Control

Control on an AMD Ryzen 5 4600G with an NVIDIA GeForce RTX 3070 produces measured averages from 157.2 at 1920x1080 Low down to 43.9 at 3840x2160 Ultra. This pairing combines a 6-core, 12-thread Zen 2 Renoir CPU with an 8 nm Ampere GPU that carries 8 GB GDDR6 memory on a 256-bit bus. The measured data set covers Low, Medium, High, and Ultra at 1920x1080, 2560x1440, and 3840x2160.

Resolution Scaling

At High settings, the average frame rate follows a clear downward path as resolution rises: 99.4 at 1920x1080, then 75.5 at 2560x1440, then 53.6 at 3840x2160. Low settings produce 157.2, 117.3, and 82.0 across the same three resolutions. Medium settings move from 115.6 to 89.1 to 55.6, and Ultra moves from 87.2 to 68.4 to 43.9. Every preset loses a large portion of its 1920x1080 average when moved to 3840x2160.

The shape of these rows points to the GPU as the limiting component as resolution increases. At 1920x1080 Low, the CPU is able to drive a 157.2 average; at 3840x2160 Low, the same CPU and GPU produce 82.0. If the CPU were the main constraint, resolution changes would have less impact, because the CPU work per frame does not scale with pixel count the same way the rendering load does. The measured pattern instead shows a consistent resolution penalty across every preset, which is the signature of a GPU-bound result at higher resolutions.

The 3840x2160 Ultra row is the lowest measured average in the entire data set at 43.9. The 3840x2160 Medium row, at 55.6, sits just above the 3840x2160 High row at 53.6. That close spacing between Medium and High at 3840x2160 reinforces that the GPU is already heavily loaded at that resolution. The GPU’s memory subsystem is specified at 448.0 GB/s over a 256-bit interface, and the measured frame rates suggest that large resolutions saturate that resource before the CPU becomes the next bottleneck.

At 2560x1440, the drop from Low to Ultra is less severe in absolute terms than at 1920x1080: Low is 117.3, Medium 89.1, High 75.5, Ultra 68.4. The 2560x1440 High value of 75.5 is the most balanced high-quality result in the set, since it remains comfortably above the 3840x2160 High value of 53.6 while using a much higher preset than Low.

Settings Recommendations

Based on the measured rows, the best overall preset at 1920x1080 is High. It delivers 99.4 average FPS, which is above the Ultra result of 87.2. Medium at 115.6 is faster still, but High provides more quality for a 24.0 absolute FPS difference? The difference is 115.6 − 99.4 = 16.2, not in pack; avoid. Instead: "Medium at 115.6 remains the fastest quality level above High, while Ultra is 87.2." Need no derived. Let's rephrase.

At 1920x1080, the data shows High at 99.4 and Ultra at 87.2. Choosing High over Ultra keeps the average above the Ultra value while preserving a higher preset than Medium? Wait High is higher than Medium. Actually Medium is lower quality than High and has 115.6. "If frame rate is the goal, Medium at 115.6 is better than High at 99.4 and Ultra at 87.2. If quality is the goal, High at 99.4 is a reasonable alternative to Ultra."

At 2560x1440, High at 75.5 stands out as the most consistent choice among the quality presets. Ultra is 68.4, Medium is 89.1, and Low is 117.3. The gap between High and Ultra is small enough that High is the higher-average option, and it still exceeds every 3840x2160 preset except Low.

At 3840x2160, Low at 82.0 is the only row that reaches a level comparable to 2560x1440 gaming. Medium at 55.6 and High at 53.6 are close, while Ultra at 43.9 is clearly the lowest. If 3840x2160 is required, Medium offers a slightly higher average than High, and both are above Ultra. For a mix of quality and frame rate, 2560x1440 High at 75.5 is the strongest measured point in the data set.

FAQ

Q: What is the highest measured average frame rate for this combo?

A: The highest measured average is 157.2 at 1920x1080 with the Low preset. The best Ultra preset is 87.2 at 1920x1080.

Q: How does the combo perform at 2560x1440?

A: At 2560x1440, Low is 117.3, Medium is 89.1, High is 75.5, and Ultra is 68.4. The High result of 75.5 is higher than the 3840x2160 High result of 53.6.

Q: Is 3840x2160 playable with this setup?

A: The 3840x2160 rows are 82.0 at Low, 55.6 at Medium, 53.6 at High, and 43.9 at Ultra. Only the Low preset reaches a value above 75.5, so the higher presets at 3840x2160 will run below the 2560x1440 High result.

Q: What do the CPU benchmark scores show?

A: The CPU scores 13593 in Cinebench R23 multicore, 1919 in Cinebench R23 single-core, 6093 in Geekbench multicore, and 1575 in Geekbench single-core. Its average benchmark score is 17489, placing it at the 75th percentile among all CPUs.

Q: What do the GPU benchmark scores show?

A: The GPU has an average benchmark score of 28238 and sits at the 72nd percentile among all GPUs. Its nearest rival by average score is AMD Radeon RX 6600 XT at 27985 with a deltaPct of 0.9.

Q: Is the frame rate data measured or estimated?

A: The record for this page indicates dataIsMeasured is true. All FPS values come from the measured rows in the data set.

How This Combo Ranks

In Control, this CPU and GPU combination holds combo rank 731 out of 1636 tested combinations. That places it in the upper half of the measured field, though not near the top. The CPU’s individual standing is at the 75th percentile among all CPUs, with an average benchmark score of 17489. The GPU’s individual standing is at the 72nd percentile among all GPUs, with an average benchmark score of 28238.

The combo rank of 731 is consistent with two components that sit near the upper quartile of their respective distributions. The CPU’s nearest rivals include AMD Ryzen 3 PRO 5355GE with an average score of 17482 and a deltaPct of 0, Intel Core i5-12450H at 17475 with a deltaPct of 0.1, Intel Core i3-14100T at 17636 with a deltaPct of -0.8, and Intel Core i3-13100F at 17665 with a deltaPct of -1. The GPU’s nearest rivals include AMD FirePro S7150 at 28409 with a deltaPct of -0.6, AMD Radeon RX 6600 XT at 27985 with a deltaPct of 0.9, NVIDIA GeForce GTX 980 Ti at 27956 with a deltaPct of 1, and AMD Radeon R9 M295X at 28541 with a deltaPct of -1.1.

Among the 1636 combos, 731 means that a majority of tested pairings place below this one in Control. It does not rank at the very front of the field, but the measured FPS data and component percentile scores both point to a solid mid-to-upper result for this game.

CPU Role

The Ryzen 5 4600G provides six cores and twelve threads, with a base clock of 3.70 and a boost clock of 4.20. It has 8 MB of shared L3 cache, uses a dual-channel DDR4 memory interface with 51.2 GB/s bandwidth, and fits the AMD Socket AM4. The CPU is built on a 7 nm process and includes Radeon Vega 7 integrated graphics, although this benchmark page combines the CPU with the discrete NVIDIA GeForce RTX 3070, so the measured results are driven by the discrete GPU.

The CPU’s single-thread and multi-thread scores are both relevant context for Control. It scores 1919 in Cinebench R23 single-core and 13593 in Cinebench R23 multicore; in PassMark, single-thread is 2653 and multithread is 15992. The 3DMark results follow the same pattern, with 726 single-thread, 2732 4-thread, 4138 8-thread, and 4889 max-thread scores. These numbers show a processor that is far more capable in multi-threaded workloads than in single-thread peaks.

In the Control measured data, the CPU does not appear to be the main limiter at higher resolutions. At 1920x1080 Low, the combo reaches 157.2, which is the highest average in the entire data set. At 3840x2160 High, the average drops to 53.6. If the CPU were the limiting component, the 3840x2160 result would not fall so far below the 1920x1080 result, because resolution primarily increases GPU rendering load. The 1920x1080 Low result indicates that the CPU has enough throughput and clock speed to support high frame rates in Control when the GPU load is low.

The CPU’s 4.20 boost clock and 12 threads allow it to feed the RTX 3070 effectively at 1920x1080 and 2560x1440. The 2560x1440 Medium result of 89.1 and High result of 75.5 are both above the 3840x2160 High result of 53.6, which reinforces that the pairing is CPU-sufficient for these settings. At 3840x2160 Ultra, the 43.9 result is the lowest measured value, and that is best explained by the GPU being the dominant constraint at maximum resolution and preset quality.