Thief

How This Combo Ranks

The AMD Ryzen 5 4600G paired with the NVIDIA GeForce RTX 5090 lands at rank 672 out of 1,642 tested combos in Thief. That places this configuration in the upper half of all tested systems, though not at the very top. The RTX 5090 is a massive GPU — it sits in the 94th percentile of all GPUs tested — but the Ryzen 5 4600G is a more modest processor, ranking in the 75th percentile of all CPUs. The gap between those two percentile positions is the story here.

This is not a balanced pairing. The GPU is capable of far more than this CPU can feed it, especially at lower resolutions. The data shows the combo ranking reflects that bottleneck. When you look at the measured FPS, the numbers are playable across the board, but they are not the kind of figures you would expect from an RTX 5090 in a 2014 stealth title. The GPU's average benchmark score of 84,306 puts it just ahead of the NVIDIA GeForce RTX 5090 D (84,241, delta 0.1%) and the RTX 5050 Mobile (84,171, delta 0.2%), and it is 1.3% ahead of the AMD Radeon PRO W6600. The CPU, meanwhile, scores 17,489 on average, nearly identical to the AMD Ryzen 3 PRO 5355GE (17,482, delta 0%) and within 1% of several Intel Core i3 parts. That CPU is the limiting factor in this combination.

FAQ

Q: Is the RTX 5090 being held back by the Ryzen 5 4600G in Thief?

A: Yes. At 1080p Low, the combo produces 128 FPS. At 4K Ultra, it produces 78 FPS. The GPU is in the 94th percentile of all GPUs, while the CPU is in the 75th percentile. The FPS ceiling at lower resolutions is clearly CPU-bound, as the GPU could deliver far higher frame rates if the processor could keep up.

Q: What is the best resolution for this combo in Thief?

A: 1440p offers the best balance. At 1440p High, the combo delivers 96.1 FPS, which is only about 10 FPS less than 1080p High (106.3 FPS). The gap between 1080p and 1440p is much smaller than the gap between 1440p and 4K, suggesting the CPU bottleneck fades as resolution increases.

Q: How does the CPU compare to its nearest rivals?

A: The Ryzen 5 4600G scores 17,489 on average, which is essentially tied with the AMD Ryzen 3 PRO 5355GE (17,482, delta 0%). It is 0.1% ahead of the Intel Core i5-12450H (17,475) and 0.8% behind the Intel Core i3-14100T (17,636). These are all very close scores, indicating the 4600G is a mid-pack performer in this range.

Q: Does the 4600G's integrated graphics matter here?

A: The CPU includes Radeon Vega 7 integrated graphics, but the system uses a discrete RTX 5090. The integrated graphics are irrelevant for gaming in this configuration. The CPU's 6 cores and 12 threads are what matter, along with its 3.70 GHz base clock and 4.20 GHz boost clock.

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

A: At 1080p, Low settings produce 128 FPS, while Ultra produces 105.7 FPS. That is a 22.3 FPS difference, or roughly 17% lower performance at Ultra. The gap is smaller at higher resolutions, with 4K Low at 98.5 FPS and 4K Ultra at 78 FPS.

Q: Is this combo better suited to 4K gaming?

A: The data suggests yes, relatively speaking. At 4K Ultra, the combo delivers 78 FPS, which is 74% of the 1080p Low result. That is a much smaller proportional drop than you would expect from a CPU-bound system. The RTX 5090's 32 GB of GDDR7 memory and 1.79 TB/s bandwidth are well-utilized at 4K, where the CPU bottleneck is less pronounced.

Resolution Scaling

The FPS scaling from 1080p to 4K reveals a classic CPU bottleneck. At High settings, the combo delivers 106.3 FPS at 1080p, 96.1 FPS at 1440p, and 79.5 FPS at 4K. That is a drop of only 9.6% going from 1080p to 1440p, but a 25.2% drop going from 1440p to 4K. The modest drop at 1440p indicates the CPU is still the primary constraint at 1080p, and even at 1440p it is not fully relieved.

Look at the Low settings row: 128 FPS at 1080p, 116.3 FPS at 1440p, and 98.5 FPS at 4K. The 1080p to 1440p gap is 9.1%, which is nearly identical to the High settings gap. This tells you the CPU is the limiting factor regardless of graphics load at these resolutions. The GPU has so much headroom that lowering settings does not meaningfully increase FPS at 1080p.

At 4K, the scaling behavior changes. The 4K Low result of 98.5 FPS is 77% of the 1080p Low result, which is a much larger drop. This suggests that at 4K, the GPU starts to become more relevant, though the CPU still holds things back. The RTX 5090's 104.8 TFLOPS of FP32 performance and 423.6 GPixel/s pixel rate are far beyond what this game needs at 1080p, but at 4K with Ultra settings, the GPU is finally doing meaningful work.

The pattern is consistent across all settings: the delta between 1080p and 1440p is always in the 9-10% range, while the delta between 1440p and 4K is always 17-18%. This two-tier scaling is a textbook sign of a CPU-limited system. The CPU cannot push more frames regardless of how much GPU power is available.

Measured FPS Breakdown

At 1920x1080, the combo produces its highest frame rates. Low settings yield 128 FPS, Medium delivers 114.7 FPS, High comes in at 106.3 FPS, and Ultra drops to 105.7 FPS. The spread from Low to Ultra is 22.3 FPS, but the gap between High and Ultra is only 0.6 FPS, which is negligible. This suggests that High and Ultra are nearly identical in performance at 1080p, likely because the CPU is the limiting factor and the GPU can handle either preset without breaking a sweat.

Moving to 2560x1440, the numbers drop but remain smooth. Low settings produce 116.3 FPS, Medium delivers 102.2 FPS, High comes in at 96.1 FPS, and Ultra yields 93.7 FPS. The gap between Low and Ultra is 22.6 FPS, almost the same as at 1080p. The 1440p results are consistently about 9-10% lower than the 1080p results, which is a small penalty for a significant resolution increase.

At 3840x2160, the frame rates are still playable but clearly lower. Low settings deliver 98.5 FPS, Medium produces 88.2 FPS, High comes in at 79.5 FPS, and Ultra yields 78 FPS. The gap between Low and Ultra is 20.5 FPS, which is slightly smaller than at lower resolutions. Interestingly, the difference between High and Ultra at 4K is only 1.5 FPS, again showing that the CPU is the wall.

The data shows a consistent pattern: the difference between High and Ultra is always tiny (0.6 to 1.5 FPS), while the difference between Low and Medium is always substantial (13.3 to 10.3 FPS). This tells you that the game's Ultra preset does not add much GPU load over High in this configuration, or the CPU simply cannot tell the difference.

Settings Recommendations

For the best experience with this combo, Medium settings at 1440p is the sweet spot. At 1440p Medium, the combo delivers 102.2 FPS, which is above the 100 FPS threshold that most players consider smooth. Dropping to Low at 1440p only gains 14.1 FPS (116.3 FPS), but the visual quality loss is not worth that small gain.

If you prefer 1080p, High settings at 106.3 FPS is the recommended preset. The jump from High to Ultra at 1080p gains only 0.6 FPS, so there is no reason to run Ultra. Medium at 1080p gives 114.7 FPS, but again, the visual improvement from Medium to High is worth the 8.4 FPS cost.

For 4K gaming, High settings at 79.5 FPS is the best choice. Ultra at 4K produces 78 FPS, a mere 1.5 FPS less, so you could argue either preset works. However, Medium at 4K delivers 88.2 FPS, which is 11% faster than High. If you want to stay above 85 FPS at 4K, Medium is the safer bet. Low at 4K gives 98.5 FPS but sacrifices too much visual fidelity for a 2014 game that still looks decent at Medium.

The data suggests avoiding Ultra across all resolutions. The performance difference between High and Ultra is consistently under 2 FPS, which means the Ultra preset is either poorly optimized or the CPU bottleneck masks any GPU gains. Stick with High for the best visual quality per frame, or drop to Medium if you want a comfortable FPS buffer.

CPU Role

The AMD Ryzen 5 4600G is a 6-core, 12-thread processor based on the Zen 2 architecture, codenamed Renoir. It has a base clock of 3.70 GHz and a boost clock of 4.20 GHz, with 8 MB of shared L3 cache. These are modest specs by 2025 standards, and the benchmark data reflects that. The CPU scores 13,593 in Cinebench R23 multicore and 1,919 in single-core, which places it in the 75th percentile of all CPUs.

In Thief, the CPU's role is decisive. The game is a 2014 title, and the data shows it does not scale well with the RTX 5090's massive compute resources. The CPU's passmark single-thread score of 2,653 and multithread score of 15,992 are adequate for the game's engine, but they cap the frame rate. The near-identical FPS between High and Ultra settings at every resolution is direct evidence that the CPU, not the GPU, is setting the ceiling.

The 4600G's memory bandwidth of 51.2 GB/s over dual-channel DDR4 is another potential constraint. The RTX 5090 has 1.79 TB/s of memory bandwidth on its side, but the CPU must feed the game's draw calls through that 51.2 GB/s pipe. The 3DMark scores show the CPU scaling from 1,430 at 2 threads to 4,889 at max threads, but Thief likely does not use more than a few threads efficiently, given its age.

Compared to its nearest rivals, the 4600G is essentially average. It matches the AMD Ryzen 3 PRO 5355GE (delta 0%), edges out the Intel Core i5-12450H by 0.1%, and trails the Intel Core i3-14100T and i3-13100F by less than 1%. This is a CPU that performs exactly in line with its price class, but it is simply out of its depth next to an RTX 5090. If you were to pair this GPU with a stronger CPU, the FPS numbers in Thief would likely be significantly higher, especially at 1080p and 1440p.