Thief

The AMD Ryzen 7 9700X paired with the AMD Radeon RX 9070 delivers a strong 1080p and 1440p experience in Thief, but the data shows a sharp performance cliff at 4K Ultra that points directly at the GPU as the limiting factor. With a combo rank of 338 out of 1,642 tested configurations, this pairing sits firmly in the top 20% of all systems, though the scaling pattern reveals that the CPU is rarely the bottleneck in this 2014 stealth title.

Resolution Scaling

The measured FPS data shows a clear and consistent drop as resolution increases, but the shape of that drop changes depending on the quality preset. At Low settings, the system delivers 226.2 FPS at 1080p, which falls to 162 FPS at 1440p—a 28.4% reduction—and then to 98.2 FPS at 4K, a further 39.4% drop from 1440p. This steep, accelerating decline at 4K Low is the signature of a GPU-bound scenario, where pixel throughput becomes the dominant constraint.

At High settings, the pattern shifts. The system produces 138.4 FPS at 1080p, 94.1 FPS at 1440p (a 32% decline), and 57.2 FPS at 4K (a 39.2% further drop). The 4K High result of 57.2 FPS is just above the 60 FPS threshold, indicating that this resolution and preset combination is right at the edge of playability. The Ultra preset tells a similar story: 129.4 FPS at 1080p, 93.2 FPS at 1440p, and a significant fall to 55.5 FPS at 4K.

The data suggests that the CPU is not the primary limiter at any resolution. The Ryzen 7 9700X’s strong single-thread performance, evidenced by a 3DMark single-thread score of 1,280 and a Cinebench R23 single-core score of 4,459, keeps frame rates high at 1080p where CPU load is proportionally larger. The consistent 28-40% drops when moving to 4K across all presets point to the RX 9070’s pixel-pushing capacity as the ceiling, not the processor.

CPU Role

The Ryzen 7 9700X brings 8 cores and 16 threads to the table, built on the Zen 5 architecture with a 4 nm process. Its base clock of 3.80 GHz and boost clock of 5.50 GHz provide the raw frequency headroom that Thief—a title from 2014—can leverage effectively. The benchmark data supports this: the CPU achieves a Cinebench R23 multi-core score of 31,588 and a single-core score of 4,459, placing it in the 89th percentile against all CPUs.

In Thief, the CPU’s role is most visible at 1080p Low, where the system hits 226.2 FPS. That figure is nearly 2.3 times the 4K Low result of 98.2 FPS, confirming that at lower resolutions the processor can feed the GPU fast enough to remove any CPU-side bottleneck. The 3DMark 16-thread score of 9,824 and max-thread score of 9,771 show that multi-threaded workloads are handled capably, but Thief’s older engine does not appear to require that full capacity; the single-thread score of 1,280 is the more relevant metric for this game’s frame pacing.

The CPU’s 32 MB of shared L3 cache and 89.6 GB/s memory bandwidth provide ample data throughput for the game’s level streaming and AI logic. Compared to its nearest rivals, the Ryzen 7 9700X posts an average benchmark score of 36,029, which is 0.2% ahead of the Intel Core i5-14500HX and 0.6% ahead of the AMD Ryzen 7 PRO 5845. It trails the Intel Core Ultra 9 185H by 0.4%, but that margin is negligible in real-world gaming scenarios.

GPU Role

The Radeon RX 9070 is the clear performance driver in this pairing, and its 16 GB of GDDR6 memory on a 256-bit bus with 644.6 GB/s bandwidth ensures that VRAM capacity is never a constraint even at 4K Ultra. The card’s boost clock of 2520 MHz and game clock of 2070 MHz, combined with 3,584 shading units and 128 ROPs, deliver a pixel rate of 322.6 GPixel/s and a texture rate of 564.5 GTexel/s.

The GPU’s benchmark profile shows a 3DMark Steel Nomad DX12 score of 6,290 and a Passmark G3D score of 25,381, placing it in the 79th percentile among all GPUs. Its nearest rival, the NVIDIA A2, scores 34,866 in average benchmark terms, a 0.2% difference, while the AMD Radeon HD 7970 sits 0.7% behind. These are synthetic scores, but the measured FPS data in Thief tells a more game-specific story.

At 4K Ultra, the RX 9070 produces 55.5 FPS, which is the lowest measured result in the entire dataset. The gap between 4K High (57.2 FPS) and 4K Ultra (55.5 FPS) is only 3%, suggesting that the GPU is already saturated and additional graphical features have diminishing returns. In contrast, at 1080p, the difference between High (138.4 FPS) and Ultra (129.4 FPS) is 6.5%, indicating more headroom. The GPU’s 16 GB VRAM is never a limiting factor; the 4K Ultra frame rate is constrained by raw compute and fill rate, not memory capacity.

How This Combo Ranks

The combination of the Ryzen 7 9700X and RX 9070 ranks 338th out of 1,642 tested combos in Thief, placing it in the top 20.6% of all configurations. This is a strong result for a pairing where both components are from the same generation and architecture family—Zen 5 CPU with RDNA 4.0 GPU.

The rank position reflects a balanced system where neither component dramatically underperforms the other. The CPU’s 89th percentile standing and the GPU’s 79th percentile standing combine to produce a result that is better than the GPU’s percentile alone would suggest, indicating that the CPU’s single-thread strength compensates for any GPU limitations at lower resolutions. At 4K, however, the combo’s rank would likely drop, as the GPU becomes the sole bottleneck and its 79th percentile performance is not enough to keep pace with higher-ranked configurations that pair the RX 9070 with even faster processors or pair faster GPUs with this CPU.

Measured FPS Breakdown

The dataset provides 12 measured FPS points across three resolutions and four quality presets. At 1920x1080, the system achieves 226.2 FPS on Low, 180.6 FPS on Medium, 138.4 FPS on High, and 129.4 FPS on Ultra. The spread from Low to Ultra is 96.8 FPS, a 42.8% reduction from the lowest preset.

At 2560x1440, the results are 162 FPS on Low, 123 FPS on Medium, 94.1 FPS on High, and 93.2 FPS on Ultra. The drop from Low to Ultra is 68.8 FPS, a 42.5% reduction—nearly identical to the 1080p scaling, which indicates the GPU is handling the extra pixel load proportionally.

At 3840x2160, the numbers fall to 98.2 FPS on Low, 72.5 FPS on Medium, 57.2 FPS on High, and 55.5 FPS on Ultra. The Low-to-Ultra reduction is 42.7 FPS, a 43.5% drop. This consistent ~43% reduction across all resolutions suggests that the quality presets scale uniformly, but the absolute frame rates at 4K drop below the 60 FPS threshold for High and Ultra.

Settings Recommendations

The data indicates that the best experience balances frame rate and visual fidelity. At 1080p, the High preset at 138.4 FPS is the sweet spot—it is only 6.5% slower than Ultra (129.4 FPS) but delivers a substantial 52.2% improvement over Medium (180.6 FPS) in terms of smoothness headroom. The 1080p Ultra result of 129.4 FPS is still well above any refresh-rate limit, so users with high-refresh monitors can choose either.

At 1440p, the High preset at 94.1 FPS is recommended. It is nearly identical to Ultra (93.2 FPS) with only a 1% difference, and both are comfortably above 60 FPS. Medium at 123 FPS is the choice for competitive play, but the visual difference between Medium and High in a stealth game like Thief—where lighting and shadow detail matter—makes High the better trade-off.

At 4K, no preset achieves a locked 60 FPS. High (57.2 FPS) and Ultra (55.5 FPS) are both below the threshold, while Medium (72.5 FPS) is the only preset that clears 60 FPS. For 4K users, Medium is the recommended preset, as it provides a 26.7% improvement over High while still maintaining playable frame rates. Low (98.2 FPS) is viable for high-refresh 4K panels, but the visual compromise is significant.

FAQ

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

A: The data shows that at 4K, the system achieves 57.2 FPS on High and 55.5 FPS on Ultra, both below 60 FPS. The Medium preset at 72.5 FPS is the only 4K setting that exceeds 60 FPS.

Q: What is the best resolution for this pairing?

A: 1440p High delivers 94.1 FPS, which is both visually rich and well above 60 FPS. At 1080p, the system can push 138.4 FPS on High, making that resolution ideal for high-refresh monitors.

Q: Is the CPU or GPU the bottleneck at 1080p?

A: At 1080p Low, the system hits 226.2 FPS, which is 2.3 times the 4K Low result of 98.2 FPS. This large gap indicates the GPU is the primary limiter even at 1080p, as the CPU’s single-thread score of 1,280 in 3DMark allows it to keep up easily.

Q: How much performance is lost going from 1080p to 4K?

A: The average FPS drops from 226.2 to 98.2 at Low settings (a 56.6% reduction), from 180.6 to 72.5 at Medium (59.9% reduction), from 138.4 to 57.2 at High (58.7% reduction), and from 129.4 to 55.5 at Ultra (57.1% reduction). The consistent ~57-60% drop indicates resolution is the dominant factor.

Q: Does the 16 GB VRAM help at 4K Ultra?

A: The 4K Ultra result of 55.5 FPS is only 3% lower than 4K High (57.2 FPS), despite Ultra typically using more memory. This small difference suggests VRAM capacity is not a constraint, and the frame rate is limited by the GPU’s compute and fill rate instead.

Q: How does this combo rank against all tested systems?

A: It ranks 338th out of 1,642 combos, placing it in the top 20.6% of configurations. The CPU’s 89th percentile and GPU’s 79th percentile combine to produce a result better than the GPU’s percentile alone would suggest.