Thief

Thief is a 2014 stealth title that mixes tight corridors with broad, light-filled scenes, and this database entry pairs AMD's Ryzen 7 3700X with AMD's Radeon RX 570. The measured FPS results indicate a pairing that ranks at 3551 out of 3720 tested combos for this game, with the metric "dataIsMeasured" confirming that these numbers come from real runs rather than estimates. The CPU is a Ryzen 3000-series part: 8 cores, 16 threads, 3.60 GHz base, 4.40 GHz boost, a 65 W TDP, 32 MB of L3 cache, and a 7 nm Zen 2 (Matisse) die. The GPU, meanwhile, is a Polaris 20 chip from 2017, built on GCN 4.0, with 4 GB of GDDR5 memory, a 256-bit bus, and a 1244 MHz boost clock. The combination yields a frame-rate profile that is far more constrained by the graphics card than the processor, as the CPU's raw processing power is an order of magnitude higher than what Thief's 30-frame console era can nominally demand. This page breaks down the CPU's role, how the combo ranks, resolution scaling, and the exact measured FPS values across the three tested resolutions, concluding with practical settings advice.

CPU Role

The AMD Ryzen 7 3700X offers 8 cores and 16 threads with a base clock of 3.60 GHz and a boost clock of 4.40 GHz. Its TDP is a modest 65 W, sitting on the AMD Socket AM4 with a PCIe Gen 4.0 interface (24 lanes). In benchmark terms, this CPU has an 84th-percentile ranking against all tested CPUs, and its average benchmark score is 34260. When compared with its closest rivals, the Ryzen 7 3700X's nearest neighbors by average score are the AMD Ryzen AI 7 350 (34222, deltaPct +0.1%), the AMD EPYC 4244P (34220, deltaPct +0.1%), the Intel Core i5-13450HX (34333, deltaPct -0.2%), and the Intel Core Ultra 7 165H (34083, deltaPct +0.5%). In native benchmarks, the 3700X achieves a Cinebench R15 multi-core score of 2092 and a single-score of 204; Geekbench gives multi-core 8899 and single-core 1593; Passmark's multithread score is 22458. The cache layout includes 64 KB L1 per core, 512 KB L2 per core, and 32 MB L3. These numbers show a CPU that has 16 threads feeding eight physical cores, a design that scales well with modern multithreaded titles but is also perfectly adequate for a lightweight engine from 2014 such as Thief.

For Thief's gameplay loop, the title is largely single-thread-bound. But this CPU's single-core performance lifts far above the game's engine requirements. The 204-point Cinebench R15 single-core score places it on the expected 1080p output with modest threads. In the measured data for this combo, the CPU never becomes a limiting factor at low resolutions. At 1920x1080 Low settings, the pair achieves 75 FPS average; at 1080p Ultra, still 41 FPS. Such a small delta (34 FPS) in a calculation at the same effective polygons is typical of how a 2019-generation 8-core chip behaves. The CPU offers 16 threads, which corroborates excellent parallel performance in synthetic body of tests per core, so the engine will run with plenty of headroom. Notably, the nearest-rival comparison for the CPU shows a difference of only a half-percentage point in aggregate performance—meaning the 3700X's single-thread performance remains essentially identical to the Intel Core i5-13400HX and the AMD Core Ultra 7 165H. In short: a robust CPU that will never be the limiting factor in Thief, regardless of the resolution chosen.

How This Combo Ranks

This specified combination of Ryzen 7 3700X + Radeon RX 570 appears on the Thief result page with a combo rank of 3551 out of 3720 total combos. That means only about 4.5% of all tested pairings perform worse in this game, while the remaining 95.5% outpace it. This is a profoundly weak rank, tempered by the fact that part of that assessment pulling down the average is the Ryder card. The GPU itself—the Radeon RX 570—earns its own 74th percentile of all tested GPUs with a baseline average score of 28766. In the generic benchmark suite, the RX 570's closest neighbors include the Radeon RX 6800M (avg score 28874, deltaPct -0.4%), the AMD Radeon R9 M280 (28580, +0.7%), the AMD Radeon RX 470 (28996, -0.8%), and the NVIDIA Quadro RTX 8000 (28421, +1.2%). The RX 570 sits in a narrow band, roughly within a percentage point of those four. Yet the combo itself falls to the bottom 4.5% of a game-specific list. The reason boils down to the 4 GB VRAM buffer and the 224 GB/s memory bandwidth, which Thaif's ambient maps and alpha effects chew through. A CPU of this calibre should be paired with a far more modern GPU to produce a strong game result. Instead, the measured data shows exactly how badly this pairing translates to Thief.

The rank 3551/3720 is a red flag for general use. The 3700X is a top-tier 8-core/16-thread CPU by average score (84th percentile), but the combo's bottleneck lies on the Adapter. The 4 GB RAM of the RX 570 and its 150 W TDP set a hard frame ceiling. Thennearest rivals of the GPU in generic benchmarks are all within a 1% margin, so Thief sees a similar spread; yet none of those alternatives would move this combo out of the bottom at 4K. To gain a higher rank, you would essentially need to swap the GPU. The data table clearly shows the impact: at 4K High, this combo averages a mere 17 FPS, and at 1440p Low, only 51 FPS. These FPS values are the actual reason for the rank, and they will always belong to the Radeon unit.

Resolution Scaling

The measured data reveals a dramatic frame rate erosion progression as the resolution climbs from 1920x1080, to 2560x1440, to 3840x2160. Let's keep the Low preset baseline: GPU at 1080p Low = 75 FPS. At 1440p Low, that drops to 51 FPS (a 32% reduction). At 4K Low, you get 30 FPS, another 41% of the 1440p value (and a 60% straight nose dive from 1080p). Medium settings follow the same perspective: at 1080p Medium = 57 FPS, at 1440p Medium = 39 FPS (down 31.6%), and at 4K 2160p Medium = 23 FPS (down of 41% from 1440p). High settings: 1080p High = 43, 1440p High = 30 (down 30.2%), 4K High = 17 (again -43% from 1440p). Ultra numbers have a near identical pattern: 41 > 28 > 17. In every one of these rows the percentage drop from 1440p to 4K is around 40%, and the drop from 1080 to 1440 is about 30-32%. This consistent geometric scaling pattern points directly to the GPU as the limiting component. If the CPU were responsible, we'd expect diminishing frame-rate penalties at 4K when the fragment shading comes to dominate, but the opposite is true. The RX 573 sends the same GPU-generated primitives to draw, and its 32 RGPs and 224 GB/s memory bus simply cannot sustain 8.3 million pixels per second. Thus, we can state decisively that 4K is beyond the graphic card’s capacity regardless of settings, and even 1440p fails to hold anywhere near 60 FPS.

When we break down the FPS scaling across the four resolutions, the relative penalty from High and Ultra settings is nearly identical, except for one notable nuance. At 4K, High and Ultra both stabilize at 17.0, indicating that the card is completely pipeline-bound at that point—raising ultra over high provides literally zero head-room in ever frame. At 1080p, High 43 vs Ultra 41 shows only a pico foot apart. The extra shaders on Ultra are not used because the GPU is at 150 W and hits its power term due to the refresh limit. Meanwhile, the gap from Low to Medium is repeated, but at lower resolutions it is only the GPU blocks, while CPU's unknown. From 1080p Low (75) to 1080p Ultra (41) means a 45% reduction in average FPS, but this is still a playable range. At 4K, the entire settings ladder is compressed into a 13 FPS window (from 30 Low to 17 Ultra), which is a catastrophic situation.

Measured FPS Breakdown

Here is the corrected, explicit measured FPS table from the database for Thief with this hardware:

• 1920x1080 (Full HD):
• Low settings: 75 FPS average (the highest single number recorded for this listing)
• Medium settings: 57 FPS average (with a minimum of 48 and a maximum of 65 FPS)
• High settings: 43 FPS average
• Ultra settings: 41 FPS average

• 2560x1440 (Whirred):
• Low settings: 51 FPS average
• Medium settings: 39 FPS average (min 33, max 45 FPS)
• High settings: 30 FPS average
• Ultra settings: 28 FPS average

• 3840x2160 (4K UHD):
• Low settings: 30 FPS average
• Medium settings: 23 FPS average (min 19, max 26 FPS)
• High settings: 17 FPS average
• Ultra settings: 17 FPS average

Across every resolution, the minimum and maximum values for the Medium preset demonstrate a visible frametime variance: at 1080p Medium, the range is 48–65 FPS, – a around 17. Against the average of 57, that shows the GPU sustains reasonable consistency. At 1440p Medium, the range is 33–45, a range of 12; 4K's is 19–26, range 7. This consistent pattern of high variance illustrates the GPU is frequently spiking closets to its hard cap and then dropping.

5. FAQ

Q: Is this CPU capable of delivering 60 FPS in Thief at any resolution?

A: Yes, if the settings are properly paired. The single measured point that reaches this threshold is at 1920x1080 Low, which reports 75 FPS average. At 1920x1080 Medium the average falls to 57 FPS—just under a 60 FPS target. At no other setting or resolution (as measured in this page) does this combination sustain a 60 FPS average level across the test.

Q: Can this setup handle 2560x1440 with acceptable performance?

A: At 2560x1440, the Low preset provides 51 FPS, Medium gives 39, High 30, and Ultra 28. These yields average fewer than 60 FPS, so a strict 60 FPS locking is not possible. For those tolerant to 30–40 FPS ranges, 1440p Low is the most chargeable option, but the graph shows the GPU reaches a forgotten limitation in this resolution.

Q: Does the extra CPU compute help the GPU in Thief?

A: the data shows no. The 3700X's multi-thread strength (16 threads, 8899 Geekbench multi) far exceeds Thief's load, and the frame performance does not change as the resolution adds—it dives. If the CPU were the limiting factor, scaling by 40% per resolution wouldn't happen; the bottleneck is distinctly the Radeon RX 570’s 224 GB/s bandwidth and 32Rop.

Q: How does the RX 570 compare to its closest rival GPU in the synthetic list?

A: On the raw GPU-pool benchmarks, the RX 570 average score is 28766. The AMD Radeon RX 6800M is merely -0.4% from this (28874), the AMD Radeon R9 M295X is +0.7% (28580), the RX 470 is -0.8% (28996), and the NVIDIA Quadro RTX 8000 is +1.2% (28421). Thus, even if you substituted one of those GPUs, Thief's FPS would remain essentially identical, given their lack of increase core count.

**Q: Are these frame rates measured or