Thief

The AMD Ryzen 7 5700G paired with the NVIDIA GeForce RTX 3080 Ti delivers a compelling 1080p and 1440p experience in Thief, but the data suggests that settings choices matter more than resolution when chasing frame rate consistency. At 4K, the Ultra preset drops below the 60 FPS threshold, making the High preset the most balanced option for maintaining fluidity without sacrificing the visual fidelity that the game's stealth-oriented lighting demands.

Settings Recommendations

The measured FPS rows reveal a clear hierarchy across all resolutions: Low, Medium, High, and Ultra, in ascending order of visual demand. At 1920x1080, the gap between Low (158.7 avg FPS) and Ultra (132.7 avg FPS) is only 26 FPS, meaning even the heaviest preset remains well above the 120 Hz comfort zone. The Medium preset at 1080p delivers 144.3 avg FPS, which is only 14.4 FPS slower than Low, suggesting that the visual uplift from Medium is nearly free in performance terms. For 1440p, High emerges as the standout choice: 102.2 avg FPS versus 93.5 on Ultra, a difference of 8.7 FPS that many users would perceive in busy scenes with particle effects. The High preset at 1440p also beats the Medium preset at 4K (79.3 avg FPS) by a massive 22.9 FPS, making it the superior experience for most displays. At 4K, the recommendation shifts decisively to High (57.6 avg FPS) over Ultra (54.7 avg FPS), because the 2.9 FPS difference between them is negligible, but High's lower shadow and texture sampling loads reduce the risk of hitches. The Low preset at 4K (103.5 avg FPS) is technically playable, but it strips away the atmospheric occlusion effects that define Thief's stealth gameplay, so it is not recommended for a first playthrough.

GPU Role

The RTX 3080 Ti's 12 GB of GDDR6X memory on a 384-bit bus provides 912.4 GB/s of bandwidth, which is crucial for Thief's texture streaming. The data shows that at 1080p, the GPU is not the limiting factor: the jump from Low (158.7 FPS) to Ultra (132.7 FPS) costs only 26 FPS, indicating the CPU is still feeding frames efficiently. However, at 4K, the GPU's compute ceiling becomes apparent. The boost clock of 1665 MHz and base clock of 1365 MHz are typical for the Ampere architecture, but the real story is the raw throughput: 10240 shading units and 34.10 TFLOPS of FP32 performance. The measured FPS drop from 1080p Ultra (132.7) to 4K Ultra (54.7) is a 78 FPS collapse, which aligns with the pixel rate of 186.5 GPixel/s being saturated by the 8.3 million pixels at 3840x2160. Interestingly, the 4K Low preset (103.5 FPS) still outperforms the 1440p Ultra preset (93.5 FPS) by 10 FPS, proving that the GPU has headroom when shader complexity is reduced. The RTX 3080 Ti's 80 RT cores are present, but Thief is a 2014 title with no ray tracing implementation, so those cores sit idle; the game relies entirely on rasterization, which is why the 112 ROPs and 532.8 GTexel/s texture rate are the dominant factors in the 4K results.

Resolution Scaling

The FPS scaling from 1080p to 4K is non-linear and reveals where the bottleneck shifts. At Low settings, the progression is 158.7 (1080p) → 143.4 (1440p) → 103.5 (4K), meaning the 1440p drop is only 15.3 FPS, but the 4K drop is 39.9 FPS from 1440p. This steep cliff indicates the GPU is hitting a memory bandwidth or fill-rate wall at 4K, even with Low shaders. At High settings, the pattern is 132.3 (1080p) → 102.2 (1440p) → 57.6 (4K), with the 1440p to 4K transition costing 44.6 FPS — the largest absolute drop in the entire dataset. This suggests that High settings at 4K require more than double the pixel throughput per frame compared to 1080p, and the RTX 3080 Ti cannot sustain that load. The Medium settings show a similar trend: 144.3 (1080p) → 131.9 (1440p) → 79.3 (4K), where the 4K result is 52.6 FPS lower than 1440p. The Ultra preset is the most punishing: 132.7 (1080p) → 93.5 (1440p) → 54.7 (4K), with the 1440p to 4K drop of 38.8 FPS. Across all settings, the 1080p to 1440p transition never exceeds a 39.2 FPS loss (Ultra), but the 1440p to 4K transition always exceeds 38.8 FPS, confirming that the GPU's fillrate is the primary limiter at 4K, not the CPU.

Measured FPS Breakdown

At 1920x1080, the data shows a tight band of performance: Low leads at 158.7 FPS, followed by Medium at 144.3 FPS, Ultra at 132.7 FPS, and High at 132.3 FPS — the latter two are virtually identical, differing by only 0.4 FPS. This near-parity between High and Ultra at 1080p suggests that Ultra's extra effects (likely ambient occlusion and shadow resolution) are not taxing the GPU at this resolution. Moving to 2560x1440, the rankings shift: Low (143.4 FPS) still dominates, but Medium (131.9 FPS) pulls ahead of High (102.2 FPS) by 29.7 FPS, revealing that High settings introduce a specific rendering cost (possibly higher-resolution shadow maps) that Medium avoids. Ultra at 1440p (93.5 FPS) is the only preset to drop below 100 FPS at this resolution. At 3840x2160, the order remains Low (103.5 FPS), Medium (79.3 FPS), High (57.6 FPS), and Ultra (54.7 FPS), but the margins widen dramatically: Medium beats High by 21.7 FPS, and High beats Ultra by only 2.9 FPS. The 4K Low result of 103.5 FPS is notable because it is higher than the 1440p Ultra result (93.5 FPS), meaning a player would get a smoother experience at 4K with reduced settings than at 1440p with maxed settings. The data also shows that the gap between Low and Ultra grows with resolution: 26 FPS at 1080p, 49.9 FPS at 1440p, and 48.8 FPS at 4K — the GPU's workload scales super-linearly with pixel count when shader complexity is high.

How This Combo Ranks

The combo ranks 491 out of 1642 tested combinations in Thief, placing it in the upper 30% of all systems. This percentile positioning is interesting because the individual components have strong benchmark scores: the CPU sits at the 83rd percentile among all processors, and the GPU also sits at the 83rd percentile among all GPUs. A combo with two 83rd-percentile parts should theoretically rank higher, but the 491st position suggests that Thief's engine is not fully utilizing the RTX 3080 Ti's capabilities at lower resolutions, where the CPU becomes the bottleneck. The RTX 3080 Ti's nearest rival, the RTX 3090, scores 41441 in average GPU benchmarks — a delta of only -0.5% — yet this combo's 4K Ultra result of 54.7 FPS is only 5.9 FPS above the 4K High result of 57.6, indicating the GPU is not being pushed to its limits at 4K Ultra. The CPU's nearest rival, the Ryzen 7 6800U, has an average score of 25706, a delta of -0.3%, which is statistically negligible, reinforcing that the 5700G's 8-core, 16-thread configuration is sufficient for Thief's threading model. The 491st rank out of 1642 (70.1% of combos are faster) is a reasonable outcome for a 2021 CPU paired with a 2021 GPU running a 2014 game, but it implies that the game's engine prefers higher clock speeds over core counts — the 5700G's boost clock of 4.60 GHz helps, but its 16 MB L3 cache is modest compared to newer chips.

FAQ

*Q: Is the RTX 3080 Ti overkill for Thief at 1080p?*

A: The data suggests yes. At 1080p, even the Ultra preset achieves 132.7 FPS, which is 72.7 FPS higher than the 4K Ultra result (54.7 FPS). The GPU is not the limiter at 1080p; the CPU's 8 cores and 16 threads are sufficient to feed frames, and the GPU's 912.4 GB/s bandwidth is underutilized.

Q: What is the biggest FPS drop in the dataset?

A: The largest single drop is from 1440p High (102.2 FPS) to 4K High (57.6 FPS), a loss of 44.6 FPS. This exceeds the drop from 1440p Medium (131.9 FPS) to 4K Medium (79.3 FPS), which costs 52.6 FPS, making the Medium transition the absolute largest at 52.6 FPS. The High transition is 44.6 FPS, but the Ultra transition from 1440p (93.5 FPS) to 4K (54.7 FPS) costs 38.8 FPS.

*Q: Does the Ryzen 7 5700G's integrated graphics affect performance in Thief?*

A: No. The 5700G includes Radeon Vega 8 integrated graphics, but the system uses the discrete RTX 3080 Ti. The benchmark data is measured with the discrete GPU, and the CPU's 3DMark scores (6629 for 16 threads) are irrelevant to the game's rendering, which is handled entirely by the NVIDIA card.

Q: Which resolution gives the best experience per the measured rows?

A: 2560x1440 with High settings provides 102.2 FPS, which is above 100 FPS and visually superior to 1080p. The 4K High preset (57.6 FPS) is risky for 60 Hz displays, while 1440p High offers a 44.6 FPS margin over 4K High, making it the most balanced option.

Q: How does the RTX 3080 Ti compare to its closest rival in this context?

A: The RTX 3090 scores 41441 in average GPU benchmarks, a delta of -0.5% relative to the 3080 Ti's 41224. This 0.5% difference is negligible, and the 3080 Ti's 12 GB VRAM is sufficient for Thief at all settings, since the game's maximum measured VRAM usage is not listed but the 4K Ultra result (54.7 FPS) indicates no memory thrashing.

Q: Is the 4K Ultra preset ever worth using?

A: Only for screenshots. The 4K Ultra result of 54.7 FPS is 2.9 FPS lower than 4K High (57.6 FPS), and the visual difference between High and Ultra in Thief is minimal at 4K, where pixel density already masks many texture detail differences.

CPU Role

The AMD Ryzen 7 5700G is a Zen 3 (Cezanne) desktop processor with 8 cores and 16 threads, base clock of 3.80 GHz, and boost clock of 4.60 GHz. Its 16 MB L3 cache and dual-channel DDR4 memory bandwidth of 51.2 GB/s are modest by 2021 standards, but the benchmark data shows this is not a hindrance in Thief. The CPU's Cinebench R23 multicore score of 20755 and single-core score of 2930 indicate balanced performance, and the game's measured FPS at 1080p Low (158.7 FPS) versus 1080p Ultra (132.7 FPS) shows that the CPU can drive the GPU to high frame rates without stuttering. The 3DMark 16-thread score of 6629 and max-thread score of 6627 are nearly identical, suggesting that the CPU's scaling plateaus at 16 threads — which is exactly what Thief needs, as it is a 2014 title that likely uses fewer than 8 threads. The Passmark multithread score of 24419 and single-thread score of 3283 further confirm that the 5700G's per-core performance is strong enough to avoid bottlenecking the RTX 3080 Ti at 1080p, where the GPU's fillrate is the limiting factor. However, at 4K, the CPU's role diminishes: the FPS drops from 1080p to 4K are driven entirely by GPU workload, as evidenced by the 4K Low (103.5 FPS) being lower than the 1440p Ultra (93.5 FPS) — a counterintuitive result that only makes sense if the CPU is not the bottleneck. The 5700G's 65 W TDP and 7 nm process node are efficiency highlights, but they do not translate to gaming performance advantages. The CPU's nearest rival, the Ryzen 7 6800U, scores 25706 in average benchmarks, a -0.3% delta, meaning the 5700G is virtually identical in synthetic tests, yet the combo ranking of 491 out of 1642 suggests that Thief's engine does not reward the 5700G's architecture with higher frame rates than a typical mid-range CPU would achieve. The CPU's PCIe Gen 3 interface (16 lanes) is also a potential bottleneck for the RTX 3080 Ti's PCIe 4.0 x16 bus, but the measured FPS at 4K High (57.6 FPS) shows no PCIe bandwidth starvation, as the GPU's 912.4 GB/s memory bandwidth is the dominant data pathway.