Steamworld Build

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

The Intel Core i7-12700KF is a 12th Gen Alder Lake desktop processor with 12 cores and 20 threads, running at a base clock of 3.60 GHz and a boost clock of 5.00 GHz. Its architecture features a shared 25 MB L3 cache, with 80 KB L1 and 1.25 MB L2 per core. This CPU sits in the 85th percentile of all processors tested, with an average benchmark score of 35,365. In synthetic multi-threaded workloads, it scores 9,983 on 3DMark max threads, 28,838 on Cinebench R23 multicore, and 34,092 on Passmark multithread. Single-thread performance is also strong: 1,043 on 3DMark single thread, 4,071 on Cinebench R23 single core, and 3,984 on Passmark single thread.

For a city builder like Steamworld Build, the CPU's role is significant because these games typically simulate many entities simultaneously — workers, transport, production chains, and building states. The data shows the 12700KF's 20 threads provide ample headroom for such simulation workloads. The nearest rival, the Intel Core i7-13700T, scores 35,403 on average, a delta of just -0.1% relative to the 12700KF, meaning the two are effectively tied in overall CPU performance. The Intel Core i5-13600T scores 35,305, a 0.2% delta, also effectively identical. This places the 12700KF in a tight performance cluster where no rival has more than a 0.2% advantage.

The boost clock of 5.00 GHz is particularly relevant for the game's frame pacing at lower resolutions. When the GPU is not the bottleneck, the CPU's single-thread speed determines how fast the simulation tick can advance. The 3DMark 2-thread score of 2,065 and 4-thread score of 4,011 indicate that the processor scales well with the limited thread counts that many game engines actually utilize. At 1080p, where the CPU is more likely to be the limiting factor, the measured 353 FPS on Low settings suggests the 12700KF can feed the RTX 5090 D at very high rates, but the simulation overhead of higher settings will pull that down.

In terms of memory support, the 12700KF supports both DDR4 and DDR5 in a dual-channel configuration, which gives system builders flexibility, though the benchmark data does not separate results by memory type. The CPU connects via PCIe Gen 4 with 20 lanes from the processor itself. The launch MSRP of $384 positions it as a mid-range-to-high-end part, though pricing considerations are outside the scope of this analysis.

FAQ — 4-6 Q&A pairs answerable from FACT PACK data

Q: How does the Intel Core i7-12700KF compare to its closest rival, the Intel Core i7-13700T?

A: The average benchmark scores are nearly identical: the 12700KF scores 35,365, while the i7-13700T scores 35,403, a delta of -0.1%. This means the 12700KF is effectively tied with the 13700T in overall CPU performance, with a negligible 38-point deficit out of roughly 35,000 points.

Q: What is the CPU's percentile ranking among all processors?

A: The 12700KF sits in the 85th percentile of all CPUs tested, meaning it outperforms 85% of processors in the database. Its average benchmark score of 35,365 places it well above the median, with the nearest rival (i7-13700T) at 35,403 and the Intel Core 5 213PE at 35,428, both within 0.2% delta.

Q: Does the CPU's performance scale well with more threads?

A: Benchmark results show strong scaling. The 3DMark scores progress from 1,043 (single thread) to 2,065 (2 threads), 4,011 (4 threads), 7,211 (8 threads), and 9,983 (max threads). This indicates the 20-thread configuration delivers meaningful multi-threaded gains, which is useful for simulation-heavy gameplay.

Q: What is the CPU's single-threaded performance relative to its multi-threaded performance?

A: The single-thread score on Cinebench R23 is 4,071, while the multicore score is 28,838, a ratio of about 7.1x. The Passmark single-thread score of 3,984 versus multithread of 34,092 shows a similar 8.6x scaling. This suggests the CPU can handle both lightly-threaded game logic and heavily-parallel background tasks.

Q: How does the GPU compare to its nearest rivals in synthetic benchmarks?

A: The RTX 5090 D scores 77,712 on average, placing it in the 92nd percentile of all GPUs. Its nearest rival, the AMD Radeon RX 6650M XT, scores 76,904, a delta of 1.1% (the 5090 D is ahead). The AMD Radeon RX 6850M XT scores 78,940, which is 1.6% higher, and the NVIDIA Tesla P100 PCIe 16 GB scores 79,605, which is 2.4% higher. So the 5090 D is competitive but not the absolute top in aggregate GPU benchmarks, despite its high-end specs.

Q: What are the GPU's key specifications and memory features?

A: The RTX 5090 D has 32 GB of GDDR7 memory on a 512-bit bus, yielding 1.79 TB/s of bandwidth. It features 21,760 shading units, 680 texture mapping units, 176 ROPs, 170 ray tracing cores, and 680 tensor cores. The boost clock is 2,407 MHz, and it delivers 104.8 TFLOPS of FP32 compute. It has a 575 W TDP and uses a 16-pin power connector.

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

The measured FPS data across resolutions reveals a clear pattern of GPU-bound scaling at higher settings and CPU-bound behavior at lower settings. At 1080p with Low settings, the combo achieves 353 FPS. Moving to 1440p Low, it drops to 304 FPS, and at 4K Low, it reaches 161 FPS. This is a 54% reduction from 1080p to 4K at Low settings, indicating that even at the lowest settings, the GPU is becoming a limiting factor as pixel count quadruples.

At Medium settings, the scaling is similar but with a steeper initial drop. The 1080p Medium average is 320 FPS, 1440p Medium is 234 FPS (a 27% drop), and 4K Medium is 124 FPS (a 61% drop from 1080p). The min-max range at 1080p Medium is 272-368 FPS, at 1440p Medium it is 199-269 FPS, and at 4K Medium it is 105-142 FPS. These ranges show that frame time consistency degrades as resolution increases, with the 4K Medium min being 85 FPS below the 1080p Medium min.

High settings show the same trend. At 1080p High, the average is 275 FPS. At 1440p High, it is 175 FPS, a 36% drop. At 4K High, it is 92 FPS, a 67% drop from 1080p. This suggests that as settings increase, the GPU load grows disproportionately with resolution, which is typical for a city builder where draw calls and object detail scale with pixel density.

Ultra settings amplify this effect dramatically. The 1080p Ultra average is 186 FPS, 1440p Ultra is 118 FPS (a 37% drop), and 4K Ultra is 62 FPS (a 67% drop from 1080p). The 4K Ultra figure of 62 FPS is the lowest measured in the entire dataset. This indicates that at Ultra settings, the GPU is heavily loaded at 4K, and the CPU's role is diminished because the GPU cannot produce frames fast enough to become CPU-bound.

The key takeaway is that the 12700KF can sustain extremely high frame rates at 1080p, up to 353 FPS on Low, which strongly suggests the CPU is not a bottleneck at lower resolutions. However, as resolution increases, the RTX 5090 D's workload grows, and the frame rate drops by roughly two-thirds from 1080p to 4K at every settings level. This is classic GPU-bound scaling. The fact that 1080p Low (353 FPS) and 1080p Ultra (186 FPS) differ by 167 FPS shows that the game's settings have a substantial impact on GPU load, independent of resolution.

Settings Recommendations — which preset gives the best experience per the measured rows

Based on the measured FPS values, the Medium preset offers the most balanced experience across all resolutions. At 1080p, Medium delivers 320 FPS average with a min of 272 and max of 368, which is only 33 FPS lower than Low (353 FPS) but 134 FPS higher than Ultra (186 FPS). The min FPS of 272 at 1080p Medium is well above any display refresh rate in common use, ensuring smooth gameplay without the visual degradation of Low settings.

At 1440p, Medium averages 234 FPS with a min of 199 and max of 269. This is a 70 FPS drop from Low (304 FPS) but a 116 FPS improvement over Ultra (118 FPS). The min of 199 FPS at 1440p Medium remains fluid even on high-refresh-rate monitors. In contrast, Ultra at 1440p drops to 118 FPS average, which, while playable, is significantly lower and may show more frame time variance.

At 4K, Medium delivers 124 FPS average with a min of 105 and max of 142. This is the only 4K preset besides Low (161 FPS) that exceeds 100 FPS on average. High at 4K drops to 92 FPS, and Ultra falls to 62 FPS. The 4K Medium min of 105 FPS ensures that even in demanding scenes, the frame rate stays above 100 FPS, which is a strong target for a city builder where camera pans and zoom operations can cause frame hitches.

The High preset is a reasonable middle ground for visual fidelity, but the data shows it sacrifices significant performance: 275 FPS at 1080p, 175 FPS at 1440p, and 92 FPS at 4K. Compared to Medium, High costs 45 FPS at 1080p, 59 FPS at 1440p, and 32 FPS at 4K. The visual uplift from Medium to High is not quantified in the data, but the performance cost is consistent.

Low settings are only recommended for competitive or maximum-refresh-rate scenarios, as they reach 353 FPS at 1080p and 304 FPS at 1440p. However, even at 4K, Low produces 161 FPS, which is more than sufficient for any display. The Ultra preset is the most demanding, with 62 FPS at 4K being the only measured result below 90 FPS, making it the only preset that could cause perceptible stutter on high-refresh displays. For most users, Medium provides the best combination of high average FPS, stable minimums, and acceptable visual quality across all three resolutions.

How This Combo Ranks — use comboRankInGame vs other tested combos

The combination of Intel Core i7-12700KF and NVIDIA GeForce RTX 5090 D ranks 30th out of 2,924 tested combos in Steamworld Build. This places the combo in the top 1.03% of all tested configurations, which is an exceptional result. The rank of 30 means that only 29 other CPU-GPU pairings achieve higher measured FPS in this game, out of nearly three thousand combinations.

This high ranking is notable because the CPU itself is in the 85th percentile of all processors, while the GPU is in the 92nd percentile. The combo outperforms what those individual percentiles might suggest, indicating that the pairing is particularly well-suited for Steamworld Build's workload. The 12700KF's strong single-thread performance (1,043 on 3DMark single thread) likely contributes to the game's simulation logic, while the RTX 5090 D's raw throughput handles the rendering side.

The nearest CPU rivals in the database — the i7-13700T, i5-13600T, and Intel Core 5 213PE — all have average scores within 0.2% of the 12700KF. This means that swapping the CPU for any of these would likely yield nearly identical in-game results, assuming the GPU remains constant. The 12700K, which shares the same architecture, also scores within 0.2% (35,287 vs 35,365), confirming that the KF variant (without integrated graphics) does not sacrifice performance.

For the GPU, the nearest rivals are all within 2.4% in aggregate benchmarks, but the 5090 D's rank of 30 out of 2,924 suggests that in this specific game, it outperforms many higher-scoring GPUs in synthetic tests. This is likely due to the game's engine characteristics: a city builder with heavy draw calls may favor the 5090 D's 32 GB memory capacity and 1.79 TB/s bandwidth, which prevents texture thrashing in large, detailed cities.

The combo rank of 30 also indicates that there is room for improvement — the top 29 combos likely include CPUs with even higher single-thread scores or GPUs with better game-specific optimizations. However, for practical purposes, a top-30 finish out of nearly 3,000 means that this pairing is elite for Steamworld Build, and users can expect maximum or near-maximum settings to be playable at high resolutions.

Measured FPS Breakdown — resolution by resolution, settings by settings, exact numbers

The measured FPS data covers three resolutions (1920x1080, 2560x1440, 3840x2160) and four settings levels (Low, Medium, High, Ultra), for a total of 12 data points. All measurements include average FPS, with min and max FPS only provided for the Medium preset at each resolution.

1920x1080 (1080p):

• Low: 353 FPS average, no min/max provided.
• Medium: 320 FPS average, 272 FPS min, 368 FPS max.
• High: 275 FPS average, no min/max provided.
• Ultra: 186 FPS average, no min/max provided.

At 1080p, the combo is clearly CPU-bound at Low settings, since 353 FPS is far beyond what most displays can show. The progression from Low to Ultra shows a 47% drop in FPS (353 to 186), indicating that the GPU becomes increasingly loaded as settings rise. The Medium min of 272 FPS is higher than the High average of 275 FPS, which highlights the importance of the min/max range: Medium offers a tighter performance envelope.

2560x1440 (1440p):

• Low: 304 FPS average, no min/max provided.
• Medium: 234 FPS average, 199 FPS min, 269 FPS max.
• High: 175 FPS average, no min/max provided.
• Ultra: 118 FPS average, no min/max provided.

At 1440p, the Low setting still achieves over 300 FPS, but the gap between Low and Medium widens to 70 FPS (304 to 234). The High setting drops below 200 FPS for the first time, and Ultra falls to 118 FPS. The Medium min of 199 FPS is exactly at the 200 FPS threshold, meaning the combo dips just below 200 FPS in the most demanding moments at this resolution and preset.

3840x2160 (4K):

• Low: 161 FPS average, no min/max provided.
• Medium: 124 FPS average, 105 FPS min, 142 FPS max.
• High: 92 FPS average, no min/max provided.
• Ultra: 62 FPS average, no min/max provided.

At 4K, the GPU becomes the clear limiting factor. Low settings still deliver 161 FPS, which is smooth on most displays. Medium drops to 124 FPS average with a min of 105 FPS, which is the lowest min across all Medium presets. High falls to 92 FPS, and Ultra drops to 62 FPS, which is the only measured average below 90 FPS in the entire dataset. The 4K Ultra result of 62 FPS is exactly at the threshold for 60 Hz displays, but it would not take full advantage of 120 Hz or higher refresh rates.

The data shows a consistent pattern: moving from 1080p to 4K reduces FPS by roughly 54-67% depending on the settings, with Low settings showing the smallest relative drop (353 to 161, a 54% reduction) and Ultra showing the largest (186 to 62, a 67% reduction). This confirms that the RTX 5090 D scales predictably with resolution, and the 12700KF provides sufficient CPU headroom to avoid becoming a bottleneck until the GPU is pushed to its limits at 4K Ultra.