramatically behind – we are talking about differences that land in the 7-9 fps range at average, which is noticeable but not catastrophic. What mattered more was the 245K’s occasional dips below 60 fps in specific scenes where the 9600X held above that threshold consistently.
Control with full ray tracing is an older implementation, and here the gap between the two chips narrowed considerably. Both processors handled the game comfortably, staying well above playable frame rates with minimal stuttering. This is partly because Control‘s RT workload is less aggressive than newer titles, and partly because at sufficient average frame rates, the architectural differences between Zen 5 and Arrow Lake matter less to the player experience.
Dying Light 2 was the closest contest of the four. The game’s open-world structure distributes CPU workload differently from the corridor-heavy scenarios in Alan Wake 2, and the 245K’s hybrid core count showed some benefit in managing background simulation tasks while the P-cores handled rendering. Average frame rates came within 4 percent of each other, with the 245K actually edging ahead in one outdoor benchmark sequence by a small margin. It is the one scenario in this test where the Intel chip’s architecture works in its favor at ray traced workloads.
Power and Thermal Differences
The 9600X runs within AMD’s 65W TDP by default, and even under sustained ray tracing loads it rarely exceeded 75W at the package level during testing. The Core Ultra 5 245K defaults to Intel’s 125W MTP, and in practice drew closer to 115W under the same workloads. For a user running a smaller case or a budget air cooler, that difference is not trivial. The 9600X stayed cooler and quieter under load without any manual power limits applied.
Intel does allow the 245K to be power-limited down to a 65W target in BIOS, which drops performance modestly but brings it closer to the 9600X’s thermal behavior. At 65W limited, the 245K loses roughly 5-6 percent of its average frame rate in ray traced titles, which narrows the gap with AMD but does not close it. It also requires deliberate BIOS intervention that most users will not bother with out of the box.
Pricing and Platform Costs
Street pricing puts the 9600X and 245K within about $20-30 of each other depending on the retailer and timing, which makes the head-to-head comparison genuinely relevant rather than one chip punching down. Where the cost picture shifts is in platform expenses. B650 motherboards for the Ryzen chip are widely available at lower price points than Z890 boards required to unlock the 245K’s full performance. A realistic B650 build with DDR5 memory undercuts the equivalent Z890 setup by $60-100 in most configurations.
That platform gap matters more when the use case is primarily gaming with ray tracing. If the workload extended into heavily multithreaded production tasks – video encoding, 3D rendering, large compile jobs – the 245K’s eight E-cores would contribute meaningfully and shift the value calculation. For gaming-first builds, the additional platform cost is harder to justify when the gaming performance consistently runs behind.
For anyone pairing either chip with a current-generation GPU that supports hardware ray tracing, the GPU choice will influence total RT performance significantly – but the CPU differences found here hold across different GPU pairings. The 9600X’s advantage in ray traced 1% lows does not disappear when you swap in a different card; the architecture is simply better matched to how modern RT workloads distribute their demand across a processor’s execution resources.
The Core Ultra 5 245K is not a bad chip, and in non-gaming workloads its hybrid core count earns its place. But if you are specifically building for ray traced PC gaming at 1080p and evaluating these two processors side by side, the 9600X wins more scenarios, runs cooler, costs less to build around, and delivers better frame time stability in the titles where RT workloads hit hardest. That is a clean enough result that the platform savings only reinforce it.
