Thief

The Intel Core i5-12400 paired with the NVIDIA GeForce RTX 3050 4 GB represents a mid-range CPU from 2022 embraced by a low-end GPU from the same generation, combined here in the 2014 stealth title Thief. The measured data, drawn from the benchmark database, covers four quality presets across 1080p, 144p, and 2160p. While the CPU comfortably powers headroom, the RTX 3050 4 GB consistently becomes the constraining factor as resolution increases, pulling average frame rates from a solid 110 at 1080p Low to just 24 at 4K Ultra. Analysis of the recorded numbers reveals a GPU-bound profile at all resolutions above 1080p, confirming that the 4 GB frame buffer and modest shader count do the real work.

CPU Role — cores, clocks, and how they relate to this game's results

The Intel Core i5-12400 is a six-core, twelve-thread desktop processor built on Alder Lake-S, with a base clock of 2.50 GHz and boost of 4.40 GHz. It features an 18 MB shared L3 cache, 1.25 MB per-core L2, and 80 KB per-core L1, making it well suited for a wide range of games. In the synthetic benchmarks, this CPU achieves a Cinebench R23 multi-core score of 15989 and a single-core score of 2257, placing it in the 72nd percentile of all CPUs. Its Geekbench 5 single-core result of 1848 and multi-core of 8564 reinforce its competence in both lightly threaded and heavier workloads.

Thief, released in February 2014, is an engine designed around a pre-2010 generation of CPUs. It does not heavily scale with thread counts, but it does benefit from high single-core performance. The i5-12400's single-thread Cinebench R23 score of 947 is among the upper tiers of its class, and the measured results from the game confirm that the CPU is not a limiter at moderate settings. For example, with the 1080p Low preset, the average 1080p frame rate reaches 110 FPS, which is likely past the CPU's share but still indicates that the 3050 4GB can take over at that point.

In the benchmark database, the i5-12400's average benchmark score is 18683, aligning it within 0.1% of the AMD EPYC 7643 (18697, delta -0.1%) and within 0.7% of the AMD Ryzen AI 5 330 (18811). This puts the i5-12400 essentially on par with EPYC workstation and mid-level mobile chips in synthetic terms, yet the game result is far more dependent on the attached GPU. The CPU's 72nd percentile and its benchmark strength suggest it could easily drive higher FPUs if paired with a more potent graphics card; however, in this pairing, the 12400 is never the bottleneck according to the resolution scaling.

GPU Role — VRAM, clocks, and how they relate to this game's results

The NVIDIA GeForce RTX 3050 4 GB uses the GA107 chip on an 8 nm Samsung process, with 2048 shading units, 64 tensor cores, and 16 ray tracing cores. Its base clock is 1545 MHz and boost 1740 MHz, while its 4 GB of GDDR6 runs at 1500 MHz effective (12 Gbps) across a 128-bit interface, delivering 192.0 GB/s bandwidth. The card's raw computational power, 7.127 TFLOPS single-precision, supports a 50th-percentile position among all GPUs, meaning half the cards are faster.

For Thief, the GPU is the clear bottleneck. As the resolution and preset rise, the FPS plagued. At 1080p Low, the card hits 110 FPS, but that is 55% of the same preset at 1440p (72 FPS) and 40% at 4K (22 FPS). The 4 GB VRAM, though not heavily risked in a 2014 title, is still a factor in the memory bandwidth constraints. The bandwidth of 192.0 GB/s is sufficient for moderate resolutions, but when scaling to 4K, the card's raw throughput (fill rate 55.68 GPixel/s) is exhausted.

Crucially, the RTX 1545 base clock and 1740 boost are unchanged by the resolution; the only difference is the constant workload per frame. The result is that at 1440p, the card achieves between 41 FPS (Ultra) and 72 FPS (Low), never reaching a 60+ average except at the Low preset. At 4K, the highest setting, Ultra, produces only 24 FPS, and even the Low preset yields 44 FPS, which is below the standard 60 FPS target. The GPU is the sole reason the combo ranks at the low end of the database (rank 8272 out of 3720), which we will discuss later.

Resolution Scaling

The measured data, the FPS trends across three resolutions clearly identify the GPU as the bottleneck. Taking the Low preset: 1080p averages 110 FPS, 1440p averages 75, and 2160p drops to 44. The ratio 1440p/1080p is 75/110 = 68.2%, and 2160p/1080p is 44/110 = 40%. This scaling factor is extremely close to the pixel-count ratio: from 1080p to 1440p (pixels scaled by (36/4) = 1.7778 times), the FPS goes down to 68% (inverse ~1/times), and to 4K (pixels 1290/1080 = 2 p, actually 38402160 = 8.5 Megapixels vs 10801920 = 2.07, so ~4.09x), FPS drops to 40%. That 4x pixel count yields a 2.5x frame-time increase, indicating that the card is ~40% scaled. At 1080p Low 110 FPS, the 4GB card is not yet fully loaded, but the gap to 4K is still ~70% reduction. For the Ultra preset: 1080p 50, 1440p 43, 4K 24, so same ratio of ~34% from 1080p to 4K. In all cases, the CPU overhead is almost identical because the CPU's speed is constant across resolutions; the changing FPS confirms that the RTX 3050 4GB is operating in power limit or bandwidth limit, not CPU limit.

We can also infer the GPU load from the medium preset: 1080p Medium 83 FPS, 1440p Medium 57, 4K Medium 34. The drop from 83 to 57 to 34 are a nominal 68% and 41% again. This parabola strongly suggests a classic fill-rate scaling – the GPU is working harder per frame as resolution grows. The minimum and maximum frames measure for Medium also show narrowing at higher resolutions; at 1080p the min/max range is 71-106, at 1440p it's 49-58, and at 4K 29-39. The 4K vari imply is narrower, but all below playability, indicating that the card is consistently at 100% utilization at 4K.

Measured FPS Breakdown

The following table list the exact average FPS and, where available, the minimum and maximum frames collected by the benchmark:

| Resolution | Preset | Average FPS | Min FPS | Max FPS |

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

| 1920x1080 | Low | 110 | - | - |

| 1920x1080 | Medium | 83 | 71 | 96 |

| 1920x1080 | High | 64 | - | - |

| 1920x1080 | Ultra | 60 | - | - |

| 2560x1440 | Low | 75 | - | - |

| 2560x1440 | Medium | 57 | 49 | 66 |

| 2560x1440 | High | 44 | - | - |

| 2560x1440 | Ultra | 41 | - | - |

| 3840x2160 | Low | 44 | - | - |

| 3840x2160 | Medium | 34 | 29 | 39 |

| 3840x2160 | High | 26 | - | - |

| 3840x2160 | Ultra | 24 | - | - |

Globally, the only combination that reliably holds 60+ average is the 1080p Low (110) and 1080p High (64) – and 1080p Ultra exactly 60. At 1440p, the Medium preset yields 57 FPS, which is within the normal range for 60 but only if one tolerates an occasional dip to 49. At 4K, no preset reaches 45 FPS; the best is Low with 44 FPS. The min/max at Medium preset indicate that 1080p drops only to 71 in the worst case, while 1440p falls to 49, causing potential stutter on 60Hz monitors.

How The Combo Ranks

In the full dataset of tested CPU/GPU combinations in Thief, the i5-12400 + RTX 3050 4GB combo is ranked 3272 out of 3720 combos. This position places it in the 12th percentile from the bottom, meaning only 12% of all configurations deliver worse performance. Given that the i5-12400 CPU alone sits in the 27th percentile, the forming gap is entirely caused by the GPU—it's the same RTX 3050 4GB that has a 50th-percentile standing on its own, but in this game context (which is moderate loads), it cannot push beyond these low averages.

The per-combination rank is a clear signal that if a user is playing Thief, this pairing is far from an ideal match. The CPU could handle significantly stronger GPUs without becoming a bottleneck, but the 3050 4GB caps. The large number of combos below (448) is a small margin, yet the vast majority of configurations produce better frame rates. The notable thing is that the RTX 3050 4GB is an end-of-life product with a 50th percentile worldwide; its performance in Thief is limited not by raw shader count, but by the modest 4 GB VRAM and 128-bit bus, which hinders high-resolution.

FAQ

Q: Does the RTX 3050 4GB run Thief at 60 FPS at 1080p?

A: At 1080p, the High preset averages 64 FPS and the Ultra preset 60 FPS, so 60+ FPS is achievable at both High and Medium (83 FPS) and low (110 FPS). The Medium preset dips to 71 FPS as a minimum, so it is still stable.

Q: What is the best FPS that 1440p can deliver?

A: At 1440p, the highest average frame rate is on the Low preset with 75 FPS and the Medium preset gives 57 FPS. Playing at Ultra yields only 41 FPS, so 1440p is only suitable for casual play on the low/medium presets.

Q: What is the performance difference between 1080p and 4K when using the Low preset?

A: The Low preset average FPS drops from 108 (1080p) to 44 (2160p), a decrease of 66 FPS (or exactly 40% of the original). This is a 2.5x reduction in frame rate.

Q: What is the maximum FPS measured in this test?

A: The maximum recorded average is 110 FPS, seen at 1080p with the Low preset. No 4K preset exceeds 44 FPS.

Q: Does 4GB of VRAM cause limitations?

A: While the game is from 2014 and likely does not use more than 3GB, the GPU's 4GB VRAM, together with a 128-bit memory bus, correlates with the steep resolution scaling. However, the substantial drop in FPS from 1080p to 1440p (up to ~33%) suggests that bandwidth and fill rate, not capacity, are the primary limiting factors.

Q: How does this combo compare to other test results in Thief?

A: The rank 3272 out of 3720 totals place it in the bottom 12% of all combos, meaning it outperforms 12% of the systems. The CPU's own average benchmark score is 18683, which is only slightly above rivals like the AMD EPYC 7643 (18697) but the GPU is the weak link.

Settings Recommendations

Form 60 FPS target, the only presets that deliver a solid average are found at 1080p. The Medium preset provides 83 FPS with a minimum of 71, which is a comfort zone for a 60Hz display. The High preset gives 64 FPS, and Ultra exactly 60 – but with no recorded minimum, it is uncertain whether stutter could occur. For 1080p, the Medium preset seems a good balance of visuals and smoothness, as the 80−70 window keeps frame times low.

At 1400p, the only way to stay above 60 FPS is to use the Low preset (75 FPS). The Medium preset – 57 FPS – is just below 60, and although its minimum 49 is not ideal, it is surprisingly playable with good V‑sync. If the user owns a 1440p monitor, the Low preset is the only way to hit 75 FPS. Ultra and High are not playable (below 45 FPS).

For 4K, none of the presets produces a stable 60 FPS. The Low preset at 44 FPS is the best, but it represents a compromised visual quality in exchange for playability. The 4K present "Medium" and above have frames that drop to sub‑30 range, making them only for immersion but not interaction. Given the massive drop from 1080p to 4K, the recommended usage of this GPU is exclusively for 1080p gameplay, using the Medium or High presets for a reliable 60+ frame rate.

Because the RTX 3050 4GB is end-of-life product, owners should consider utilising 1080p Medium to avoid the frame rate dips that appear on the Ultra and 1440p presets.