Overpower

The pairing of the Intel Core i7-13700K with the AMD Radeon RX 570 in Overpower produces a dataset that clearly separates the CPU’s substantial headroom from the GPU’s hard performance ceiling. The measured frames per second across three resolutions and four settings presets tell a consistent story: this is a configuration where the graphics card determines the outcome, while the processor operates with considerable capacity to spare. The following analysis breaks down the measured data to explain how the system scales, where its limits lie, and what the numbers suggest for actual gameplay.

Resolution Scaling

The most direct way to observe the bottleneck in this pairing is to track the average FPS as resolution increases. At 1920x1080 with Low settings, the system produces 61.1 FPS. Moving to 2560x1440 at the same Low preset drops the average to 54.7 FPS, a reduction of 6.4 frames. Further scaling to 3840x2160 yields 47.5 FPS, another 7.2-frame decline. The pattern is linear and consistent: each step up in resolution costs roughly 6-7 FPS at Low settings, which indicates the rendering load is scaling directly with pixel count.

The scaling becomes more dramatic as settings quality increases. At High settings, the FPS drops from 38.5 at 1080p to 31.3 at 1440p, then to 28.1 at 4K. The percentage drop from 1080p to 4K at High is about 27% fewer frames. At Ultra settings, the same resolution climb produces 31.2, 25.7, and 20.4 FPS respectively — a total loss of 10.8 FPS from 1080p to 4K, which represents a 35% reduction. This steeper decline at higher presets suggests that the GPU is not only resolving more pixels but also applying more per-pixel work, compounding the stress on the Polaris 20 chip.

The data also reveals that the GPU is the limiting component at every resolution. If the CPU were the constraint, we would expect FPS to remain relatively flat across resolutions because the CPU’s workload (draw calls, physics, game logic) does not change with pixel count. Instead, FPS falls monotonically with every resolution increase, a classic sign of GPU-bound performance. The 4K Ultra figure of 20.4 FPS is particularly telling — it is below the 1080p Low result by a factor of three, which is far more than the CPU’s core count or clock speed could explain. The Intel Core i7-13700K, with its 16 cores and 5.40 GHz boost clock, is not breaking a sweat; the Radeon RX 570 is.

FAQ

Q: What is the highest average FPS achieved in the measured data?

A: The best result is 61.1 FPS, recorded at 1920x1080 with Low settings. This is the only measured configuration that exceeds the 60 FPS threshold.

Q: How does the system perform at 4K resolution?

A: At 3840x2160, the average FPS ranges from 47.5 on Low to 20.4 on Ultra, with Medium at 30.5 and High at 28.1. The 4K Ultra result is below 30 FPS, which typically indicates a playability concern.

Q: Is the CPU or GPU more responsible for the frame rates?

A: The data indicates the AMD Radeon RX 570 is the primary constraint. FPS drops consistently as resolution increases, which is characteristic of GPU-bound performance. The CPU’s benchmark percentile of 93 versus the GPU’s 74 also supports this interpretation.

Q: What is the difference between Low and Ultra settings at 1440p?

A: At 2560x1440, Low produces 54.7 FPS, while Ultra drops to 25.7 FPS. That is a 29.0 FPS gap, meaning the Ultra preset more than halves the frame rate compared to Low.

Q: How does the RX 570 compare to its nearest rivals in general benchmarks?

A: The RX 570’s average benchmark score of 30158 is nearly identical to the NVIDIA GeForce RTX 5060 (30109, deltaPct 0.2) and the AMD Radeon RX 7700 XT (30097, deltaPct 0.2). It sits slightly ahead of the NVIDIA Tesla M60 (29969, deltaPct 0.6) and the RTX 5070 Mobile (29928, deltaPct 0.8).

Q: Does the i7-13700K have any performance advantage over its closest rivals?

A: The i7-13700K’s average benchmark score is 47282. It trails the Intel Core i7-13700KF by 0.1% (score 47334) and the AMD Ryzen 9 PRO 5945 by 0.5% (score 47527), but leads the AMD EPYC 4364P by 0.3% (score 47131) and the Intel Core i9-12900F by 0.4% (score 47081).

Settings Recommendations

For a target of 60 FPS, the measured data offers only one option: 1920x1080 with Low settings, which delivers 61.1 FPS. Every other configuration falls short of that threshold. If the user is willing to accept a lower frame rate, the next best balance of quality and performance appears at 2560x1440 with Low settings, producing 54.7 FPS. This is close to 60 FPS and provides a higher resolution than 1080p, though the visual fidelity from the Low preset may be a trade-off.

At 1920x1080, the Medium preset yields 38.3 FPS, which is nearly identical to the High preset’s 38.5 FPS. This suggests that Medium and High have similar rendering costs in this game, so choosing High over Medium is essentially free in terms of performance. However, the jump to Ultra drops the frame rate to 31.2 FPS, a 7.3 FPS penalty that does not appear justified given the modest difference in settings tier.

For 1440p, the Medium preset at 36.3 FPS is arguably the sweet spot for quality per frame, as it stays above 30 FPS while offering more detail than Low. High at 31.3 FPS is still playable, but Ultra at 25.7 FPS falls below the 30 FPS line. At 4K, no preset is particularly viable for smooth gameplay; the Low preset at 47.5 FPS is the only one above 30 FPS, but 4K Low still requires a significant visual compromise.

The overall recommendation from the data is to prioritize resolution over settings. Dropping from High to Low at 1080p gains 22.6 FPS (from 38.5 to 61.1), while moving from 1080p to 1440p at Low settings only costs 6.4 FPS. This suggests that the GPU has more headroom in pixel throughput than in shader or texture work, so lowering quality settings is more effective than lowering resolution for improving frame rates.

Measured FPS Breakdown

At 1920x1080, the full range of settings produces a spread from 31.2 to 61.1 FPS. The Low preset achieves 61.1 FPS, Medium and High are nearly identical at 38.3 and 38.5 FPS, and Ultra drops to 31.2 FPS. The close grouping of Medium and High is notable — a 0.2 FPS difference is within measurement noise, implying the game’s High preset adds little graphical load over Medium on this hardware.

Moving to 2560x1440, the same pattern holds but with lower absolute values. Low delivers 54.7 FPS, Medium 36.3 FPS, High 31.3 FPS, and Ultra 25.7 FPS. The gap between Low and Medium here is 18.4 FPS, which is larger than the 22.8 FPS gap at 1080p between Low and Medium. This indicates that the Medium preset’s additional effects are more costly at higher resolutions, likely due to increased memory bandwidth pressure on the 224.0 GB/s bus.

At 3840x2160, the numbers compress further. Low produces 47.5 FPS, Medium 30.5 FPS, High 28.1 FPS, and Ultra 20.4 FPS. The difference between Medium and High shrinks to 2.4 FPS, while the drop from High to Ultra is 7.7 FPS. The 4K Ultra result of 20.4 FPS is the lowest measured value in the entire dataset, confirming that this configuration is not suitable for maximum settings at 4K.

The data also shows that the performance delta between Low and Ultra grows with resolution. At 1080p, the gap is 29.9 FPS (61.1 minus 31.2). At 1440p, it is 29.0 FPS (54.7 minus 25.7). At 4K, it is 27.1 FPS (47.5 minus 20.4). While the absolute gap narrows slightly, the percentage difference widens — Ultra delivers 51% of Low’s FPS at 1080p, but only 43% at 4K. This suggests the Ultra preset’s effects scale superlinearly with pixel count, making it increasingly punishing at higher resolutions.

GPU Role

The AMD Radeon RX 570 is the clear performance limiter in this configuration. Its specifications provide context for the measured results: 2048 shading units, 128 texture mapping units, and 32 render output units. The memory subsystem consists of 4 GB of GDDR5 on a 256-bit bus, providing 224.0 GB/s of bandwidth. The chip runs at a base clock of 1168 MHz and a boost clock of 1244 MHz, with a transistor count of 5,700 million on a 14 nm process.

The benchmark data places this GPU at the 74th percentile among all GPUs, with an average score of 30158. Its nearest rivals are clustered tightly: the RTX 5060 scores 30109 (deltaPct 0.2), the RX 7700 XT scores 30097 (deltaPct 0.2), and the Tesla M60 scores 29969 (deltaPct 0.6). This tight grouping suggests that the RX 570’s general compute performance is competitive with much newer cards, but the Overpower results tell a different story. The 4 GB VRAM capacity is likely a factor at higher resolutions, as the measured FPS drops disproportionately when moving from 1080p to 4K.

The GPU’s FP32 throughput of 5.095 TFLOPS and pixel rate of 39.81 GPixel/s are modest by modern standards. The texture rate of 159.2 GTexel/s is sufficient for 1080p but becomes strained at 4K, which aligns with the observed scaling. The 4K Ultra result of 20.4 FPS is roughly 33% of the 1080p Low result, and while pixel count increases by 4x, the FPS does not drop by 4x, indicating some efficiency in the rendering pipeline. However, the absolute numbers are too low for smooth play.

CPU Role

The Intel Core i7-13700K provides a stark contrast to the GPU’s limitations. With 16 cores, 24 threads, and a boost clock of 5.40 GHz, this processor has ample compute headroom for Overpower. Its benchmark scores reinforce this: 39000 in Cinebench R23 multi-core, 5505 in single-core, and a Passmark multi-thread score of 45887. The CPU’s percentile rank of 93 among all CPUs places it well above the GPU’s 74th percentile, confirming that the processor is not the bottleneck.

The CPU’s nearest rivals show how close the performance envelope is. The i7-13700KF scores 47334 (deltaPct -0.1), the AMD EPYC 4364P scores 47131 (deltaPct 0.3), and the Intel Core i9-12900F scores 47081 (deltaPct 0.4). The differences are all within one percent, meaning the i7-13700K is effectively tied with its closest competitors in general compute. This consistency suggests that in Overpower, any of these CPUs would produce identical frame rates when paired with the RX 570, as the GPU would still be the limiting factor.

The CPU’s 30 MB of shared L3 cache and 2 MB per-core L2 cache provide fast access to game data, but this advantage is not visible in the measured FPS because the GPU cannot process frames fast enough to benefit. The 3.40 GHz base clock and 5.40 GHz boost clock are both high enough to avoid any CPU-bound scenarios in the tested configurations. Even at 1080p Low, where the GPU produces 61.1 FPS, the CPU is likely waiting on the GPU to finish rendering each frame. The data shows no evidence of CPU throttling, and the benchmark results indicate a processor that is capable of far more than this GPU pairing demands.