The CPU Factor in Ray Tracing
Ray tracing performance at 1080p sits at an awkward intersection where the GPU does most of the heavy lifting but the CPU still matters more than most people expect. At lower resolutions, the processor has to feed frames fast enough to keep the GPU busy, and any bottleneck shows up immediately in frame times and minimum fps figures. That tension becomes especially visible when you put two of the most popular mid-range desktop CPUs side by side – AMD’s Ryzen 5 9600X and Intel’s Core Ultra 5 245K – and push them through ray-traced workloads.
Both chips target roughly the same buyer: someone building or upgrading a gaming PC without spending flagship money on the processor. The 9600X uses AMD’s Zen 5 architecture on a 4nm TSMC process, while the 245K is Intel’s Arrow Lake design built on Intel 20A and TSMC N5 tiles. They’re priced close enough that the choice comes down to real-world performance rather than budget constraints, which makes a direct comparison worth doing properly.
Test Setup and Methodology
Both CPUs were paired with the same RTX 5060 Ti to isolate processor influence as cleanly as possible. Memory was set to the recommended XMP/EXPO profiles – DDR5-6000 CL30 on the AMD platform and DDR5-6400 CL32 on the Intel side, which reflects what each platform actually runs best at rather than forcing artificial parity. The GPU driver version was identical across both systems, and Windows power settings were left at Balanced to reflect a realistic user environment rather than an optimized benchmark lab scenario.
Games selected for the ray tracing suite included Cyberpunk 2077 with path tracing enabled via RT Overdrive, Alan Wake 2 at its Ultra RT preset, and Control with all ray tracing effects active. These three titles put different kinds of pressure on both the CPU and GPU. Cyberpunk’s path tracing is brutally GPU-bound even at 1080p, Alan Wake 2 has complex scene geometry that stresses draw call throughput, and Control’s ray tracing implementation is older but still sensitive to CPU scheduling efficiency.
All results were captured using CapFrameX over a consistent 60-second scene in each title, repeated three times and averaged. Frame time graphs were reviewed alongside average and 1% low fps figures, because ray tracing workloads are notorious for producing smooth averages that hide ugly frame time spikes. A CPU that delivers 90 fps average but drops to 40 fps momentarily is not the same experience as one delivering 85 fps with rock-solid frame pacing.
Average Frame Rate Results
In Cyberpunk 2077 with RT Overdrive at 1080p, the gap between the two CPUs was small but consistently favored the Core Ultra 5 245K. The 245K produced around 4-6 fps more on average across the test scene, which in a heavily GPU-bound path tracing workload is actually meaningful. When the GPU is already being strangled by ray tracing calculations, the CPU’s ability to prepare command lists quickly and handle async compute scheduling becomes more visible. Arrow Lake’s improved memory subsystem and higher bandwidth to the GPU appear to give it a modest but real edge here.
Alan Wake 2 told a more interesting story. The Ryzen 5 9600X closed the gap significantly in that title, with differences falling within a couple of frames per second in most runs. Zen 5’s per-core improvements – particularly in branch prediction and instruction-level parallelism – seem to suit Alan Wake 2’s workload well. The game’s engine distributes CPU tasks differently than Cyberpunk’s REDengine, and the 9600X’s efficiency in single-threaded bursts helps it stay competitive despite the 245K’s architectural advantages elsewhere.
Control was the closest matchup of the three. Both CPUs averaged within 2 fps of each other, and neither produced a clear winner in frame time consistency either. The game’s Northlight engine is well-optimized for multi-core workloads and doesn’t push either CPU toward its ceiling at 1080p with ray tracing active. This is actually the scenario where the choice between these two chips matters least – the GPU is doing the work, and both processors are feeding it adequately.
Where the 245K pulled ahead more noticeably was in 1% low performance during the Cyberpunk path tracing test. The 9600X showed occasional frame time spikes that pushed 1% lows down to figures roughly 12-15% below the 245K’s equivalent metric. This wasn’t a constant stutter – it appeared during scene transitions and moments of heavy particle and lighting load. For competitive gaming this would be disqualifying, but ray-traced single-player games run at 60-90 fps target ranges where brief spikes are felt but not catastrophic. It’s worth knowing about before committing to a platform, though, particularly if you run RT-heavy titles frequently. For GPU-side comparisons at similar resolutions and ray tracing settings, the Radeon RX 9060 XT vs Arc B770 1080p ray tracing results provide useful context on how much headroom these CPUs have to work with depending on which GPU you pair them with.
Power and Thermals Under RT Load
Ray tracing workloads are GPU-intensive enough that CPU power draw is a secondary concern for most people, but the numbers are still worth stating. The 9600X ran at 65-75W package power during the Cyberpunk test scene. The 245K ran significantly hotter and pulled 110-125W under the same conditions. If you’re building in a small form factor case or working with a modest cooling budget, that gap matters. A 65W CPU running cool and quiet inside a compact build is a different user experience from a 120W chip that demands a high-end tower cooler or AIO to stay out of thermal throttling.
The 245K’s performance advantage in ray tracing comes with a real energy cost, and the return on that cost is modest at best. In pure fps-per-watt terms, the 9600X is substantially more efficient, and its smaller performance gap in most titles means you’re paying a disproportionate power penalty for the gains Intel’s chip actually delivers. That calculation changes if you’re also using the CPU for video rendering or streaming while you game, where the 245K’s higher thread count can matter, but as a dedicated gaming chip judged on ray tracing output alone, the efficiency story doesn’t flatter it.
Platform Considerations and Verdict
AM5 motherboards have stabilized in pricing and offer a longer upgrade runway – AMD has committed to AM5 support through at least 2027, meaning a future Zen 6 drop-in upgrade is plausible. Intel’s LGA1851 socket with Arrow Lake is newer and less predictable in terms of long-term support, though Intel has historically offered at least two CPU generations per socket. If the CPU is a platform investment rather than just a one-time purchase, AM5’s trajectory is easier to plan around.
The Core Ultra 5 245K edges ahead in ray-traced frame rates at 1080p, particularly in the most demanding path tracing scenarios, and its 1% low advantage in Cyberpunk is a real differentiator. But the Ryzen 5 9600X keeps pace in most other RT titles, runs cooler, draws less power, and sits on a platform with clearer upgrade options. The 245K’s win here is real but narrow – and whether that margin is worth the power and cooling trade-off depends entirely on how much of your gaming time is spent in Cyberpunk with RT Overdrive cranked up.
