Thief

FAQ

Q: How does the AMD Ryzen 7 5700 + NVIDIA GeForce RTX 5060 combo rank in Thief?

A: This combination ranks 781st out of 1642 tested hardware combinations in Thief, placing it in the upper-middle portion of the field. The CPU sits at the 84th percentile among all processors, while the GPU is at the 73rd percentile among all graphics cards.

Q: Can this system run Thief at 4K?

A: Yes, but with compromises. At 3840x2160, the average frame rate is 44.5 FPS on High settings and 42.5 FPS on Ultra. Dropping to Medium yields 59.1 FPS, while Low settings reach 79.6 FPS. For smooth 4K gameplay, Medium settings are the practical ceiling.

Q: What is the best resolution for high refresh rate monitors?

A: At 1920x1080, the combo delivers 146.6 FPS on Medium and 163.7 FPS on Low, making it well-suited for 144Hz displays. At 2560x1440, Medium settings produce 100.9 FPS, which still suits most high-refresh panels, but 1080p is the only resolution where Ultra exceeds 100 FPS.

Q: How does the RTX 5060 compare to its nearest rivals in synthetic benchmarks?

A: The RTX 5060's average benchmark score of 30109 is essentially tied with the AMD Radeon RX 7700 XT at 30097 (0% delta) and the AMD Radeon RX 570 at 30158 (-0.2% delta). It edges out the NVIDIA Tesla M60 (29969, +0.5%) and the RTX 5070 Mobile (29928, +0.6%).

Q: Is the Ryzen 7 5700 a bottleneck for the RTX 5060 in this game?

A: The data suggests the GPU becomes the limiting factor at higher resolutions. Frame rates drop by roughly half from 1080p to 4K at the same settings, which is typical of GPU-bound scenarios. The CPU's 8 cores and 16 threads provide ample headroom for a 2014 stealth title.

Q: What settings preset should I avoid?

A: Ultra at 4K is the weakest option, delivering only 42.5 FPS. Compared to Low at 4K (79.6 FPS), Ultra is 46.6% slower. The performance gap between High and Ultra is small at every resolution, suggesting Ultra's extra visual fidelity demands a disproportionate frame rate cost.

Resolution Scaling

The measured FPS data reveals a clear pattern of GPU-bound scaling as resolution increases. Moving from 1920x1080 to 2560x1440 on High settings drops the average from 111.8 FPS to 76.9 FPS, a 31.2% reduction. The jump from 1440p to 3840x2160 on High is even steeper, falling to 44.5 FPS — another 42.1% decrease. Across all settings, the total loss from 1080p to 4K is substantial: Low goes from 163.7 to 79.6 FPS (-51.4%), Medium from 146.6 to 59.1 (-59.7%), and Ultra from 107.4 to 42.5 (-60.4%).

This scaling behavior indicates that the RTX 5060's pixel throughput is the dominant constraint at higher resolutions. The GPU's pixel rate of 119.9 GPixel/s and 48 ROPs are modest figures for 4K workloads, explaining why 2160p frame rates hover near or below 60 FPS on most presets. The memory subsystem — 8 GB of GDDR7 on a 128-bit bus with 448.0 GB/s bandwidth — is sufficient for Thief's era, but the raw fill rate limits 4K performance.

At 1080p, the frame rates are high enough that the CPU's role becomes more visible. The Ryzen 7 5700's single-thread score of 901 in 3dmark and 2916 in Cinebench R23 suggest it can feed the GPU effectively at lower resolutions. The relatively small gap between Medium (146.6 FPS) and Low (163.7 FPS) at 1080p — just 11.7% — hints that the CPU is starting to cap performance at the very highest frame rates, though the 8-core Zen 3 design keeps this from becoming severe.

The 1440p data sits in a sweet spot: the GPU is still strong enough to deliver 75.3 to 134.9 FPS across presets, while the CPU has enough headroom that no setting combo appears CPU-limited. For players targeting 144Hz, 1080p is the only resolution that consistently reaches that threshold, and only on Low or Medium settings.

GPU Role

The NVIDIA GeForce RTX 5060 is the decisive component in Thief's frame rates. Its 8 GB of GDDR7 memory is ample for a 2014 game, even at 4K, where texture demands are modest by modern standards. The 128-bit memory bus and 448.0 GB/s bandwidth are not stressed by this title, which explains why the game scales predictably with resolution rather than hitting memory capacity walls.

The GPU's clocks — 2280 MHz base and 2497 MHz boost — are relatively high, and the Blackwell 2.0 architecture's 19.18 TFLOPS of FP32 performance provides solid compute headroom. However, the rendering pipeline is the limiting factor: 3840 shading units, 120 TMUs, and 48 ROPs produce a pixel rate of 119.9 GPixel/s and a texture rate of 299.6 GTexel/s. In Thief, which relies heavily on shadow rendering and geometry, these figures translate to the measured frame rates.

Comparing to synthetic benchmarks, the RTX 5060's PassMark G3D score of 20891 and average benchmark score of 30109 place it essentially level with the RX 7700 XT (30097, 0% delta) and RX 570 (30158, -0.2%). This parity is reflected in real-world performance: the GPU is neither a standout nor a laggard in its class. The predecessor GeForce 40 series would likely show similar scaling, though no direct comparisons are available in the data.

At 4K, the GPU's limitations become obvious. Ultra settings at 3840x2160 yield 42.5 FPS, which is below the 60 FPS threshold many players consider playable. The 5 nm process node and 21,900 million transistors pack enough compute, but the 48 ROPs are the bottleneck — a common constraint for mid-range GPUs at high pixel counts. The 73rd percentile ranking among all GPUs confirms this is a capable but not top-tier part.

Settings Recommendations

For 1080p gameplay, Medium settings at 146.6 FPS offer the best balance. The jump from Medium to High costs 34.8 FPS (down to 111.8), while High to Ultra costs only 4.4 FPS (down to 107.4). Low at 163.7 FPS is faster but sacrifices visual quality for a 11.7% gain over Medium — likely not worth it for this title's atmospheric lighting.

At 1440p, Medium at 100.9 FPS is the recommended starting point. High at 76.9 FPS is still smooth for most monitors, and Ultra at 75.3 FPS is nearly identical — a 2.1% drop from High, making Ultra a viable choice if you want maximum fidelity without meaningful frame loss. Low at 134.9 FPS is the choice for competitive play or high refresh rate panels.

At 4K, the choices are more constrained. Medium at 59.1 FPS is the only preset that approaches 60 FPS, and it should be the default. High at 44.5 FPS and Ultra at 42.5 FPS are playable but noticeably less smooth. Low at 79.6 FPS is the only preset that clears 60 FPS comfortably, but it undermines the visual experience at a resolution where detail is the point.

The data shows diminishing returns above Medium at every resolution. The average FPS gap between Medium and Ultra is consistently smaller than between Low and Medium, suggesting the game's Ultra preset adds effects that tax the GPU without proportional visual benefit. For most players, Medium or High will deliver the best experience.

Measured FPS Breakdown

At 1920x1080, the full range of presets is usable. Low produces 163.7 FPS, Medium 146.6 FPS, High 111.8 FPS, and Ultra 107.4 FPS. The gap from Low to Medium is 17.1 FPS (10.4%), from Medium to High is 34.8 FPS (23.7%), and from High to Ultra is just 4.4 FPS (3.9%). This compression at the top suggests Ultra is barely more demanding than High, making it a low-cost upgrade.

At 2560x1440, the frame rates are 134.9 FPS on Low, 100.9 FPS on Medium, 76.9 FPS on High, and 75.3 FPS on Ultra. The Medium-to-High drop of 24.0 FPS (23.8%) is the largest single step, while High-to-Ultra again shows a minimal 1.6 FPS (2.1%) difference. Low retains a significant 33.7% advantage over Medium, indicating that the Low preset cuts substantial GPU work.

At 3840x2160, the numbers are 79.6 FPS on Low, 59.1 FPS on Medium, 44.5 FPS on High, and 42.5 FPS on Ultra. The scaling is steeper here: Low-to-Medium costs 20.5 FPS (25.8%), Medium-to-High costs 14.6 FPS (24.7%), and High-to-Ultra costs 2.0 FPS (4.5%). The 4K data confirms that Ultra is nearly indistinguishable from High in performance terms, while Low is the only preset that keeps the GPU from being the bottleneck.

Across all resolutions, the pattern is consistent: Ultra adds negligible cost over High, while Low provides a substantial speed boost at the cost of visual fidelity. The CPU's role is minimal, as evidenced by the consistent 4K scaling regardless of preset.

CPU Role

The AMD Ryzen 7 5700 is a Zen 3 desktop processor with 8 cores and 16 threads, clocked at 3.70 GHz base and 4.60 GHz boost. Its 16 MB of L3 cache and 51.2 GB/s dual-channel DDR4 memory bandwidth are more than sufficient for a 2014 stealth title that predates modern multi-threaded engine demands.

Synthetic benchmarks show the CPU's strength. The Cinebench R23 multicore score of 20655 and single-core score of 2916 indicate strong per-thread performance. The 3dmark scores — 901 single-thread, 1762 for 2 threads, 3382 for 4 threads, 5585 for 8 threads, and 6617 for max threads — reveal that scaling flattens after 8 threads, which is typical for games of Thief's era. The PassMark multithread score of 24303 and single-thread score of 3296 confirm the CPU is well-balanced.

The CPU's 84th percentile ranking among all processors places it ahead of the Intel Core i9-9900K (27043, +0.7%), the Intel Core i7-12700H (27006, +0.8%), and the Intel Core i7-1370P (26991, +0.8%), while trailing the AMD Ryzen AI 5 340 (26909, +1.2%) by a narrow margin. These deltas are all under 1.2%, meaning the Ryzen 7 5700 is effectively in a performance tie with its nearest rivals — none of which would meaningfully change Thief's frame rates.

In Thief specifically, the CPU's role is secondary to the GPU. At 1080p, where the GPU has the least work, the frame rates reach 163.7 FPS on Low — but the CPU's 8 cores and high boost clock prevent any significant bottleneck. The 65 W TDP and 7 nm process also mean the CPU runs cool enough to sustain boost clocks during extended sessions, which is relevant for maintaining consistent minimum frame rates.

The 3dmark 16-thread score of 6628 and max-thread score of 6617 show that the CPU's extra threads beyond 8 provide almost no additional performance — a direct reflection of Thief's engine limitations. Players with this CPU should not expect multi-threading benefits from this title; instead, the single-thread performance (901 in 3dmark single-thread, 293 in Cinebench R15 single-core) is what matters, and it is strong enough to keep the RTX 5060 fed at all tested resolutions.