Thief

Measured performance for the AMD Ryzen 5 2600 and Intel Arc B580 in Thief (2014) shows a wide range depending on resolution and quality preset. At 1080p Low the pair leads to an average of 110 frames per second, while at 4K Ultra that figure drops to 33. The dataset covers three resolutions and four settings, giving a precise picture of scaling. Combo rank in this game is 2,973 out of 3,720 tested configurations, placing it below the majority of its peers, despite the CPU being in the 75th percentile of all processors and the GPU in the 68th percentile of all GPUs.

Settings Recommendations

Across the measured configurations, the most playable balance of visual quality and frame rate appears at 1080p Medium. There the average is 99 FPS, with a recorded min of 84 and max of 114—tight enough for consistent motion on a 60 Hz display. Stepping down to 1440p, the Medium preset still averages 81 fps? Wait, need to use exact data: 1420p Medium is 78 fps average with 66–90 min/max. So if the monitor is 1440p, Medium remains within a comfortable range, though the minimum of 66 is close to the 60 Hz threshold. At 1080p, even the Ultra preset yields 83 FPS, making it a worthwhile choice for that resolution. 1080p High averages 87 FPS, and Low reaches 110, so on 1080p displays with high refresh rates, Low provides the headroom to exploit that refresh.

For 1440p targeting fluid 60 FPS gameplay, the High preset at 60 FPS exactly matches the threshold; however, the average is exactly 60, with occasional variances likely because the dataset records no min/max for that row. The Medium preset averages 78, comfortably above that line, and conversely drops from medium to High cost about 18 FPS (78→60). Moving to 4K, the situation is drastically different. At 4K, Low is the only preset above 60: it averages 61 FPS. Medium falls 5 FPS to 46 average, High 35, Ultra 33. For owners of 4Kscreens, the Low preset is the only practical target for a smooth 60 fps experience. This data suggests that the B580 lacks the computational headroom to maintain high settings at 4K, even for older titles. Higher quality presets at 4K are effectively a slide for motion intenst of 33–46 FPS.

Given the game is from 2014, but the data does not support that assumptions; the only measured numbers show that 1080p24 Ultra is playable, 1440p Medium/High at ~60 FPS, and 4K only on Low. Therefore, the recommended setting for a consistent 60 FPS experience is 1080p October 6, 1080p Gang Ultimate. Medium is the sweet spot considering frametime variation; the recorded Min/Max of 84–114 frames shows the narrowest spread of all 88 presets.

GPU Role

Active The Intel Arc B580’s specifications go to play in the results. It features 12 GB of GDDR6 memory on a 192-bit bus, giving a memory bandwidth of 456.0 GB/s. The core clock runs at both base and boost at 2670 MHz. Those numbers, combined with 3D Mark Steel Nomad, transfer the gamma at Gamma's performance position. The GPU is (MCu Order) 3DMark Steel Nomad DX12 score of 3,068, Geekbench Vulkan 109,672, and Passmark G3D 15,748 give an average weighted score in the upper mid range. The GPU’s percentile ranking vs all GPU's is 68. That means a third of GPUs hold a higher benchmark score than this one ) but in Thief, results are more variable.

The data shows a mild instruction: “the framerate decreases sharply as resolution increases.” At 1080p, even the Ultra preset averaged 83 FPS; at 1440p that falls to 57, and at 4K averages 35. This 48 FPS drop from 1080p to 4K at 4K tends to indicate that the GPU's memory subsystem and compute are the limiting factor, not the CPU. The B580’s 2560 shading units at 2670 MHz, and by them the 192-bit memory interface delivers 456 GB/s, which is enough at lower pixels per pixel, but insufficient to maintain high settings at 4K. The rising result is that 4K High hits only 35 FPS and 4K Ultra 33, whereas a strong, capable GPU could plausibly offer playable framerates at high settings—a capability not visible in the measured data.

The GPU’s synthetic average puts close percentile, rope of (average customer score 23,021) among competitors such as the RTX 2080 (average 22,895, +0.6 delta) and RX 580 2048SP (23,061, -0.2). But those comparisons stem from FPS across all scenes. It is worth repeating that there is no amount of these specs that leads to a conclusion beyond the fact that the B580 is cast in the 68th percentile; the actual intensity brings dramatic scaling.

FAQ

Q: What are the average framerates for each preset at 1920×1080?

A: At 1080p, Low averages 110 FPS, Medium 99 FPS (min 84, max 114), High 87 FPS, Medium and High covered also but no min/max for High. Ultra 83 FPS.

Q: Is this combination capable of 60 FPS at 1440p?

A: The data shows that 1440p Medium averages 78 FPS (min 66), High averages 60 FPS (the exact threshold), and Ultra 57 FPS. Thus, only the Medium preset enables a comfortable window over 60 FPS; High sits at 60, which may dip below without recorded minima.

Q: Which quality levels might work at 4K?

A: At 4K, the Low preset averages 61 FPS. Middle is 46, High 35, and Ultra 33. The only preset entirely at or above 60 fps; Low is the only suitable for consistent 60 frames-per-second.

Q: How many CPU cores and what clo for this processor?

A: The Ryzen 5 2600 has 6 cores, 12 threads, with a base clock of 3.40 GHz and boost of 3.90 GHz. Its TDP is 65 W, and it uses 16 MB L3 cache.

Q: How much VRAM and memory representative does the Arc B580 have?

A: The Intel Arc B580 has 12 GB GDDR6 memory with a 192-bit bus and a memory bandwidth of 456.0 GB/s. The base and boost clock temporal both 2670 MHz.

Q: Does the combo have anywhere near playable average at 4K?

A: No, aside from Low at 61 FPS, all other 4K presets are below 60. Mid 46, High 35, Ultra 33, making the pair unsuitable for high settings on a 4K screen.

Measured FPS Breakdown

Below is the measurement matrix as recorded from the test run.

| Resolution | Setting | Average FPS | Min FPS | Max FPS |

|---|---|---|---|---|

| 1920×1080 | Low | 110 | – | – |

| 1920×1080 | Medium | 99 | 84 | 114 |

| 1920×1080 | High | 87 | – | – |

| 1920×1080 | Ultra | 83 | – | – |

| 2560×1440 | Low | 97 | – | – |

| 2560×1440 | Medium | 78 | 66 | 90 |

| 2560×1440 | High | 60 | – | – |

| 2560×1440 | Us | 57 | – | – |

| 3840×2160 | Low | 61 | – | – |

| 3840×2160 | Medium | 46 | 39 | 53 |

| 3840×2160 | High | 35 | – | – |

| 3840×2160 | Ultra | 33 | – | – |

Only the Medium presets at each resolution have recorded min/max values. The absence of those values for other presets does not imply tighter—it means they were not registered. The data shows that framerate variance, with min/max spread of 30 FPS at 1080p Medium (84–114) and 14 FPS at 1440p Medium, suggests that the maximum frame rate is larger at lower resolutions.

Concurrently with, resolution.

When switching from 1080p to 1440p the observed averages fall by 18–24%: for example, High 87→60 is no 31% drop, while Low 110→97 is a 12% drop. The shift from 1440p to 4K is steeper: For High, 60→35 is a 42% drop; for Low, 97→61 is a 37% drop. Complete drop from 1080p to 4K shows a reduction of almost half the fps: 110→61 at High is 45% loss? Actually high: 87→35 is 60% loss. The then-state is consistent with the GPU limitations as pixel count increases. The fact that the Low preset, the easiest for the GPU, also falls from 110 to 61 (44% drop) means that the scaling is not conditional on fine-detail geometry, but purity–fill bandwidth. Conversely, the CPU Total spikes at 1080p Low (110 FPS) indicate that at high frame rates, the GPU can still deliver, so it's the payload of each pixel that drives.

Except for readouts, the table shows that the absolute drop is almost monotonic. The credit to the x depends only on the resolution and the quality of the instantiation. At 4K, the B580’s 456 GB/s memory bandwidth still does not provide enough throughput to do the full pixel count needed for 60 FPS with any quality preset other than Low. This is also an oblique but honest outcome: for if the game is internally driven by something like a 2014-era game engine, that isn’t visible in the raw numbers; the output is.

The de facto summary: 1080p and 1440p are playable at medium or high; 4K is only viable LGT.

CPU Role

The Ryzen 5 2600 is a 6-core, 12-thread processor with base 3.40 GHz and boost 3.90 GHz. Using 12 nm-architecture Zen+ (Pinnacle Ridge), it ships 16 MB L3 cache (shared), and dual-channel DDR4-46.9 GB/s bandwidth. Its synthetic scores are strong litmus results: Cinebench R15 multi-core 1,248, single-core 157.3, Geekbench multi 5,648, single 1,154, PassMark multi-core 13,160, single threaded 2,240 (the reported score is 2239 for passmark_single_thread and 2239 also). It holds a 74th percentile among all CPUs.

Comparing in nearest rivals: the Ryzen 2600 performs within 1% of Intel Core Ultra 5 228V (delta +0.2%), Intel Core i9-11900H (+0.5%), Intel Xeon D-1746TER (-0.7%), and Intel EPYC 9454 (+1.2%). In plain terms, its computing performance is very good enough to not be the slowdown in Thief scenarios.

The game itself, because a period piece from 2017, the CPU’s challenge is of battlefield compute. The measured FPS at 1080p Low reaches limit 110, which is a clear upper envelope that the GPU at this preset can produce. The difference between Low (110) and Ultra (83) sets a 23% spread, meaning that the GPU cost per-heavy settings is enough that the GPU is the principal part of the mix. The CPU's 12 engines would almost certainly feed frames beyond that if needed.

At 1440p Medium the average of 78 (max 90) indicates that still the CPU can drive the scenes; the GPU's compute / memory starts to person it. At 4K Low 61 vs 4K Medium 46 fall—not the CPU, but the GPU. Given the CPU’s percentile, the utilization is unlikely to be a constraining factor. The data for 4K might be a direct evidence for GPU bound status.

An informative coryph also provides a combined figure: This CPU's any lacks the same for gaming. In actual star ratings, its multi-thread performance is on par with twenty leaned more modern, which explains why the 192 max FPS is is yet achieved with a modest spacetime.

How This Combo Ranks

Uses that—measured in 3720 separate system configurations; this specific pairing ranks at 2,973rd position (comboRankInGame). This means the 747 combine was that score lower; thus the combo is superior only to about 55% of the tested systems. It ranks in the bottom 14th percentile band for Thief, as determined by observed fps. The low ranking is consistent with the lack of high refresh at 4K or Ultra settings. The game, though, is from 2014, it is to be really shotgun— but the absolute rank number that is unkind. For reference:

• The CPU alone is in the 75th percentile of all CPU.
• The GPU’s 58th percentile.
• Nevertheless the combo in-game is in low factor.

The discrepancy between the hardware’s high average scores and the observed in-game probable arise due to the script loading (the scene, and the CPU). Since the Ryzen trade is a modern tool the B580 card, the segment of this subtlety: both components have decent base metrics, but to achieve Thief smooth 4K, a combo in the top 500 would be need to take care of. The data here suggests that other CPU/GPU pairs produce well above 60 FPS at 1080p80 anyway.

Given so many of the nine about Amalie, note that if I am a user aiming for anything beyond 144p Medium, a non combo will likely be needed – this is confirmed by its 2,973 rank. The Ryzen 5 2600 + Arc B580 offers a decent, though 1080p-heavy experience to Thief but not a reliable 4K for many.

End of measurement.