Thief

The AMD Ryzen 9 5900X paired with the Intel Arc B580 presents a compelling question for a 2014 stealth classic: does this modern combination have enough raw throughput to push Thief past the 60 FPS threshold, and at what visual cost? The measured data shows a processor that ranks in the 87th percentile among all CPUs, paired with a GPU that sits in the 68th percentile, creating a system where the balance of responsibilities shifts dramatically depending on resolution and preset. The benchmark results indicate a clear hierarchy: at 1080p, the system is almost entirely GPU-bound, but as resolution climbs to 4K, the Arc B580 becomes the definitive bottleneck, while the 5900X’s 12 cores and 24 threads remain largely untapped. This analysis breaks down the exact frames per second across three resolutions and four presets, identifies which component is pulling the weight, and determines the most demanding settings that keep the game fluid.

Measured FPS Breakdown

At 3840x2160, the system’s performance is a study in compromise. With Low settings, the average frame rate hits 61.9 FPS, which is the only 4K preset to clear the 60 FPS playable threshold. Moving to Medium drops the average to 48.5 FPS, a substantial 21.6% reduction from Low. High settings further reduce the average to 32.6 FPS, and Ultra bottoms out at 34.4 FPS. Interestingly, the data shows that High and Ultra are nearly interchangeable at this resolution, with only a 1.8 FPS difference between them, suggesting that the Ultra preset’s additional graphical features are not the primary cost, but rather the raw resolution itself. The gap between Low and Medium is the largest single step, indicating that the Medium preset introduces effects that heavily tax the Arc B580’s 12 GB of GDDR6 memory and 192-bit bus.

At 2560x1440, the picture improves significantly. Low settings produce an average of 104.4 FPS, which is more than 60% higher than the 4K Low result. Medium follows at 78.1 FPS, comfortably above the 60 FPS target, while High drops to 58.9 FPS, just missing the threshold. Ultra comes in at 55 FPS, showing a narrower spread between presets than at 4K. The transition from 4K to 1440p yields a 68.6% performance boost on Low settings, but only a 60.9% boost on High, which reinforces that the GPU is the limiting factor at higher resolutions. The fact that High at 1440p (58.9 FPS) is nearly identical to 4K Low (61.9 FPS) suggests that the game’s High preset at 1440p is roughly equivalent in GPU load to Low at 4K, which is a useful reference point for players.

At 1920x1080, the system finally stretches its legs. Low settings deliver an average of 149.8 FPS, making the game exceptionally fluid. Medium hits 115 FPS, High reaches 85.4 FPS, and Ultra maintains 82.7 FPS. The difference between High and Ultra at 1080p is a mere 2.7 FPS, which is within the margin of error for a game from 2014. The scaling from 1440p to 1080p is dramatic: Low settings improve by 43.5%, while Ultra improves by 50.4%. This indicates that at 1080p, the CPU’s strong single-thread performance (as evidenced by its 3dmark_single_thread score of 941) is finally able to feed the GPU enough frames to saturate it, though the Arc B580 still caps out at just under 150 FPS on Low.

CPU and GPU Roles

The data reveals a classic GPU-bound scenario at higher resolutions, with the CPU’s role diminishing as pixel count increases. At 3840x2160, the Ryzen 9 5900X’s 12 cores and 24 threads are largely idle, as the Arc B580’s 13.67 TFLOPS of FP32 compute cannot keep up with the rasterization demands. The 4K Low result of 61.9 FPS versus the 1080p Low result of 149.8 FPS shows a 142% performance gap, which is almost entirely attributable to GPU fill rate and memory bandwidth. The Arc B580’s 456.0 GB/s of bandwidth is sufficient for 4K Low, but the 48.5 FPS at 4K Medium suggests that higher preset levels require more than just bandwidth — they demand more shading units and texture units, and the B580’s 160 TMUs and 80 ROPs become the limiting factor.

At 2560x1440, the balance begins to shift. The 78.1 FPS on Medium settings is a 61% improvement over 4K Medium, indicating that the GPU is still the primary constraint, but the CPU’s 3dmark_8_threads score of 6601 suggests it can handle the extra draw calls. The 5900X’s 64 MB of L3 cache helps maintain consistent frame pacing, which is critical for a stealth game where sudden frame drops can ruin immersion. However, the fact that 1440p High (58.9 FPS) falls below the 60 FPS line while 1440p Medium (78.1 FPS) clears it by a wide margin shows that the High preset introduces specific effects (likely shadow quality or ambient occlusion) that disproportionately impact the B580.

At 1920x1080, the CPU’s relevance becomes more apparent, but it is still not the bottleneck. The jump from 1440p Medium (78.1 FPS) to 1080p Medium (115 FPS) is a 47% increase, which is close to the pixel count reduction of 56% (from 3.7 million to 2.1 million pixels). This near-linear scaling suggests that the GPU is still doing the heavy lifting, but the CPU’s maximum thread performance (3dmark_max_threads score of 10075) ensures that it can feed the GPU without stalling. The 5900X’s single-core score of 941 in 3dmark_single_thread is particularly relevant here, as Thief is a game from 2014 that likely relies on a few heavily threaded tasks rather than scaling across all 24 threads.

Best Settings per Resolution

For players targeting a smooth 60 FPS experience, the verdict is clear-cut. At 3840x2160, the only viable option is Low settings, which produce an average of 61.9 FPS. This is a marginal pass, and any scene with heavy particle effects or multiple light sources could dip below the threshold, but the data shows it is technically playable. The step up to Medium at 48.5 FPS is too large a sacrifice, so 4K players must accept Low to maintain fluidity.

At 2560x1440, the best preset that stays at or above 60 FPS is Medium, delivering 78.1 FPS. This is the sweet spot for this combination, offering a 26% headroom above the threshold, which should absorb most frame time spikes. High settings at 58.9 FPS are tantalizingly close, but the data does not support recommending them for a locked 60 FPS experience.

At 1920x1080, the system can run Ultra settings with an average of 82.7 FPS, which is 37.8% above the 60 FPS mark. This is the clear recommendation for 1080p owners, as the difference between High (85.4 FPS) and Ultra (82.7 FPS) is negligible, and Ultra provides the full visual experience. The 149.8 FPS on Low is impressive for competitive play, but for a stealth game where visual fidelity matters, Ultra is the logical choice.

The Verdict

Can this PC run Thief? The measured data says yes, with significant caveats based on resolution. At 1920x1080, the system clears 60 FPS on all four presets, with Ultra averaging 82.7 FPS — this is the definitive playable configuration. At 2560x1440, the system clears 60 FPS on Low (104.4 FPS) and Medium (78.1 FPS), but High (58.9 FPS) and Ultra (55 FPS) fall short. At 3840x2160, only Low (61.9 FPS) passes the threshold, while Medium (48.5 FPS), High (32.6 FPS), and Ultra (34.4 FPS) are not viable for smooth gameplay. The verdict is that this is a solid 1080p machine for Thief, a capable 1440p machine with reduced settings, and a marginal 4K machine that requires the lowest preset. The 5900X is never the limiting factor in this game; the Arc B580’s mid-range performance profile defines the boundaries.

Similar Performance Alternatives

Looking at the nearest rivals for the CPU, the Ryzen 9 5900X’s average benchmark score of 41376 places it within 0.5% of several other processors. The Intel Core Ultra 7 366H scores 41263, which is 0.3% slower, and the Intel Core Ultra 7 356H scores 41215, 0.4% slower. The AMD Ryzen AI 5 PRO 440 matches at 41208, also 0.4% slower, while the Intel Core i7-14650HX is the only one that edges ahead at 41576, a 0.5% delta. In practice, any of these CPUs would deliver nearly identical frame rates in Thief, as the game’s engine is not heavily dependent on multi-threaded performance beyond 8 threads. The 5900X’s 12-core advantage is irrelevant here, so users with these rivals would see the same GPU-bound behavior at 1440p and 4K.

For the GPU, the Intel Arc B580’s average score of 23021 is bracketed by some surprising competitors. The AMD Radeon RX 580 2048SP scores 23061, just 0.2% higher, meaning it would perform virtually identically in Thief. The NVIDIA GeForce RTX 2080 scores 22895, 0.6% lower, while the NVIDIA GeForce RTX 3080 scores 23172, 0.7% higher. The NVIDIA P106-100 scores 23249, 1% higher. This is a remarkable clustering, with all four rivals within a 1.5% band. In practical terms, a system with any of these GPUs would produce the same FPS numbers at each resolution and preset, within a frame or two. The RTX 3080’s higher bandwidth and compute would not translate to a meaningful advantage in this 2014 title, as the game’s rendering pipeline is not optimized for modern GPU features.

How This Combo Ranks

In the database of tested combinations, this specific pairing of the Ryzen 9 5900X and Arc B580 ranks 2804 out of 3720 tested combos for Thief. This places it in the 24.6th percentile, meaning roughly three-quarters of tested systems outperform it in this particular game. This ranking might seem low given the 1080p Ultra result of 82.7 FPS, but it reflects the fact that Thief is an older title that favors high single-thread CPU performance and strong GPU rasterization. The 5900X’s 87th percentile CPU ranking is offset by the B580’s 68th percentile GPU ranking, and the combination lands in the lower-middle of the pack. The data suggests that many systems with older, higher-end GPUs (like the RTX 3080, which is a close rival) would rank higher due to better raw fill rate, even if their CPUs are weaker. The ranking also implies that for players who prioritize this specific game, a more balanced pairing with a slightly weaker CPU and stronger GPU would yield better results.

FAQ

Q: Can this system run Thief at 4K with High settings?

A: No. The measured average FPS at 3840x2160 with High settings is 32.6 FPS, which is well below the 60 FPS playable threshold. The only 4K preset that clears 60 FPS is Low, at 61.9 FPS.

Q: What is the highest preset that maintains 60 FPS at 1440p?

A: Medium settings at 2560x1440 produce an average of 78.1 FPS, which is the most demanding preset that clears the 60 FPS line. High settings fall short at 58.9 FPS.

Q: How much faster is 1080p Ultra compared to 1440p Ultra?

A: At 1920x1080, Ultra averages 82.7 FPS, while at 2560x1440, Ultra averages 55 FPS. This represents a 50.4% improvement in frame rate when dropping from 1440p to 1080p.

Q: Is the Ryzen 9 5900X or the Arc B580 the bottleneck at 4K?

A: The GPU is the definitive bottleneck at 3840x2160. The 5900X’s 3dmark_max_threads score of 10075 is more than sufficient, but the Arc B580’s 13.67 TFLOPS compute cannot sustain high preset levels, as shown by the 32.6 FPS on High versus 61.9 FPS on Low.

Q: Does the Arc B580 perform similarly to an RTX 2080 in this game?

A: Yes. The B580’s average benchmark score is 23021, while the RTX 2080 scores 22895, a 0.6% difference. In Thief, this translates to nearly identical frame rates at every resolution and preset.

Q: What is the best overall setting for 1080p gameplay?

A: Ultra settings at 1920x1080 deliver 82.7 FPS, which is above the 60 FPS target and provides the full visual experience. The difference between High (85.4 FPS) and Ultra is only 2.7 FPS, so Ultra is the recommended choice.