Thief

The Intel Core Ultra 5 225 paired with the NVIDIA GeForce GTX 1650 produces a constrained experience in Thief, with the benchmark data clearly indicating that the GPU is the primary bottleneck across all tested resolutions. The measured frame rates show a steep decline as resolution increases, a classic symptom of a graphics card operating beyond its comfortable performance envelope. This analysis walks through the measured FPS data, settings behavior, and the individual roles of the CPU and GPU in shaping the final results.

Resolution Scaling

The measured FPS data reveals a pronounced and consistent drop in performance as the resolution scales from 1080p to 4K, which strongly indicates that the NVIDIA GeForce GTX 1650 is the limiting component. At 1080p with Low settings, the combo achieves an average of 74 FPS, which is a playable baseline. However, moving to 1440p at the same Low settings reduces the average to 51 FPS, a 31% decrease. At 4K with Low settings, the average falls further to 30 FPS, representing a 59% drop from the 1080p Low figure. This scaling pattern is far steeper than what would be expected if the CPU were the primary constraint, since the Intel Core Ultra 5 225 has ample processing headroom for a 2014 title.

The pattern holds across all settings presets. For High settings, the average FPS goes from 43 at 1080p, to 29 at 1440p, and finally to 17 at 4K. Medium settings show a similar trajectory: 56 FPS at 1080p, 39 FPS at 1440p, and 23 FPS at 4K. The Ultra preset follows with 41 FPS at 1080p, 28 FPS at 1440p, and 17 FPS at 4K. The fact that both High and Ultra settings land at exactly 17 FPS at 4K suggests that the GPU is hitting a hard computational ceiling, unable to differentiate between the two presets at that resolution because it is fully saturated. The data shows that the GTX 1650’s rendering capabilities are exhausted well before the CPU’s limits are approached.

The implication is clear: resolution scaling in this configuration is governed almost entirely by the GPU’s pixel throughput and memory bandwidth. The GTX 1650’s 128.1 GB/s memory bandwidth and 2.984 TFLOPS FP32 performance are simply insufficient for 4K rendering in Thief, regardless of the quality preset. For users targeting higher resolutions, the CPU is effectively idle, waiting on the graphics card to finish each frame. The benchmark results indicate that this combo is only viable for smooth gameplay at 1080p, with 1440p being a marginal stretch and 4K being impractical for anything other than static scenes or screenshots.

FAQ

*Q: What is the best resolution for this CPU and GPU combo in Thief?*

A: The data shows that 1920x1080 is the only resolution where the combo achieves consistently playable frame rates. At 1080p Low, the average is 74 FPS, and even at 1080p Ultra, the average is 41 FPS. At 1440p, only the Low preset (51 FPS average) crosses the 50 FPS threshold, while 4K fails to reach 30 FPS on any preset.

Q: How does the GTX 1650’s performance compare to its nearest rivals based on the benchmark data?

A: The GTX 1650 has an average benchmark score of 7472 and sits in the 40th percentile of all GPUs. Its nearest rival, the Intel Arc A310, scores 7550, which is 1% higher. The AMD Radeon HD 8850M scores 7447, which is 0.3% lower, and the Intel UHD Graphics 750 scores 7441, 0.4% lower. The GTX 1650 is essentially in a dead heat with these parts, though none of them are modern high-performance cards.

Q: Is the CPU or GPU more responsible for the low frame rates at 4K?

A: The GPU is entirely responsible. The Core Ultra 5 225 has a high single-thread score of 3655 in Cinebench R23 and a multi-core score of 25891, indicating it is far more powerful than what Thief requires. The GPU, however, has only 4 GB of GDDR5 memory and a 128-bit bus, which is insufficient for 4K workloads. The FPS drop from 1080p to 4K is steep (e.g., 74 to 30 on Low), confirming the GPU bottleneck.

Q: What is the difference in average FPS between Low and Ultra settings at 1080p?

A: At 1920x1080, the Low preset achieves an average of 74 FPS, while the Ultra preset achieves an average of 41 FPS. This represents a 45% reduction in performance when moving from the lowest to the highest quality preset. The Medium preset lands in between at 56 FPS, and High is close to Ultra at 43 FPS.

Q: Does the GTX 1650’s memory capacity affect the game’s performance?

A: Yes, the 4 GB VRAM capacity is a likely constraint at higher resolutions and settings. While the data does not show explicit VRAM usage, the fact that High and Ultra both produce 17 FPS at 4K suggests that the card is running out of memory or bandwidth, causing it to stall. The 128.1 GB/s bandwidth is also low by modern standards, further limiting texture throughput.

*Q: How does this combo rank among all tested configurations for Thief?*

A: The combo ranks 2766 out of 2851 tested combinations, placing it in the bottom 3% of all systems. This ranking reflects the GTX 1650’s low absolute performance, as the CPU is more than capable, but the overall system is dragged down by the graphics card's limitations.

GPU Role

The NVIDIA GeForce GTX 1650 is the decisive component in this pairing, and its specifications explain the measured results. The card is based on the TU117 chip, built on a 12 nm process, and features 896 shading units, 56 TMUs, and 32 ROPs. It operates at a base clock of 1485 MHz and a boost clock of 1665 MHz. In synthetic benchmarks, the GTX 1650 scores 7880 in Passmark G3D and 3048 in Passmark GPU Compute, which places it in the 40th percentile of all GPUs. These are modest figures, and the card’s performance in Thief reflects this.

The memory subsystem is a significant weakness. The GTX 1650 has 4 GB of GDDR5 memory on a 128-bit bus, yielding a bandwidth of 128.1 GB/s. In a game like Thief, which was released in 2014, this amount of memory is technically sufficient for 1080p, but the bandwidth becomes a limiting factor at higher resolutions. The data supports this: at 1080p Medium, the average FPS is 56, but at 1440p Medium, it drops to 39, and at 4K Medium, it falls to 23. The min FPS values at Medium settings also show instability, with 48 FPS at 1080p, 33 FPS at 1440p, and 19 FPS at 4K, indicating frame time spikes when the memory bus is saturated.

The GTX 1650’s clock speeds are not the issue; the architecture’s raw throughput is. With an FP32 performance of 2.984 TFLOPS and a pixel rate of 53.28 GPixel/s, the card is designed for entry-level 1080p gaming, not for high-resolution rendering. The data confirms this: the card cannot maintain 30 FPS at 4K on any preset, and even at 1440p, only the Low preset (51 FPS) provides a fluid experience. The GPU’s role in this system is to cap the performance ceiling, and the benchmark results show that ceiling is firmly at 1080p. For users with this card, the settings recommendations are clear, but the hardware itself is the limiting factor.

Settings Recommendations

Based on the measured FPS data, the optimal settings for this combo depend on the user’s resolution and frame rate target. At 1920x1080, the Low preset delivers an average of 74 FPS, which is the smoothest experience available. However, the Medium preset, at 56 FPS average, offers a substantial visual improvement over Low while still maintaining a frame rate that is above the 50 FPS threshold, making it the best balance of quality and performance at 1080p. The High preset (43 FPS) and Ultra preset (41 FPS) are nearly identical in performance, but both dip below the 45 FPS mark, which may feel less fluid in a stealth game that rewards smooth camera movement. Therefore, the recommended setting at 1080p is Medium, as it provides a significant visual upgrade over Low without the steep performance cost of High or Ultra.

At 2560x1440, the situation is more restrictive. The Low preset achieves an average of 51 FPS, which is playable, but the Medium preset drops to 39 FPS, and High and Ultra fall to 29 and 28 FPS, respectively. Given the data, the only sensible choice at 1440p is the Low preset, as it is the only one that stays above 50 FPS. The Medium preset’s 39 FPS average is borderline, but the min FPS of 33 suggests noticeable stutter in demanding scenes. For 3840x2160, the data shows that no preset is viable: the Low preset averages 30 FPS, which is technically playable but suboptimal, and all other presets fall to 23 FPS or below. The recommendation for 4K is to avoid it entirely or to accept the Low preset with the understanding that frame rates will be inconsistent.

The measured rows indicate that the jump from Low to Medium at 1080p costs 18 FPS (74 to 56), which is a 24% reduction. The jump from Medium to High costs 13 FPS (56 to 43), and from High to Ultra costs only 2 FPS (43 to 41). This suggests that the Ultra preset offers negligible visual benefits over High for the performance cost, and that the Medium preset is the sweet spot for this GPU. Users prioritizing frame rate above all else should select Low, but for a balanced experience, Medium is the clear winner in the data.

Measured FPS Breakdown

The measured FPS data provides a comprehensive view of the combo’s performance across three resolutions and four settings presets. At 1920x1080, the full range of results is as follows: Low averages 74 FPS, Medium averages 56 FPS with a min of 48 and a max of 65, High averages 43 FPS, and Ultra averages 41 FPS. The 1080p results show a clear hierarchy, with Low being the only preset above 60 FPS and Medium being the only one above 50 FPS. The min FPS for Medium (48) is close to the average for High (43), indicating that the Medium preset can occasionally dip to levels comparable to lower quality settings.

At 2560x1440, the averages are: Low at 51 FPS, Medium at 39 FPS (min 33, max 45), High at 29 FPS, and Ultra at 28 FPS. The 1440p data shows a steeper decline than 1080p, with the gap between Low and Medium being 12 FPS (51 to 39), and the gap between High and Ultra being negligible at 1 FPS. The min FPS for Medium at 1440p (33) is below the average for High (29), suggesting that the Medium preset can experience significant frame drops during intense scenes.

At 3840x2160, the results are uniformly poor: Low averages 30 FPS, Medium averages 23 FPS (min 19, max 26), High averages 17 FPS, and Ultra averages 17 FPS. The 4K data reveals that the GPU is completely overwhelmed, with High and Ultra producing identical averages. The min FPS for Medium at 4K (19) is barely above the average for High and Ultra, confirming that the card is running at its absolute limit. Across all resolutions, the Low preset is the most consistent, while the higher presets show diminishing returns and increasing instability.

CPU Role

The Intel Core Ultra 5 225 is a powerful processor that is not the bottleneck in this configuration. It features 10 cores and 10 threads, with a base clock of 3.30 GHz and a boost clock of 4.90 GHz. The CPU is built on a 3 nm process (TSMC) and has a 20 MB shared L3 cache, along with 3 MB of L2 cache per core. In synthetic benchmarks, the Core Ultra 5 225 scores 25891 in Cinebench R23 multi-core and 3655 in single-core, placing it in the 85th percentile of all CPUs. The Passmark multi-thread score is 30459, and the single-thread score is 4412. These numbers indicate a CPU that is far more capable than what Thief needs, even at high frame rates.

The CPU’s nearest rivals in the benchmark database include the AMD Ryzen 5 5600F, which scores 36945 on average (a 0% delta), and the AMD Ryzen 5 5500X3D, which scores 37018 (a -0.2% delta). The Intel Core i9-12900T scores 37112, which is -0.5% relative to the Core Ultra 5 225’s average benchmark score of 36938. These comparisons show that the Core Ultra 5 225 is competitive with mid-range and older high-end CPUs, but its performance in Thief is irrelevant because the GPU limits the frame rate. The benchmark results indicate that the CPU is never the limiting factor, as the FPS drops are entirely attributable to the GPU’s inability to keep up with increasing resolution.

The CPU’s memory support includes DDR5 with a dual-channel bus and a bandwidth of 102.4 GB/s, which is more than sufficient for gaming. The 20 MB L3 cache and 10 cores provide ample headroom for the game’s AI and physics calculations, which are light by modern standards. The data shows that the CPU’s role in this system is to sit idle, waiting for the GPU to render each frame. Even at 1080p Low, where the GPU is least stressed, the CPU does not appear to be the constraint, as the FPS is 74, which is well below what the CPU could theoretically push. In essence, the Core Ultra 5 225 is a wasted resource in this pairing, as the GTX 1650 holds the entire system back.

How This Combo Ranks

The combination of the Intel Core Ultra 5 225 and the NVIDIA GeForce GTX 1650 ranks 2766 out of 2851 tested combos in Thief, placing it in the bottom 3% of all configurations. This low ranking is almost entirely due to the GPU, as the CPU is in the 85th percentile of all processors. The benchmark results show a massive disconnect between the two components: the CPU is capable of far higher frame rates, but the GPU caps the system’s performance at a level consistent with entry-level hardware from several years ago.

The ranking reflects the measured FPS data, which shows that the combo cannot achieve 60 FPS at 1080p on any preset other than Low, and cannot achieve 30 FPS at 4K on any preset. In the context of the database, this places the combo in the company of other low-end GPU pairings, rather than with systems that can fully utilize the Core Ultra 5 225’s capabilities. The GPU’s percentile rank of 40 and its average benchmark score of 7472 are consistent with the observed in-game performance, confirming that the ranking is not an anomaly but a reflection of the hardware’s true capabilities. For users considering this combo, the data is unambiguous: the GTX 1650 is the sole determinant of performance, and the CPU’s high-end specs do little to boost the overall experience in Thief.