The Medium

The Medium, a fixed-camera horror title built around a dual-reality mechanic, presents an unusual load profile. With the Intel Core i5-11400F and NVIDIA GeForce RTX 4090, the data reveals a system where the GPU’s raw power is frequently restrained by the CPU’s architectural limits, particularly at lower resolutions. This analysis relies exclusively on the measured frame rates and benchmark scores provided.

CPU Role — cores, clocks, and how they relate to this game's results

The Intel Core i5-11400F is a 6-core, 12-thread processor based on the Rocket Lake architecture, built on Intel’s 14 nm process. Its base clock of 2.60 GHz and boost clock of 4.40 GHz define its operational envelope. In the context of The Medium, this CPU’s single-thread performance is a critical factor. The game’s rendering logic, which often manages two simultaneous camera views, places a premium on per-core throughput rather than sheer core count. The benchmark data for the processor shows a Cinebench R23 single-core score of 2029 and a multi-core score of 14372, indicating a significant gap between its strongest single-thread capability and its multi-threaded throughput.

This gap is relevant because The Medium’s engine does not scale perfectly with the 12 threads available. The PassMark multi-thread score of 16908 and the 3DMark 16-thread score of 5619 suggest that while the processor can handle parallel workloads, its performance in a gaming scenario is often pinned to the speed of its fastest cores. At 1080p, the measured average FPS of 89.4 on High settings is notably lower than the 124.9 FPS seen on Low settings. This delta indicates that as settings increase, the CPU is not the sole bottleneck; however, the fact that the 1080p Low result does not reach the RTX 4090’s potential ceiling points to a CPU limitation in frame pacing.

The processor’s Cinebench R20 single-core score of 852 and Geekbench single-core score of 1979 place it in a competitive position for its generation, but the data shows it trails more modern parts. Compared to its nearest rivals, the average benchmark score of 17181 is nearly identical to the AMD Ryzen 7 5700U (17189, 0% delta) and the Intel Core i7-1260P (17212, -0.2% delta). This parity in synthetic averages suggests that the i5-11400F’s gaming performance in a title like The Medium will be similar to those processors, though the game’s specific engine quirks may favor the higher boost clock of the i5 in single-threaded scenarios. The L3 cache of 12 MB (shared) is adequate, but the dual-reality rendering in The Medium may cause frequent cache misses, as the engine alternates between two scenes.

GPU Role — VRAM, clocks, and how they relate to this game's results

The NVIDIA GeForce RTX 4090 is the dominant component in this pairing, equipped with 24 GB of GDDR6X memory on a 384-bit bus, yielding a massive 1.01 TB/s of bandwidth. Its Ada Lovelace architecture, built on a 5 nm process, includes 128 RT cores and 512 Tensor cores. The base clock of 2235 MHz and boost clock of 2520 MHz allow it to handle the most demanding rasterization and ray tracing tasks. In The Medium, the GPU’s role is to render the game’s two simultaneous worlds, a task that is heavily dependent on memory bandwidth and raw fill rate.

The measured FPS data shows the RTX 4090’s strength. At 4K Ultra, the average FPS is 49, while at 4K High it is 64.3, and at 4K Low it jumps to 109.4. These figures demonstrate that the GPU scales predictably with settings, but the performance is far from the card’s theoretical maximum. The PassMark G3D score of 38194 and the 3DMark Steel Nomad DX12 score of 9223 confirm the card’s top-tier status, yet the game’s engine does not fully utilize its capabilities. The GPU’s pixel rate of 443.5 GPixel/s and texture rate of 1,290.2 GTexel/s are more than sufficient for the game’s visual fidelity, but the fixed-camera perspective and pre-rendered lighting effects mean that the card is often idling, waiting for the CPU to deliver draw calls.

The 24 GB VRAM buffer is never a limiting factor in this title, as even the most detailed textures at 4K Ultra are unlikely to exceed that allocation. The memory clock of 1313 MHz (21 Gbps effective) provides the bandwidth needed for the dual-scene rendering, but the data suggests that the GPU’s frame output is constrained by the CPU at lower resolutions. The RTX 4090’s nearest rivals in synthetic tests are not consumer gaming cards—the NVIDIA Tesla T4 (66733, -0.4% delta) and NVIDIA Quadro P6000 (67320, -1.3% delta) are professional compute parts. This underscores that the RTX 4090’s gaming performance is exceptional, but its lead over such rivals in a game like The Medium is less pronounced due to engine limitations.

Resolution Scaling — how FPS drops from 1080p to 4K and what it says about the limiting component

The FPS scaling across resolutions reveals a classic bottleneck shift. At 1080p with High settings, the average FPS is 89.4. Moving to 1440p High, the FPS drops to 82.9, and at 4K High, it falls to 64.3. This represents a 28% decrease from 1080p to 4K. The pattern is similar at other settings: 1080p Low yields 124.9 FPS, while 4K Low yields 109.4 FPS, a 12% drop. The smaller percentage drop at Low settings indicates that the CPU is the primary limiter at 1080p Low, as the GPU has ample headroom to render frames quickly, but the CPU cannot feed it fast enough.

At Ultra settings, the scaling is more pronounced. 1080p Ultra averages 77.5 FPS, but 4K Ultra drops to 49 FPS, a 37% decrease. This larger drop at Ultra settings suggests that the GPU becomes the dominant bottleneck as the increased visual fidelity (shadows, reflections, and texture filtering) consumes more of the card’s resources. The data shows a clear trend: the RTX 4090’s performance is more consistent across resolutions at Low settings, where the CPU limits the frame rate, whereas at Ultra settings, the GPU’s workload increases enough to make the resolution change more impactful.

The 1440p Medium result of 95.4 FPS versus the 1080p Medium result of 103.3 FPS shows a 7.6% drop, which is modest. This indicates that at Medium settings, the system is balanced, with neither component fully saturated. However, the 4K Medium result of 88 FPS versus 1440p Medium shows a 7.7% drop, suggesting that the GPU is beginning to feel the pressure of the higher pixel count. The data implies that the i5-11400F is sufficient to drive the RTX 4090 to playable frame rates at all resolutions, but it prevents the card from achieving its maximum potential at 1080p, where the CPU’s single-thread performance is the limiting factor.

FAQ

Q: What is the average FPS for this combo at 1080p with High settings?

A: The measured average FPS is 89.4. This is lower than the 1080p Low result of 124.9 FPS, indicating that the CPU is the limiting component at this resolution.

Q: How does the CPU’s multi-threaded performance compare to its single-threaded performance?

A: The Cinebench R23 multi-core score is 14372, while the single-core score is 2029. This ratio suggests that the processor’s multi-threaded capability is strong, but the game’s engine relies more on single-thread speed, which is why the 1080p results are capped.

Q: Is the RTX 4090’s 24 GB VRAM sufficient for The Medium at 4K Ultra?

A: Yes, the 24 GB buffer is more than adequate. The 4K Ultra average FPS of 49 is limited by the GPU’s rendering workload, not memory capacity, as the game’s textures are unlikely to exceed this allocation.

Q: What is the difference in FPS between 1080p and 4K at Low settings?

A: At 1080p Low, the average FPS is 124.9, while at 4K Low, it is 109.4. This 12% drop is smaller than the drops at higher settings, showing that the CPU is the primary bottleneck at Low settings.

Q: How does this combo rank among all tested combos in The Medium?

A: This combo ranks 163rd out of 1689 tested combinations. This places it in the top 10% of all systems tested, though the ranking shows that many other configurations deliver higher average frame rates.

Q: What is the GPU’s benchmark score relative to its nearest rival?

A: The RTX 4090 has an average benchmark score of 66473. Its nearest rival, the NVIDIA Tesla T4, scores 66733, a delta of -0.4%, meaning the RTX 4090 is essentially tied in synthetic benchmarks, despite being a gaming-focused card.

How This Combo Ranks

The combo rank of 163 out of 1689 tested systems places this configuration in the top 9.6% of all tested combinations for The Medium. This is a strong ranking, but it is not at the very top, which is notable given the RTX 4090’s position as the fastest consumer GPU. The ranking is likely dragged down by the CPU’s performance in this specific title. The i5-11400F’s 75th percentile ranking among all CPUs indicates that it is a solid processor, but its single-thread performance (Cinebench R23 single-core score of 2029) is not sufficient to fully feed the RTX 4090.

The data shows that the GPU’s raw power is not the limiting factor in this game; the CPU is. The 1080p High average of 89.4 FPS is far below what the RTX 4090 is capable of in other titles, and the fact that the 1080p Low result of 124.9 FPS is still capped suggests a hard CPU limit. The ranking of 163rd implies that many systems with newer, higher-clock CPUs achieve better frame rates despite having weaker GPUs. For example, a system with a CPU that has a higher single-thread score would likely push the RTX 4090 to higher FPS at 1080p, improving the overall combo rank.

At 4K, the rank is more favorable, as the GPU becomes the bottleneck. The 4K Ultra average of 49 FPS is respectable, and the 4K High average of 64.3 FPS is playable. The combo’s rank of 163 reflects a system that is extremely capable at high resolutions but leaves performance on the table at lower resolutions. The RTX 4090’s percentile ranking of 91 among all GPUs is elite, but the pairing with the i5-11400F creates a mismatch that prevents the system from achieving its full potential in The Medium. The data indicates that this combo is best suited for 4K gaming, where the GPU’s advantages are fully utilized, rather than 1080p esports-style play, where the CPU’s limitations are exposed.