Thief

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

The measured frame rates across resolutions tell a clear story about where the bottleneck sits for this AMD Ryzen 7 9700X + NVIDIA GeForce RTX 4060 pairing in Thief. At 1920x1080 with Low settings, the combo delivers 184.6 FPS average. Moving to 2560x1440 Low drops that to 123.6 FPS, a 33% reduction. At 3840x2160 Low, the average falls to 77.9 FPS, another 37% cut from the 1440p figure. The scaling pattern is steep and consistent, which points toward the GPU becoming the primary constraint as pixel count rises.

Look at the High preset curve: 1080p yields 108.9 FPS, 1440p drops to 73.1 FPS, and 4K falls to 43.1 FPS. The 4K-to-1080p ratio here is roughly 0.40 — meaning 4K High runs at only 40% of the 1080p High frame rate. That is a heavy penalty, but it is also exactly what you would expect from an 8 GB GPU with a 128-bit memory bus when pushed to 8.3 million pixels. The RTX 4060's 272.0 GB/s memory bandwidth and 15.11 TFLOPS FP32 compute are simply not enough to keep frame rates high at 4K when image quality settings climb.

The Ultra preset tells an even more telling story. At 1080p Ultra, the average is 100.1 FPS. At 1440p Ultra, it sinks to 69.0 FPS. At 4K Ultra, it bottoms out at 42.7 FPS. The gap between Ultra and Low at 4K is substantial: 42.7 versus 77.9 FPS, a difference of 35.2 FPS. At 1080p, that same Ultra-to-Low gap is 84.5 FPS (100.1 versus 184.6). The shrinking absolute gap at higher resolutions strongly suggests the GPU is saturating — when the render target is massive, even lowering settings cannot fully compensate because the memory bandwidth ceiling is being hit regardless.

What does this imply about the limiting component? The CPU, with 8 cores and 16 threads running at up to 5.50 GHz boost, is clearly not the constraint here. The Ryzen 7 9700X's 3dmark single-thread score of 1280 and multi-thread score of 9771 indicate ample headroom for a 2014-era game like Thief. The data shows that resolution scaling follows the classic GPU-bound pattern: frame rate drops nearly linearly with pixel count, and settings changes have diminishing returns as resolution increases.

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

The measured rows offer four presets at each of three resolutions, and the data reveals a sweet spot that balances smoothness with visual fidelity. At 1920x1080, all four presets exceed 100 FPS average: Low hits 184.6, Medium reaches 141.8, High manages 108.9, and Ultra stays above the triple-digit mark at 100.1. For a stealth game where fluid camera movement matters, any preset works at this resolution. The jump from Low to Medium costs 42.8 FPS but likely delivers noticeably better shadows and textures; the jump from Medium to High costs another 32.9 FPS. The High-to-Ultra step is a relatively modest 8.8 FPS loss, which makes Ultra the most sensible choice at 1080p if the display supports it — the measured 100.1 FPS average remains well above typical 60 Hz refresh thresholds.

At 2560x1440, the picture changes. Low runs at 123.6 FPS, Medium at 99.1 FPS, High at 73.1 FPS, and Ultra at 69.0 FPS. Medium is the strongest recommendation here: it clears 60 FPS comfortably with a 99.1 average, leaving headroom for drops. High and Ultra both dip below 75 FPS, which may cause judder on 144 Hz monitors. The data suggests that choosing Medium at 1440p provides the best balance — the 24.5 FPS penalty from Low to Medium is acceptable, but the 26.0 FPS penalty from Medium to High is harder to justify when the visual gains are incremental for a 2014 title.

At 3840x2160, every preset struggles to maintain high frame rates. Low averages 77.9 FPS, Medium 56.6 FPS, High 43.1 FPS, and Ultra 42.7 FPS. Only Low stays above 60 FPS, and even that is marginal. Medium at 56.6 FPS is playable but not smooth by modern standards. The measured data indicates that 4K gaming with this combo in Thief requires accepting either Low settings or frame rates in the 40-57 FPS range. Given that High and Ultra produce nearly identical results (43.1 versus 42.7 FPS), there is no reason to choose High over Ultra at 4K — the 0.4 FPS difference is negligible, so Ultra is strictly better if visual quality is the priority.

GPU Role — VRAM, clocks, and how they relate to this game's results

The NVIDIA GeForce RTX 4060 brings 8 GB of GDDR6 memory on a 128-bit bus, yielding 272.0 GB/s of bandwidth. Its base clock of 1830 MHz and boost clock of 2460 MHz drive 3072 shading units, 96 TMUs, and 48 ROPs. The pixel rate of 118.1 GPixel/s and texture rate of 236.2 GTexel/s are modest figures that align with the 4K performance observed. The GPU's 15.11 TFLOPS FP32 throughput is the same as its FP16 rate (1:1), which is unusual and suggests the architecture is not heavily optimized for half-precision workloads.

How do these specs relate to Thief's results? The 8 GB VRAM capacity is not the limiting factor at these settings — Thief is a 2014 title unlikely to exceed even 4 GB at Ultra 4K. Instead, the 128-bit memory bus and 272.0 GB/s bandwidth appear to be the critical constraint. At 4K Ultra, the GPU must move massive amounts of texture and geometry data, and the measured 42.7 FPS suggests memory bandwidth starvation. The fact that 4K High (43.1 FPS) and 4K Ultra (42.7 FPS) are nearly identical reinforces this: the GPU is not compute-limited but rather bandwidth-limited, so adding more settings overhead barely changes the outcome.

The RTX 4060's 24 RT cores and 96 tensor cores are irrelevant for Thief, which predates ray tracing and DLSS. What matters is the raw rasterization throughput. The passmark_directx_11 score of 175 and passmark_directx_12 score of 76 indicate the GPU handles older DX11 titles like Thief reasonably well relative to its class. The 3dmark_steel_nomad_dx12 score of 2302 is a modern benchmark, but the game's performance aligns with the GPU's overall percentile rank of 66 among all GPUs — solidly mid-pack, which matches the 1080p and 1440p results while explaining the 4K shortfall.

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

Q: Can this combo run Thief at 4K Ultra smoothly?

A: No. The measured average FPS at 3840x2160 Ultra is 42.7 FPS, which is below the 60 FPS threshold most players consider smooth. Even at 4K High, the average is 43.1 FPS. Only 4K Low (77.9 FPS) exceeds 60 FPS.

Q: What is the best preset for 1440p gaming?

A: Medium at 2560x1440 yields 99.1 FPS average, which clears 60 FPS with substantial headroom. High (73.1 FPS) and Ultra (69.0 FPS) are playable but leave less margin for frame time spikes.

Q: How much faster is the RTX 4060 than its nearest rival in the GPU benchmark?

A: The RTX 4060's average benchmark score is 23181, which is 0.7% higher than the Intel Arc B580 (23021) and 0.8% higher than the AMD Radeon 760M (22999). It trails the AMD Radeon RX 6600M by 0.6% and the NVIDIA GeForce GTX 780 Ti by 0.7%.

Q: Does the CPU bottleneck the GPU in this game?

A: No. The Ryzen 7 9700X has a 3dmark single-thread score of 1280 and a max-thread score of 9771, placing it at the 89th percentile among all CPUs. The frame rates scale with resolution in a GPU-bound pattern, indicating the CPU has ample headroom.

Q: What is the FPS difference between High and Ultra at 1080p?

A: At 1920x1080, High averages 108.9 FPS while Ultra averages 100.1 FPS, a difference of 8.8 FPS. This is the smallest preset-to-preset gap at 1080p, making Ultra a reasonable choice.

Q: How does the combo rank among all tested combinations in Thief?

A: It ranks 834th out of 1642 tested combos, placing it in the upper-middle portion of the distribution.

How This Combo Ranks — use comboRankInGame vs other tested combos

The AMD Ryzen 7 9700X + NVIDIA GeForce RTX 4060 combination ranks 834th out of 1642 tested combos in Thief, putting it at roughly the 49th percentile of all pairings. This mid-pack positioning aligns with the GPU's own percentile rank of 66 among all GPUs, though the combo ranks slightly lower than the GPU's relative standing. The discrepancy suggests that the CPU is not the differentiator — the RTX 4060's mid-tier status is what drags the overall rank down.

Look at the GPU's nearest rivals to understand the competitive landscape. The Intel Arc B580 has an average benchmark score of 23021, just 0.7% below the RTX 4060's 23181. The AMD Radeon 760M scores 22999, 0.8% lower. On the other side, the AMD Radeon RX 6600M edges out the RTX 4060 by 0.6%, and the NVIDIA GeForce GTX 780 Ti leads by 0.7%. These are all extremely close scores — within a 1.5% band — yet the measured FPS in Thief could differ meaningfully due to driver optimizations and architecture-specific strengths. The RTX 4060's Ada Lovelace architecture, built on a 5 nm process at TSMC with 18,900 million transistors, likely delivers strong DX11 performance, which matters for a 2014 title.

The CPU side tells a similar story. The Ryzen 7 9700X's average benchmark score is 36029, placing it at the 89th percentile of all CPUs. Its nearest rival, the Intel Core Ultra 9 185H, scores 36165 (0.4% higher), while the Intel Core i5-14500HX scores 35940 (0.2% lower). The AMD Ryzen 7 PRO 5845 scores 35802 (0.6% lower), and the Intel Core Ultra 5 225H scores 36328 (0.8% higher). The CPU is clearly not holding back the combo — its 8 cores and 16 threads at up to 5.50 GHz boost are far more than Thief requires.

The 834th rank out of 1642 combos means there are 808 combinations that perform better and 833 that perform worse. Given that Thief is a 2014 stealth game with modest system requirements, this mid-pack position is somewhat surprising — it suggests that many tested combos include higher-tier GPUs that dominate the top half of the rankings. The RTX 4060's 8 GB VRAM and 128-bit bus are the likely culprits for the combo not climbing higher.

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

The AMD Ryzen 7 9700X is a Zen 5 architecture processor built on a 4 nm process at TSMC, with a die size of 70.6 mm² and 8,315 million transistors. It features 8 cores and 16 threads, with a base clock of 3.80 GHz and a boost clock of 5.50 GHz. The 65 W TDP is notably low for an 8-core desktop chip, and the L3 cache is 32 MB shared. The CPU supports DDR5 memory with dual-channel configuration and 89.6 GB/s memory bandwidth, which is ample for gaming.

In Thief, the CPU's role appears minimal based on the resolution scaling data. If the CPU were the bottleneck, frame rates would stay flat across resolutions or drop only slightly. Instead, the measured FPS drops sharply from 1080p to 4K at every preset — a classic GPU-bound signature. At 1080p Low, the combo hits 184.6 FPS, which is high enough that even a strong CPU might start to show limits, but the Ryzen 7 9700X's single-thread performance (3dmark single-thread score of 1280, cinebench R23 single-core of 4459) is more than sufficient for a 2014 game.

The CPU's multi-threaded capabilities are also far beyond what Thief needs. The cinebench R23 multicore score of 31588 and geekbench multicore score of 17002 indicate that the 8-core/16-thread configuration is not stressed by a game from 2014-02-24. Even the passmark physics score of 2241, which often correlates with gaming physics workloads, suggests no constraint. The CPU's 89th percentile ranking among all CPUs means it outperforms the vast majority of processors, and yet the combo still only manages 834th place in Thief — the GPU is the limiting factor, plain and simple.

What about the CPU's memory bandwidth of 89.6 GB/s? In theory, slow memory could starve the GPU, but the DDR5 dual-channel configuration is standard for modern platforms. The PCIe Gen 5 interface with 24 lanes (CPU only) provides ample bandwidth for the RTX 4060's PCIe 4.0 x8 bus. The data shows no signs of CPU-induced stuttering or frame pacing issues — the average FPS figures are consistent and scale predictably with resolution.

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

The complete measured FPS dataset for Thief with the AMD Ryzen 7 9700X and NVIDIA GeForce RTX 4060 covers three resolutions and four settings presets.

At 1920x1080, the Low preset averages 184.6 FPS. Medium averages 141.8 FPS, which is a 42.8 FPS reduction from Low. High averages 108.9 FPS, another 32.9 FPS drop from Medium. Ultra averages 100.1 FPS, just 8.8 FPS below High. The preset scaling at 1080p is nearly linear from Low to High, but the High-to-Ultra step is compressed, suggesting diminishing returns from the final quality tier.

At 2560x1440, Low averages 123.6 FPS. Medium averages 99.1 FPS, a 24.5 FPS drop. High averages 73.1 FPS, a 26.0 FPS drop from Medium. Ultra averages 69.0 FPS, only 4.1 FPS below High. The pattern repeats: the High-to-Ultra gap shrinks drastically at higher resolution. The 1440p Low-to-High spread is 50.5 FPS, compared to 75.7 FPS at 1080p, indicating the GPU is becoming more saturated.

At 3840x2160, Low averages 77.9 FPS. Medium averages 56.6 FPS, a 21.3 FPS drop. High averages 43.1 FPS, a 13.5 FPS drop. Ultra averages 42.7 FPS, a negligible 0.4 FPS below High. The 4K Low-to-High spread is only 34.8 FPS, and the Medium-to-Ultra range is just 13.9 FPS. This compression of preset differences at 4K is the strongest evidence of GPU bandwidth saturation — when the render target overwhelms the 272.0 GB/s memory bandwidth, settings tweaks can no longer produce large FPS swings.

Cross-resolution comparisons for the same preset show the scaling penalties. For Low, 1080p to 1440p costs 61.0 FPS, and 1440p to 4K costs 45.7 FPS. For Medium, the drops are 42.7 FPS and 42.5 FPS respectively. For High, the drops are 35.8 FPS and 30.0 FPS. For Ultra, the drops are 31.1 FPS and 26.3 FPS. The trend is clear: higher presets have smaller absolute FPS drops when resolution increases, because the GPU is already working near its limit and has less headroom to lose. The RTX 4060's 8 GB VRAM is never exceeded in these measurements, so the constraint is purely bandwidth and fill rate, not capacity.