The Medium

The AMD Ryzen 5 5500 paired with the Intel Arc B390 presents a highly constrained gaming platform, and The Medium’s benchmark data illustrates exactly where its limits lie. Across all tested resolutions and settings, frame rates remain low, with the configuration ranking 2581st out of 2911 tested combos in this title. The data shows a system that struggles to maintain playable performance, and the relationship between resolution, settings, and the CPU/GPU pair reveals a clear bottleneck profile.

Resolution Scaling

The measured FPS values show a steep and consistent drop as resolution increases from 1920x1080 to 3840x2160. At the Low settings preset, the average frame rate falls from 51 FPS at 1080p to 43 FPS at 1440p, and then to 34 FPS at 4K. This represents a 33% reduction from 1080p to 4K, which is a substantial decline but not the catastrophic collapse that would indicate a purely GPU-limited scenario at the lowest settings. The scaling from 1440p to 4K at Low shows an 21% drop (from 43 to 34 FPS), which is proportionally similar to the 1080p-to-1440p decline of 16% (from 51 to 43 FPS).

At the Medium preset, the same pattern emerges. The average FPS drops from 32 at 1080p to 27 at 1440p, and then to 21 at 4K. The percentage losses are 16% from 1080p to 1440p, and 22% from 1440p to 4K. This relatively uniform scaling suggests that neither the CPU nor the GPU is completely saturating at any single resolution; instead, both components are sharing the workload, with the GPU becoming progressively more stressed as pixel count rises.

The High and Ultra presets show even more dramatic scaling. High settings yield 23 FPS at 1080p, 19 FPS at 1440p, and 15 FPS at 4K. That is a 35% total drop from 1080p to 4K. Ultra settings produce 18 FPS at 1080p, 15 FPS at 1440p, and 12 FPS at 4K, a 33% decline. Interestingly, the relative percentage drop from 1080p to 4K is nearly identical across all four presets (ranging from 33% to 35%), which indicates that the limiting component is consistent regardless of settings. The data does not show a sudden cliff at any resolution; rather, performance degrades in a steady, linear fashion.

This pattern is telling. If the GPU were the sole bottleneck, the frame rate would typically plummet at 4K, with a much larger delta between 1440p and 4K than between 1080p and 1440p. The observed uniform scaling suggests that the Intel Arc B390, with its modest specifications, is the primary constraint, but the CPU’s single-thread performance also plays a role in keeping absolute numbers low. The 1080p Low result of 51 FPS is the only point where the system approaches a playable level, and even that is far from smooth. The resolution scaling data implies that reducing resolution alone will not rescue this combo for The Medium; even at 1080p, only the Low preset clears 30 FPS.

GPU Role

The Intel Arc B390 is the bottleneck in this configuration. Its benchmark results are meager, with a 3DMark Steel Nomad DX12 score of 1482 and a percentile ranking of just 9th among all GPUs. The nearest rivals in the data are NVIDIA GeForce GT 520MX (score 1463, deltaPct 1.3), GeForce 800M (1460, deltaPct 1.5), GT 625 OEM (1446, deltaPct 2.5), and GT 710 (1443, deltaPct 2.7). The Arc B390 is only 1.3% to 2.7% faster than these ancient, low-end parts, which contextualizes its performance level as essentially entry-level from a decade ago.

The GPU’s architecture is Xe3-LPG on a 3 nm process, with a base clock of 300 MHz and a boost clock of 2500 MHz. It features 1536 shading units, 48 texture mapping units, and 24 raster output pipelines, along with 12 ray tracing cores. The memory configuration is "System Shared," meaning it relies on the system’s DDR4 RAM rather than dedicated VRAM. Memory bandwidth is listed as "System Dependent," which is a crucial limitation for a game like The Medium that may stream large texture data.

The frame rates reflect this GPU’s weakness. At 4K Ultra, the average FPS is just 12, which is effectively a slideshow. Even at 1080p Low, the best case scenario, the GPU manages only 51 FPS. The texture rate of 120 GTexel/s and pixel rate of 60 GPixel/s are the theoretical limits, and the measured data shows that the card cannot sustain high fill rates under load. The fact that the GPU’s boost clock is 2500 MHz but the base is only 300 MHz suggests significant thermal or power throttling potential, but the data does not include clock readings during gameplay. The shared memory architecture is likely the single largest handicap, as the GPU must compete with the CPU for memory bandwidth, and the system’s 51.2 GB/s memory bandwidth figure is already modest.

The comparison to the GT 710 (deltaPct 2.7) is particularly damning. The Arc B390 is only 2.7% faster than a card that was considered low-end at its 2014 release. This means the GPU is fundamentally incapable of handling The Medium’s visual demands at any reasonable quality level. The ray tracing cores are present (12 RT cores), but given the overall performance, enabling any RT feature would likely halve the already-low frame rates. The data does not include RT-specific benchmarks, but the raw raster performance is insufficient for this title.

Settings Recommendations

Based on the measured rows, the only viable settings preset for this combo is Low. At 1080p Low, the average FPS is 51, which is the only measured point above 50 FPS. This is the minimum threshold for a somewhat playable experience, though the lack of min/max data for this row means we cannot assess frame pacing. At 1440p Low, the average drops to 43 FPS, which is still technically playable but marginal. At 4K Low, it falls to 34 FPS, which is below the standard playability threshold.

The Medium preset at 1080p yields an average of 32 FPS, with a minimum of 27 FPS and maximum of 37 FPS. The min FPS of 27 is concerning, as it indicates dips below 30 FPS during heavy scenes. At 1440p Medium, the average is 27 FPS with a min of 23 FPS, and at 4K Medium, it is 21 FPS with a min of 18 FPS. These numbers are all below the playable threshold, and the min FPS values suggest noticeable stuttering.

High settings at 1080p produce 23 FPS average, which is unplayable for a horror game requiring reaction times. Ultra settings at 1080p yield only 18 FPS. The data clearly shows that any preset above Low results in average FPS below 32, and at 4K, even Low is barely above 30 FPS. The recommendation is unambiguous: users must run The Medium at Low settings, and even then, only at 1080p will they see frame rates in the 50s. At 1440p, they will need to accept 43 FPS, and 4K is only viable at 34 FPS, which is still below the 60 FPS target that most players expect.

The Medium preset’s measured min/max data (27-37 FPS at 1080p) is the only other data point with variance information. The spread of 10 FPS indicates that the game is not consistently bound to a single component; the CPU and GPU are alternately limiting depending on the scene. For a smoother experience, the Low preset is the only safe choice, but the data suggests that even that preset will have frame time spikes given the overall hardware constraints.

Measured FPS Breakdown

At 1920x1080, the measured average FPS values are as follows: Low settings deliver 51 FPS, Medium delivers 32 FPS (min 27, max 37), High delivers 23 FPS, and Ultra delivers 18 FPS. The progression from Low to Ultra shows a 65% reduction in frame rate (from 51 to 18 FPS), which is a steep penalty for visual quality. The Medium preset’s min FPS of 27 is only 4 FPS above the Low preset’s average, indicating that the game’s performance is highly sensitive to the preset level.

At 2560x1440, the numbers are: Low at 43 FPS, Medium at 27 FPS (min 23, max 31), High at 19 FPS, and Ultra at 15 FPS. The drop from 1080p to 1440p at each preset is consistent: 8 FPS for Low, 5 FPS for Medium, 4 FPS for High, and 3 FPS for Ultra. The absolute differences shrink as settings increase, which is typical of a GPU-bound scenario where the GPU is already saturated. The Medium preset’s min FPS of 23 at 1440p is barely above the Ultra preset’s average of 15 FPS.

At 3840x2160, the results are: Low at 34 FPS, Medium at 21 FPS (min 18, max 24), High at 15 FPS, and Ultra at 12 FPS. The 4K Low result of 34 FPS is the only 4K data point above 30 FPS, but it is still 17 FPS lower than the 1080p Low result. The Ultra preset at 4K produces 12 FPS, which is the lowest measured value across all configurations. The min FPS of 18 at 4K Medium is lower than the average FPS at 4K Low, showing that the Medium preset’s frame drops are severe.

The data shows a clear pattern: each step up in settings (Low to Medium, Medium to High, High to Ultra) costs roughly 8-10 FPS at 1080p, 8 FPS at 1440p, and 6-7 FPS at 4K. The cost of increasing resolution from 1080p to 4K at the same settings costs 17 FPS for Low, 11 FPS for Medium, 8 FPS for High, and 6 FPS for Ultra. This indicates that the GPU’s fill rate limits become more pronounced at lower settings where the CPU is less of a factor.

FAQ

Q: What is the best average FPS this combo can achieve in The Medium?

A: The highest measured average FPS is 51, achieved at 1920x1080 resolution with Low settings.

Q: Is the Intel Arc B390 comparable to any NVIDIA GPU in terms of benchmark scores?

A: Yes, the Arc B390’s 3DMark Steel Nomad DX12 score of 1482 is only 2.7% higher than the NVIDIA GeForce GT 710 (1443) and 1.3% higher than the GT 520MX (1463).

Q: What are the minimum FPS values for the Medium preset at different resolutions?

A: At 1080p Medium, the min FPS is 27. At 1440p Medium, it is 23. At 4K Medium, it is 18.

Q: Does the combo perform better at 1080p Low or 4K Low?

A: 1080p Low averages 51 FPS, while 4K Low averages 34 FPS, a difference of 17 FPS in favor of the lower resolution.

Q: What is the combo’s rank among all tested configurations in this game?

A: The combo ranks 2581 out of 2911 tested combos, placing it in the bottom 11% of all configurations.

Q: Does the GPU have dedicated VRAM?

A: No, the Intel Arc B390 uses System Shared memory, meaning it relies on the system’s DDR4 RAM for graphics data.

CPU Role

The AMD Ryzen 5 5500 is a 6-core, 12-thread processor based on the Zen 3 architecture, manufactured on a 7 nm process. It has a base clock of 3.60 GHz and a boost clock of 4.20 GHz, with a TDP of 65 W. The CPU’s benchmark scores are modest: a Cinebench R23 multicore score of 16429 and a single-core score of 2319. Its single-thread performance is particularly relevant for The Medium, which is a horror game that likely has a single main game thread.

The CPU’s percentile ranking is 73rd among all CPUs, which places it in the upper-mid range, but its nearest rivals tell the story. The AMD Ryzen AI 5 430 has an average score of 19617 (deltaPct -0.1), meaning the Ryzen 5 5500 is essentially tied with it. The Intel Core i5-12500 scores 19668 (deltaPct -0.4), and the Intel Core i7-10700F scores 19499 (deltaPct 0.5). The deltaPct values are all under 1%, indicating that the Ryzen 5 5500 is statistically equivalent to these other CPUs in synthetic benchmarks.

The data shows that the CPU is not the primary bottleneck in this combo at higher resolutions. At 1080p Low, the 51 FPS result is likely close to the CPU’s frame rate limit for this game, but at 1440p and 4K, the GPU takes over as the limiting factor. The CPU’s 16 MB of L3 cache and 51.2 GB/s memory bandwidth are adequate for feeding the GPU, but the CPU’s single-thread score of 836 in 3DMark (single-thread) and 2319 in Cinebench R23 single-core suggest it can handle the game’s logic without being the dominant constraint.

The fact that the combo ranks 2581 out of 2911 in this game, despite the CPU ranking in the 73rd percentile, confirms that the GPU is the weak link. The CPU’s 6 cores and 12 threads are sufficient for a game like The Medium, which is not known for heavy multi-threading. However, the CPU’s boost clock of 4.20 GHz is not exceptionally high, and the measured FPS at 1080p Low (51 FPS) suggests that the CPU may be limiting frame rates in CPU-bound scenes, even if the GPU is the overall constraint. The CPU’s performance relative to its rivals (all within 0.7% delta) means that swapping the CPU for a different model would not materially change the outcome; the GPU would still dominate the performance profile.

How This Combo Ranks

The combo’s rank of 2581 out of 2911 tested configurations places it in the bottom 11.3% of all combos tested in The Medium. This is a poor ranking, but it is consistent with the individual component rankings: the GPU is in the 9th percentile among all GPUs, while the CPU is in the 73rd percentile. The GPU’s low percentile is the dominant factor in the combo’s overall ranking.

The data indicates that this combo is severely GPU-limited. The GPU’s nearest rivals are all NVIDIA parts from the early 2010s, and the Arc B390’s 3DMark score of 1482 is only 2.7% higher than the GT 710. This means that the combo’s performance is roughly equivalent to a system using a GT 710 with a modern mid-range CPU. The CPU’s high percentile does not compensate for the GPU’s abysmal absolute performance.

In practical terms, this ranking means that the vast majority of the 2911 tested combos will outperform this configuration in The Medium. The only combos ranked lower would have even weaker GPUs or CPUs, but the data shows that the GPU is the primary reason for this low rank. The 1080p Low result of 51 FPS is the only bright spot, but even that is barely above the threshold for playable frame rates. The 4K Ultra result of 12 FPS is the worst-case scenario and is representative of the GPU’s inability to handle high-resolution rendering.

The ranking also implies that any user with a discrete GPU from the last decade would likely see better performance than this combo. The Arc B390’s shared memory architecture and low fill rate are fundamental limitations that cannot be overcome by the CPU’s relatively strong performance. The combo’s rank of 2581 is a clear signal that this configuration is not suitable for The Medium at any settings above Low, and even at Low, the experience will be suboptimal. The data suggests that the only way to improve performance would be to replace the GPU entirely, as the CPU is not the limiting factor in the vast majority of the measured scenarios.