Thief

The AMD Ryzen 5 2600 and NVIDIA GeForce RTX 5090 D pairing in Thief produces a playable, but heavily CPU-bound, experience. The data shows that while the RTX 5090 D is a top-tier GPU, it is severely constrained by the 2018-era Ryzen 5 2600, resulting in a narrow performance band across all resolutions. This combination ranks 1666th out of 2851 tested combos in this title, placing it in the lower-middle portion of the database. The most significant takeaway is that the CPU limits the frame rate to a degree where upgrading the graphics card would yield minimal gains, while lowering graphical settings does not provide the expected scaling.

FAQ

*Q: Is the AMD Ryzen 5 2600 + NVIDIA GeForce RTX 5090 D combination capable of running Thief at 60 FPS?*

A: Yes. At 4K with Ultra settings, the average frame rate is 65 FPS. At 1080p with Low settings, the average rises to 110 FPS. All tested settings and resolutions achieve an average above 65 FPS.

Q: How does the CPU’s performance compare to its nearest rivals?

A: The Ryzen 5 2600 has an average benchmark score of 21,484. This is 0.2% higher than the Intel Core Ultra 5 228V (21,440) and 0.5% higher than the Intel Core i9-11900H (21,367). It trails the Intel Xeon D-1746TER (21,635) by 0.7% and the AMD EPYC 9454 (21,223) by 1.2%.

Q: How does the GPU’s performance compare to its nearest rivals?

A: The RTX 5090 D has an average benchmark score of 77,712. This is 1.1% higher than the AMD Radeon RX 6650M XT (76,904) but 1.6% lower than the AMD Radeon RX 6850M XT (78,940). It also trails the NVIDIA Tesla P100 PCIe 12 GB (79,396) by 2.1% and the Tesla P100 PCIe 16 GB (79,605) by 2.4%.

Q: What is the best average frame rate achieved in the measured data?

A: The highest average frame rate is 110 FPS, recorded at 1920x1080 resolution with Low settings. The lowest average is 65 FPS, recorded at 3840x2160 with Ultra settings.

Q: What is the difference in average FPS between the lowest and highest settings at 1440p?

A: At 2560x1440, the average FPS ranges from 97 FPS on Low settings to 78 FPS on Ultra settings. This represents a 19 FPS difference between the two extremes.

*Q: Does the RTX 5090 D’s 32 GB of VRAM provide any benefit in Thief?*

A: While the GPU has 32 GB of GDDR7 memory, the benchmark results indicate the CPU is the limiting factor. The frame rates do not scale significantly with resolution, which suggests that the memory capacity is not a bottleneck in this specific title for this pairing.

CPU Role

The AMD Ryzen 5 2600 is a 6-core, 12-thread processor from the 2000 series, based on the Zen architecture. It has a base clock of 3.40 GHz and a boost clock of 3.90 GHz. Its 16 MB of shared L3 cache and dual-channel DDR4 memory support at 46.9 GB/s are typical for its generation. In the context of Thief, a game released in 2014, the CPU's single-threaded performance is a critical factor.

The CPU's benchmark scores reveal its capability profile. Its Cinebench R15 single-core score is 157.3, and its Geekbench single-core score is 1,154. These figures are modest by modern standards. In contrast, its multi-threaded scores are relatively stronger: Cinebench R15 multi-core is 1,248, and Geekbench multi-core is 5,648. The PassMark single-thread score is 2,239, while the multi-thread score is 13,160. This disparity suggests that while the CPU can handle multi-threaded workloads adequately, it struggles with tasks that rely on high per-core performance.

In Thief, this translates to a significant bottleneck. The game's logic and physics engine appear to rely heavily on single-core speed. The data demonstrates this: at 1080p, the average frame rate only increases from 88 FPS on Ultra to 110 FPS on Low. This 22 FPS gain from a massive reduction in graphical load indicates that the CPU is unable to feed frames to the GPU any faster. The CPU is the primary limiter, preventing the RTX 5090 D from reaching its full potential. The processor's 65W TDP and PCIe Gen 3 interface are additional constraints, but the core clock speed is the dominant factor here.

GPU Role

The NVIDIA GeForce RTX 5090 D is a colossal GPU, built on the Blackwell 2.0 architecture with a 5 nm process. It features 21,760 shading units, 680 texture mapping units, and 176 ROPs. Its memory subsystem is equally impressive: 32 GB of GDDR7 on a 512-bit bus, providing a bandwidth of 1.79 TB/s. The GPU's boost clock is 2,407 MHz. In synthetic benchmarks, it scores 44,065 in PassMark G3D and 14,326 in 3DMark Steel Nomad.

Given this hardware, the RTX 5090 D should dominate a 2014 title like Thief. Its raw compute power, with 104.8 TFLOPS of FP32 performance, is more than sufficient to render the game at maximum settings at any resolution. However, the measured FPS data tells a different story. The GPU is effectively idle, waiting for the CPU to provide draw calls and game state updates.

The data shows that the GPU's role is almost entirely negated by the CPU bottleneck. At 4K Ultra, the average FPS is 65. At 1080p Ultra, it is 88 FPS. The difference is only 23 FPS, which is remarkably small for a four-fold reduction in pixel count. This indicates that the GPU is not being utilized to its full capacity. The 32 GB of VRAM is far more than Thief requires, and the 1.79 TB/s bandwidth is similarly excessive for this workload. The RTX 5090 D is capable of far higher frame rates, but it is being held back by the Ryzen 5 2600.

Measured FPS Breakdown

The measured FPS data provides a clear picture of the system's performance across three resolutions and four settings presets. The highest performance is seen at 1080p with Low settings, where the average FPS is 110. This is the only configuration to break the 100 FPS mark.

At 1080p, the average FPS scales from 88 on Ultra, to 90 on High, 99 on Medium, and 110 on Low. The step from Ultra to Low is a 22 FPS increase. The Medium preset at 1080p has a min FPS of 84 and a max FPS of 114.

At 1440p, the average FPS figures are compressed. The Ultra preset averages 78 FPS, High averages 79 FPS, Medium averages 88 FPS, and Low averages 97 FPS. The Medium preset at this resolution shows a min FPS of 74 and a max FPS of 101.

At 4K, the average FPS is even tighter. The Ultra preset averages 65 FPS, High averages 66 FPS, Medium averages 73 FPS, and Low averages 81 FPS. The Medium preset at 4K has a min FPS of 62 and a max FPS of 84. This shows that even at the most demanding resolution, the average FPS stays above 60, but the scaling between settings is minimal.

Resolution Scaling

The scaling from 1080p to 4K is a critical indicator of the system's bottleneck. When moving from 1920x1080 Low (110 FPS) to 3840x2160 Low (81 FPS), the frame rate drops by 29 FPS, a 26% reduction. Similarly, moving from 1080p Ultra (88 FPS) to 4K Ultra (65 FPS) results in a 23 FPS drop, a 26% reduction.

This pattern of a roughly 25% drop in performance when quadrupling the pixel count is the signature of a CPU-bottlenecked system. In a GPU-bottlenecked scenario, the frame rate drop would be much more severe. For instance, a GPU-bound system might see its frame rate halve or more when moving from 1080p to 4K, as the GPU's fill rate is saturated.

Here, the GPU is so powerful relative to the CPU that the resolution increase does not significantly impact the frame rate. The limiting component is the CPU's ability to process the game's simulation. The data suggests that the RTX 5090 D is capable of rendering Thief at far higher frame rates than what is being measured. The 4K Ultra average of 65 FPS is not a reflection of the GPU's limits, but rather of the CPU's inability to issue draw calls quickly enough. The narrow FPS band across all settings and resolutions confirms that the CPU is the sole bottleneck.

Settings Recommendations

Given the extreme CPU bottleneck, the choice of settings has a muted effect on performance. However, the data still provides a basis for optimizing the experience. The primary goal should be to maximize visual fidelity without sacrificing frame rate stability, as the CPU will limit performance regardless.

For 4K, the Ultra preset yields an average of 65 FPS. The High preset is nearly identical at 66 FPS. The Medium preset offers a slight improvement to 73 FPS, while Low jumps to 81 FPS. Since the visual difference between High and Ultra is often minimal in a 2014 title, the High preset is the most efficient choice at 4K. It provides the same average frame rate as Ultra with potentially better frame pacing, and it reserves headroom for demanding scenes. The Medium preset is a viable alternative for users who prioritize a higher average frame rate over visual detail.

At 1440p, the High preset (79 FPS) is again nearly identical to Ultra (78 FPS). The Medium preset (88 FPS) offers a more noticeable boost. The Low preset (97 FPS) provides the highest average. For a smoother experience, the Medium preset is recommended, as it balances a 10 FPS gain over Ultra with better visual quality than Low.

At 1080p, the situation is similar. The Ultra preset averages 88 FPS, and High averages 90 FPS. Medium provides a more substantial jump to 99 FPS, and Low reaches 110 FPS. For users with high refresh rate monitors, the Low preset would be the only way to approach 110 FPS. However, given the game's age and the CPU bottleneck, the Medium preset at 1080p offers a strong balance of 99 FPS average and improved textures. In all cases, the data suggests that the High preset is the least beneficial option, as it offers no tangible performance advantage over Ultra while providing less visual fidelity. The Medium preset is the sweet spot for improving frame rate without a significant visual downgrade.