Steamworld Build

Steamworld Build is a city builder where simulation logic and population scaling often dictate performance, and this particular pairing of the Intel Core i7-14700KF and the NVIDIA GeForce RTX 5090 D produces a fascinating dataset. The benchmark results show a configuration that is overwhelmingly GPU-bound at lower resolutions, but which begins to balance the load as pixel count rises. With a combo rank of 15 out of 2924 tested configurations, this system sits near the top of the pile, yet the frame rate curves reveal that the CPU is doing more heavy lifting than one might expect from a 20-core processor. The data indicates a system capable of triple-digit frame rates at 1080p and 1440p, with 4K remaining comfortably playable, though the scaling pattern exposes the architectural limits of the simulation engine.

CPU Role

The Intel Core i7-14700KF is a 20-core, 28-thread processor based on Raptor Lake architecture, running on a 10 nm process node. Its base clock of 3.40 GHz can boost up to 5.60 GHz, which is critical for a city builder that must handle thousands of individual agent calculations per frame. The processor houses 33 MB of shared L3 cache, with 2 MB of L2 and 80 KB of L1 per core, providing ample fast storage for the game's simulation state. In synthetic benchmarks, this CPU scores 44167 in Cinebench R23 multicore and 6235 in single-core, placing it in the 94th percentile of all CPUs tested. Its nearest rival, the Intel Core Ultra 7 265K, scores 70879 in the aggregate benchmark, making the i7-14700KF about 1% slower, while the AMD Ryzen 9 7950X trails by 0.9%.

For Steamworld Build specifically, the CPU's role is most evident in the 1080p Low preset, where the average frame rate hits 397 FPS. This is an extraordinary number for a city builder, and it suggests that the game's simulation thread is not the primary bottleneck at this setting. However, the jump from Low to Medium at 1080p sees a drop to 360 FPS, and High falls to 275 FPS, indicating that graphics settings begin to engage the GPU more heavily. The Ultra preset at 1080p yields 186 FPS, which is still remarkably high but shows a clear progression of GPU load. The data suggests that the i7-14700KF's single-thread performance, evidenced by its 6235 Cinebench R23 single-core score, is more than sufficient to feed the RTX 5090 D even at high frame rates. The 28 threads provide headroom for background tasks and the complex AI routines typical of the genre, but the measured FPS values indicate that the game rarely stresses all cores simultaneously.

When comparing to the nearest rivals, the AMD Ryzen 7 9700F scores 69996 in the aggregate benchmark, which is 0.2% lower than the i7-14700KF's 70163. This marginal difference is unlikely to manifest as a meaningful FPS gap in Steamworld Build, given that the game appears to be more GPU-limited at higher settings. The AMD Ryzen 9 7940HX, a mobile part, scores 69875, 0.4% behind, while the Intel Core Ultra 7 265K is 1% ahead. These deltas are all within noise for a title like this, where the simulation tick rate and draw call overhead matter more than raw compute throughput. The CPU's 125 W TDP and unlocked multiplier suggest it can sustain high clocks under load, which is beneficial for the frame pacing seen in the Medium preset at 1440p, where the minimum FPS of 199 stays well above the 60 FPS threshold.

GPU Role

The NVIDIA GeForce RTX 5090 D is a monstrous GPU built on the Blackwell 2.0 architecture with a 5 nm process. It packs 21760 shading units, 680 texture mapping units, and 176 ROPs, with 170 ray tracing cores and 680 tensor cores. The memory subsystem is equally imposing: 32 GB of GDDR7 memory on a 512-bit bus, delivering 1.79 TB/s of bandwidth. The GPU's base clock is 2017 MHz with a boost of 2407 MHz, and memory runs at 1750 MHz, effective 28 Gbps. This configuration produces a pixel rate of 423.6 GPixel/s and a texture rate of 1,636.8 GTexel/s, with FP32 compute reaching 104.8 TFLOPS. In the Passmark G3D benchmark, the RTX 5090 D scores 44065, placing it in the 92nd percentile of all GPUs, with an average benchmark score of 77712.

The GPU's role in Steamworld Build becomes apparent when examining the resolution scaling. At 1080p Ultra, the GPU is clearly the limiting factor, producing 186 FPS, but this drops to 118 FPS at 1440p Ultra and 62 FPS at 4K Ultra. The 4K Ultra result is particularly telling: the RTX 5090 D, with its 32 GB of VRAM and massive bandwidth, should theoretically crush a city builder, yet it falls to 62 FPS. This suggests that the game's renderer has scaling inefficiencies at higher resolutions, or that the simulation logic becomes more demanding as the viewport expands. The nearest GPU rivals include the AMD Radeon RX 6850M XT, which scores 78940 in the aggregate benchmark, 1.6% higher than the RTX 5090 D's 77712. The NVIDIA Tesla P100 variants also edge out the RTX 5090 D by 2.1% and 2.4%, respectively, though these are compute-oriented cards with different driver optimizations.

The 32 GB VRAM capacity is almost certainly not a bottleneck in Steamworld Build, as the game's textures and geometry are unlikely to exceed even 8 GB at 4K. Instead, the GPU's raw rasterization power is what drives the frame rates. The 4K High preset yields 92 FPS, which is a 33% drop from the 4K Medium preset's 124 FPS. This scaling pattern indicates that the Ultra preset introduces effects that are disproportionately expensive, likely related to shadows or volumetric lighting. The RTX 5090 D's 680 TMUs and 176 ROPs handle fill-rate-heavy scenes well, but the game's engine seems to hit a wall when combining high resolution with maxed settings. The GPU's 575 W TDP is a clear indicator of its performance class, and the data shows that it delivers where it matters, though the 4K Ultra result is surprisingly low for such a flagship part.

Resolution Scaling

The measured FPS data reveals a clear and consistent scaling pattern as resolution increases. At 1080p Low, the system achieves 397 FPS, which drops to 305 FPS at 1440p Low and 161 FPS at 4K Low. This represents a 23% drop from 1080p to 1440p and a 47% drop from 1440p to 4K. The scaling is roughly linear with pixel count, suggesting that the GPU is the primary limiter even at Low settings. However, the absolute numbers are so high that the CPU must be keeping pace, as a slower processor would bottleneck well before 397 FPS.

The Medium preset tells a similar story: 360 FPS at 1080p, 234 FPS at 1440p, and 124 FPS at 4K. The drop from 1080p to 1440p is 35%, while the drop from 1440p to 4K is 47%. Interestingly, the High preset shows 275 FPS at 1080p, 175 FPS at 1440p, and 92 FPS at 4K, with a 36% drop between the first two and a 47% drop between the latter two. The consistency of the 47% drop from 1440p to 4K across all presets indicates that the GPU's memory bandwidth and pixel throughput are the limiting factors at that resolution. The Ultra preset, however, breaks this pattern: 186 FPS at 1080p, 118 FPS at 1440p, and 62 FPS at 4K. The drop from 1080p to 1440p is 37%, which is in line with the other presets, but the drop from 1440p to 4K is 47%, matching the trend. Yet, the absolute 4K Ultra value of 62 FPS is exactly half of the 1440p Ultra value, suggesting a hard scaling limit.

This data indicates that the limiting component shifts depending on resolution and settings. At 1080p Low, the frame rate of 397 FPS is so high that the CPU's single-thread performance, with a Cinebench R23 single-core score of 6235, is likely the ceiling. At 4K Ultra, the GPU's rendering pipeline is the bottleneck, as evidenced by the 62 FPS result. The middle ground, 1440p High with 175 FPS, shows a balanced configuration where neither component is fully saturated. The frame time consistency, visible in the Medium preset's min/max values, also varies: at 1080p Medium, the range is 306 to 414 FPS, a 35% spread, while at 1440p Medium, the range is 199 to 269 FPS, a 35% spread as well. At 4K Medium, the min/max values are 105 to 142 FPS, a 35% spread, indicating that the game's frame pacing is consistent regardless of resolution.

FAQ

Q: Why does the 4K Ultra preset only hit 62 FPS when the GPU is so powerful?

A: The RTX 5090 D is clearly the bottleneck at 4K Ultra, as the drop from 1440p Ultra (118 FPS) to 4K Ultra (62 FPS) is exactly 50%. This suggests the Ultra preset introduces rendering effects that scale poorly with pixel count, likely heavy post-processing or shadow techniques. The GPU's 1.79 TB/s bandwidth is not the issue, as the 4K Medium preset achieves 124 FPS, indicating the Ultra settings add a significant computational load beyond just resolution.

Q: Is the Intel Core i7-14700KF holding back the RTX 5090 D in this game?

A: At 1080p Low, the system hits 397 FPS, which is near the practical limit for most displays. The CPU's single-core score of 6235 in Cinebench R23 is high enough to feed the GPU at these frame rates. However, the 1080p Ultra result of 186 FPS suggests that the CPU is not the primary limiter at any settings, as the GPU's rendering load scales more steeply. The data shows the CPU is sufficient, with its aggregate benchmark score of 70163 placing it in the 94th percentile.

Q: How does this system compare to its nearest rivals in terms of overall performance?

A: The CPU's average benchmark score of 70163 is 0.2% higher than the AMD Ryzen 7 9700F (69996) and 0.4% higher than the AMD Ryzen 9 7940HX (69875). The Intel Core Ultra 7 265K leads by 1% with a score of 70879. For the GPU, the RTX 5090 D's average score of 77712 is 1.6% lower than the AMD Radeon RX 6850M XT (78940) and 2.1% lower than the NVIDIA Tesla P100 PCIe 12 GB (79396). These deltas are minor and unlikely to produce noticeable FPS differences in Steamworld Build.

Q: What is the best resolution to play at for a 60 FPS experience?

A: The 4K Medium preset delivers an average of 124 FPS, with a minimum of 105 FPS, which is well above the 60 FPS threshold. For 1440p, all presets exceed 118 FPS, including Ultra. At 1080p, even the most demanding Ultra preset hits 186 FPS. The data shows that 4K Ultra is the only preset that approaches 60 FPS, with an average of 62 FPS, so any resolution below that with any preset will comfortably maintain 60 FPS.

Q: Does the 32 GB VRAM matter for this game?

A: The measured FPS data does not indicate any VRAM-related bottlenecks. At 4K Ultra, the game uses a significant amount of memory, but the RTX 5090 D's 32 GB GDDR7 is far more than needed. The performance drop from 4K Medium to 4K Ultra is 50%, which is more likely due to shader complexity than memory capacity. The GPU's 512-bit bus and 1.79 TB/s bandwidth ensure that memory bandwidth is not a constraint.

Q: How does the CPU's multi-threaded performance affect the game?

A: The i7-14700KF's Cinebench R23 multicore score of 44167 is impressive, but Steamworld Build likely does not scale perfectly with all 28 threads. The 1080p Low result of 397 FPS suggests the game is not heavily multi-threaded, as a single-thread bottleneck would cap FPS lower. The 20 cores provide headroom for background tasks, but the game's simulation loop appears to be more dependent on single-thread speed, which is why the 5.60 GHz boost clock is beneficial.

Measured FPS Breakdown

The raw frame rate data shows a clear hierarchy across resolutions and settings. At 1920x1080, the Low preset averages 397 FPS, which is the highest measured value in the entire dataset. The Medium preset follows at 360 FPS, with a minimum of 306 FPS and a maximum of 414 FPS, indicating a 35% variance in frame times. The High preset averages 275 FPS, and the Ultra preset drops to 186 FPS. This progression from Low to Ultra represents a 53% reduction in average FPS, which is substantial but expected given the increased graphical fidelity.

Moving to 2560x1440, the Low preset averages 305 FPS, a 23% drop from 1080p Low. The Medium preset averages 234 FPS, with a minimum of 199 FPS and a maximum of 269 FPS, maintaining the 35% variance seen at 1080p. The High preset averages 175 FPS, and the Ultra preset averages 118 FPS. The drop from 1080p to 1440p is consistent across presets, ranging from 23% at Low to 37% at Ultra, showing that the GPU's pixel throughput is the limiting factor as resolution increases.

At 3840x2160, the Low preset averages 161 FPS, which is a 47% drop from 1440p Low. The Medium preset averages 124 FPS, with a minimum of 105 FPS and a maximum of 142 FPS, again showing a 35% variance. The High preset averages 92 FPS, and the Ultra preset averages 62 FPS. The drop from 1440p to 4K is consistently 47% for Low, Medium, and High presets, but the Ultra preset's drop is 47% as well, from 118 to 62 FPS. The 4K Ultra result of 62 FPS is the only measured average below 90 FPS, making it the most demanding configuration in the dataset.

The Medium preset is the only one with complete min/max data across all resolutions, providing insight into frame pacing. At 1080p, the range is 306 to 414 FPS, at 1440p it is 199 to 269 FPS, and at 4K it is 105 to 142 FPS. The consistent 35% spread between minimum and maximum indicates that the game's frame time variance is stable regardless of resolution, suggesting a well-optimized render loop.

Settings Recommendations

Based on the measured FPS data, the High preset at 1440p offers the best balance of visual quality and performance, averaging 175 FPS. This is well above the 60 FPS threshold and provides a smooth experience for a city builder where camera panning and zooming are frequent. The 1440p High preset also leaves headroom for population growth, which can increase simulation load over time. For players with 144Hz displays, the 1440p Medium preset, averaging 234 FPS with a minimum of 199 FPS, ensures that frame rates never dip below the refresh rate, providing a stutter-free experience.

At 4K, the Medium preset is the recommended choice, averaging 124 FPS with a minimum of 105 FPS. This maintains a comfortable margin above 60 FPS while offering improved detail over the Low preset. The 4K High preset, at 92 FPS, is also viable for players who prioritize visual fidelity over absolute smoothness, but the 4K Ultra preset at 62 FPS is borderline, as any increase in simulation complexity could push it below 60 FPS. The data suggests that the Ultra preset's 50% performance penalty over Medium at 4K is not justified by the visual improvements, which are likely subtle in a game with a stylized art style.

For 1080p users, the High preset at 275 FPS is the sweet spot, as it delivers excellent visual quality without the performance cost of Ultra. The 1080p Ultra preset at 186 FPS is also perfectly playable and would be the choice for players with 240Hz monitors who want the highest settings. However, the Medium preset at 360 FPS is the best option for competitive or high-refresh-rate play, as it maximizes frame rates while still providing acceptable visuals. Given the data, the High preset across all resolutions offers the most consistent experience, with scaling that matches the GPU's capabilities. The Ultra preset at 4K should be avoided unless the player is willing to accept 62 FPS average, which may not hold up during large city construction.