Thief

Resolution Scaling

The measured FPS data for Thief reveals a clear and predictable scaling pattern as resolution increases, with the most dramatic drop occurring between 1080p and 4K. At 1080p with Ultra settings, the combo of the Intel Core Ultra 7 265K and NVIDIA GeForce RTX 3080 Ti delivers 143 FPS. Moving to 1440p at Ultra drops that to 94.5 FPS, a reduction of roughly 34%. The jump to 4K at Ultra produces 58.8 FPS, which represents another 38% decline from the 1440p figure. This progressive decline is characteristic of a GPU-bound scenario at higher resolutions, where pixel throughput becomes the dominant constraint.

The scaling behavior under Low settings is even more telling. At 1080p Low, the system reaches 244 FPS, which drops to 175.7 FPS at 1440p and then to 103.5 FPS at 4K. The 4K-to-1080p ratio at Low is about 42%, meaning 4K delivers less than half the frame rate of 1080p. Compare that to Ultra settings, where 4K delivers 58.8 FPS versus 143 FPS at 1080p, a ratio of roughly 41%. The consistency of this ratio across settings strongly indicates that the RTX 3080 Ti is the limiting component even at Low settings; if the CPU were the bottleneck, lower settings would show a much flatter resolution curve because the GPU would have spare capacity.

The gap between High and Ultra at 4K is minimal: 59.4 FPS versus 58.8 FPS, a difference of only 0.6 FPS. This near-identical performance suggests that at 4K, the Ultra preset's additional graphical effects are largely overshadowed by raw pixel count, and the GPU is saturated regardless. Conversely, at 1080p, the difference between High (149.1 FPS) and Ultra (143 FPS) is 6.1 FPS, indicating that at lower resolutions the extra Ultra effects do have measurable cost. The data shows that resolution scaling in Thief follows an expected curve, but the severity of the 4K drop—cutting frame rates by more than half compared to 1080p in every setting tier—underscores that 4K gaming on this combo will require careful settings management to maintain smooth play.

GPU Role

The NVIDIA GeForce RTX 3080 Ti is the primary driver of the frame rates observed in Thief, and its specifications explain the performance ceilings seen at each resolution. The GPU features 12 GB of GDDR6X memory on a 384-bit bus, yielding 912.4 GB/s of bandwidth. For a game released in 2014, this memory configuration is far beyond what Thief requires, which is why the data shows no evidence of VRAM capacity limiting performance even at 4K Ultra. The card's 10,240 shading units and 112 ROPS provide substantial fill-rate capability, translating to the high pixel rates necessary for 4K rendering.

The clock behavior matters for understanding the FPS numbers. The RTX 3080 Ti runs at a base clock of 1365 MHz and boosts to 1665 MHz. These clocks, combined with the 34.10 TFLOPS FP32 throughput, explain why the card can push 103.5 FPS at 4K Low—a scenario where the pixel workload is reduced but the geometry and draw-call processing still demand significant compute. The GPU's benchmark scores place it in the 83rd percentile of all GPUs, with an average benchmark score of 41,224, making it a high-tier but not flagship part.

Comparing the RTX 3080 Ti to its nearest rivals in the benchmark database shows a tightly clustered group. The RTX 3090 scores 41,441, a delta of -0.5% relative to the 3080 Ti. The AMD Radeon Pro 5300 scores 41,610 (-0.9%), and the RTX 5070 scores 41,687 (-1.1%). These sub-2% deltas indicate that the 3080 Ti is statistically indistinguishable from these competitors in synthetic benchmarks, yet in Thief the measured FPS tells a more nuanced story. The 4K Ultra result of 58.8 FPS is right at the edge of playability, and the fact that the 3080 Ti cannot maintain 60 FPS at 4K Ultra despite its high memory bandwidth suggests that Thief's rendering pipeline is sensitive to raw shader throughput rather than memory speed alone.

The GPU's 8 nm process and 350 W TDP—with a suggested PSU of 750 W—indicate a power-hungry design, but the benchmark data does not show thermal throttling affecting results, as the 4K High score of 59.4 FPS is nearly identical to Ultra, implying the card is consistently hitting its boost behavior rather than degrading over time.

How This Combo Ranks

The Intel Core Ultra 7 265K paired with the RTX 3080 Ti ranks 193rd out of 1,642 tested combos in Thief, placing it in the top 11.8% of all configurations. This is a strong showing, but the rank relative to the combo's component scores reveals an interesting discrepancy. The CPU sits in the 97th percentile of all CPUs with an average benchmark score of 79,102, while the GPU is in the 83rd percentile. The combo's rank in Thief is better than what the GPU percentile alone would predict, suggesting that Thief favors the CPU's strengths more than typical titles.

The CPU's nearest rivals show how close the top of the CPU market is. The Intel Core Ultra 7 265KF scores 79,240, a delta of -0.2% from the 265K, making them effectively identical. The Intel Core i9-14900KF scores 79,245 (-0.2%), and the Core i9-14900K scores 79,824 (-0.9%). All of these are within 1% of each other, meaning any of these CPUs would produce nearly identical Thief performance when paired with the same GPU. The AMD EPYC 7413 scores 80,041 (-1.2%), which is a server part and not a direct gaming comparison, but its inclusion in the nearest rivals list indicates the 265K's multicore throughput is competitive with enterprise hardware.

The combo's 193rd rank out of 1,642 means there are 192 configurations that outperform it in Thief. Given that the GPU is only in the 83rd percentile, many of those higher-ranked combos likely pair the 265K with faster GPUs, such as the RTX 3090 or RTX 5070, which the benchmark data shows are within 1.1% of the 3080 Ti's synthetic score. This narrow gap suggests that the combo's rank is not limited by the CPU but rather by the GPU's position relative to newer or higher-tier cards. In practical terms, this combo is well above average for Thief, and the data indicates the CPU has headroom to support a GPU upgrade without becoming a bottleneck.

FAQ

Q: Can this combo run Thief at 4K 60 FPS?

A: Only with settings adjustments. The measured data shows 4K High delivers 59.4 FPS and 4K Ultra delivers 58.8 FPS, both just below 60 FPS. Dropping to 4K Medium raises the average to 77.2 FPS, which clears the 60 FPS threshold comfortably.

Q: What is the biggest performance difference between settings at 4K?

A: The gap between Low and High at 4K is substantial. Low settings produce 103.5 FPS, while High produces 59.4 FPS—a difference of 44.1 FPS. Medium lands at 77.2 FPS, showing a steep curve where each preset tier costs roughly 20-25 FPS.

Q: How does this combo compare to the RTX 3090 in synthetic benchmarks?

A: The RTX 3080 Ti scores an average benchmark of 41,224, while the RTX 3090 scores 41,441, a delta of -0.5%. This places them within statistical noise, meaning real-world game performance should be nearly identical.

Q: Is the CPU or GPU the limiting factor in Thief?

A: The GPU is the limiting factor. Evidence comes from the resolution scaling: at 1080p Low the system hits 244 FPS, but at 4K Low it drops to 103.5 FPS—a 57.6% reduction. If the CPU were limiting, the 4K Low result would be much closer to the 1080p Low result.

Q: What is the best resolution to play at for a 144 Hz monitor?

A: At 1440p, all settings exceed 94 FPS, with Low hitting 175.7 FPS, Medium 132.9 FPS, and High 102.1 FPS. Only Ultra at 1440p drops to 94.5 FPS, which is still above 90 FPS. For a 144 Hz display, 1440p High or Medium offers the best balance.

Q: How large is the CPU's benchmark lead over its nearest rival?

A: The Intel Core Ultra 7 265K scores 79,102 on average, while the 265KF scores 79,240 (-0.2%) and the i9-14900KF scores 79,245 (-0.2%). The 265K is statistically tied with both, with a negligible 0.2% deficit.

CPU Role

The Intel Core Ultra 7 265K plays a supporting but essential role in Thief, and its specifications reveal why the combo achieves the measured frame rates. The CPU features 20 cores and 20 threads, with a base clock of 3.90 GHz and a boost clock of 5.50 GHz. This high boost clock is critical for Thief's single-threaded rendering tasks, as the game's engine from 2014 does not scale perfectly across many cores. The 30 MB of shared L3 cache provides ample room for game data, reducing the need to fetch from system memory, which runs on dual-channel DDR5 with 102.4 GB/s of bandwidth.

The CPU's benchmark results show exceptional multicore performance, with a Cinebench R23 multicore score of 50,009 and a single-core score of 7,060. These figures place the CPU in the 97th percentile of all CPUs, confirming that it is a top-tier part for gaming workloads. In Thief, the CPU's role is to feed the GPU with draw calls and physics calculations without becoming a bottleneck. The data confirms this, as the FPS at 1080p Low (244 FPS) is high enough that even a modest CPU would likely show limitations, yet the 265K maintains that rate, indicating it is not the constraint.

The CPU's architecture, Arrow Lake-S, built on a 3 nm process from TSMC, delivers high efficiency, but the TDP of 125 W means it draws less power than the GPU's 350 W. The 20 cores are arranged with the boost clock of 5.50 GHz enabling strong single-thread performance, which is what Thief's engine primarily utilizes. The PassMark single-thread score of 4,904 and multithread score of 58,838 illustrate the balance: the CPU is nearly 12 times faster in multithreaded workloads than single-threaded, but for gaming, the single-thread score is the better predictor of performance.

Comparing the 265K to its nearest rivals, the delta values are all within -1.2% to -0.2%, meaning the CPU choice between the 265K, 265KF, i9-14900KF, or i9-14900K would not meaningfully change Thief's FPS. The CPU's 97th percentile rank versus the GPU's 83rd percentile means the CPU has headroom; in Thief, this manifests as the ability to sustain high frame rates at 1080p without the CPU limiting the GPU.

Settings Recommendations

Based on the measured FPS data, the optimal settings for Thief depend heavily on the target resolution and desired frame rate. For 4K gaming, the data shows that Medium settings at 77.2 FPS is the sweet spot, as it provides a 30% improvement over High (59.4 FPS) while maintaining visual quality that is likely to be noticeably better than Low (103.5 FPS). Ultra at 4K is not recommended because it produces 58.8 FPS, which is below the 60 FPS threshold and only 0.6 FPS lower than High—the extra visual fidelity is not worth the performance cost.

At 1440p, the choice is more flexible. Ultra settings deliver 94.5 FPS, which is above 90 FPS and suitable for high-refresh gaming, while High offers 102.1 FPS. If the monitor supports 144 Hz, then Medium at 132.9 FPS is the best option to stay above 120 FPS, leaving headroom for frame time spikes. Low at 175.7 FPS is unnecessary unless the user prioritizes maximum responsiveness over image quality.

For 1080p, the data shows that even Ultra settings deliver 143 FPS, which is well above 60 FPS and close to 144 Hz. High at 149.1 FPS is nearly identical, so the recommendation is to use Ultra if the GPU can handle it—the 6.1 FPS difference between High and Ultra is minor. Medium at 194.5 FPS and Low at 244 FPS are available for competitive play or very high refresh rate displays, but they sacrifice visual fidelity for frame rates that most monitors cannot display.

The overall recommendation is to prioritize Medium at 4K, Medium or High at 1440p depending on refresh rate, and Ultra at 1080p. The data does not support using Low settings at any resolution unless the user has a 240 Hz or higher monitor, as even 4K Low (103.5 FPS) exceeds the refresh rate of most standard displays.

Measured FPS Breakdown

At 1920x1080, the measured average FPS across settings is as follows: Low produces 244 FPS, Medium produces 194.5 FPS, High produces 149.1 FPS, and Ultra produces 143 FPS. The spread from Low to Ultra is 101 FPS, indicating that the Ultra preset's additional effects cost significant performance even at 1080p. The difference between High and Ultra is only 6.1 FPS, suggesting that Ultra is a minor upgrade over High at this resolution.

At 2560x1440, the results are: Low at 175.7 FPS, Medium at 132.9 FPS, High at 102.1 FPS, and Ultra at 94.5 FPS. The spread from Low to Ultra is 81.2 FPS, slightly narrower than at 1080p, reflecting the GPU becoming more of a limiting factor. The gap between High and Ultra grows to 7.6 FPS, indicating that the Ultra preset's effects become relatively more expensive as pixel count rises.

At 3840x2160, the measured averages are: Low at 103.5 FPS, Medium at 77.2 FPS, High at 59.4 FPS, and Ultra at 58.8 FPS. The spread from Low to Ultra is 44.7 FPS, the smallest of all resolutions, confirming that the GPU is heavily saturated at 4K. Notably, the difference between High and Ultra shrinks to 0.6 FPS, meaning the Ultra preset is essentially free at 4K in terms of performance—both are below 60 FPS. Medium is the only 4K setting that clears 60 FPS with a margin of 17.2 FPS, making it the clear choice for playable 4K performance.

Comparing across resolutions, the 1080p Low result of 244 FPS is the highest measured value, while the 4K Ultra result of 58.8 FPS is the lowest. The scaling from 1080p to 4K at each setting is: Low drops by 57.6%, Medium by 60.3%, High by 60.2%, and Ultra by 58.9%. These consistent percentages reinforce that the GPU is the bottleneck across all settings, as the frame rate reduction tracks with the pixel count increase (4K has 4 times the pixels of 1080p, but the FPS does not drop to 25% of the 1080p value because the CPU and memory bandwidth provide partial compensation).