Thief

The combination of AMD's 12-core Ryzen 9 7900X and NVIDIA's RTX 3070 Ti produces a competent 1440p and 1080p experience in Thief, with playable 4K at medium and high settings. The combo's rank in the database is 1102 out of 3720 tested combinations, placing it in the top third of all pairings. Across 12 measured scenarios, the GPU—a GA104 Ampere part with 24 GB of GDDR6X running at 1188 MHz (19 Gbps effective) and 8 GB of VRAM—delivers 208 fps at 1080p Low, dropping to 49 fps at 4K Ultra. The CPU's 24 parallel threads and 64 MB of shared L3 cache support the engine's draw-call processing, though at lower resolutions the GPU becomes the limiting factor.

How This Combo Ranks

In the overall benchmark database, this CPU+GPU pair ranks 1100 of 3,720 tested combos for Thief. That places it in the 71st percentile of all 3,720 combinations tested, meaning roughly 29% of pairings are faster. The ranking is a function of both the CPU's high single-core throughput and the GPU's mid-high tier placement. According to the data, the CPU sits in the 91st percentile among all CPUs, while the GPU is in the 75th percentile among all GPUs. The impedance between the two parts is not severe: at 1080p and High, the average FPS is 127, which is 5.4% higher than the 120 FPS threshold that many consider a smooth gold standard, and even at 4K Ultra the average stays at 49 FPS, comfortably above 60% of the tested combos' ability to hold 60 FPS on the same quality level.

GPU Role

The RTX 3080 Ti's memory subsystem is the defining spec for a game like Thief, which places stress on texture throughput in its detailed environments. The GPU is equipped with 8 GB of GDDR6X on a 256-bit bus, resulting in a bandwidth of 608.3 GB/s. Its base clock is 1575 MHz, boosting to 1770 MHz. In Thief, the GPU scalability across settings is clear: at 1080p, raising the preset from Low to Ultra drops the average FPS from 208 to 121, a difference of 87 frames. At 4K, the same setting jump reduces average FPS from 89 to 49, a gap of 40 frames. The medium preset at 4K yields 67 FPS, with a stored minimum of 57 and maximum of 77, indicating that the frame time is relatively stable. The 4K performance is likely driven by the GPU's shading capability; the card provides an FP32 throughput of 21.75 TFLOPS. Compare this to the nearest GPU rival from the database, the AMD RX 6800, which posts an average benchmark score 0.5% lower but has a 0.5% delta against the RTX 3080 Ti's 0.5% delta in the rival list, and the data shows the RTX 3080 Ti holds a decisive edge in Thief's measured results.

The GPU's rasterization units (96 ROPs) and TMU count (192) feed a pixel rate of 169.9 GPixel/s and texture rate of 339.8 GTexel/s. In a game like Thief, which is not computationally heavy per pixel, this ensures that even at 4K Ultra the GPU can produce 49 frames per second, which is only 23% slower than the 4K Medium average. When the settings are at Low, the 4K runs hit 89 fps, which is above a 60 Hz refresh. That result, in fact, exceeds the 1080p High average of 127 fps, indicating the GPU is still bounds waiting for the CPU in lower-resolution scenarios.

FAQ

*Q: What is the combined ranking for this CPU/GPU configuration in Thief?*

A: The Ryzen 9 7900X + RTX 3080 Ti combo ranks 1100 out of the 3,720 tested combinations in Thief, which makes it a top-third performer.

Q: How much GPU memory does the RTX 3080 Ti have and what is its effect?

A: It features 8 GB of GDDR6X. At 4K Ultra the average FPS is 49, and at 4K Low it is 89; the memory bandwidth of 608.3 GB/s supports the game's texture needs, though 8 GB can become a factor in future titles that require more.

Q: What is the base and boost clock of the GPU?

A: The RTX 3080 Ti runs between a base clock of 1575 MHz and a boost clock of 1770 MHz. In Thief, this allows a boost from a 208 FPS average at 1080p Low to 121 FPS at 1080p Ultra.

Q: Which CPU benchmark demonstrates the single-thread performance of this processor?

A: Cinebench R23 scores 2016.5 in single-thread and 29,300 in multi-thread. The high single-core throughput or the 6.30 GHz boost is largely why the CPU is in the 91st percentile among all CPUs.

Q: How does the CPU compare to its nearest rival in the database?

A: The Ryzen 9 7900X has an average benchmark score of 53,288, which is 0.2% higher than the AMD EPYC 7313P (53,206) but 0.3% lower than the Intel Xeon Phi 7290 (53,469). The most comparable desktop part is the Intel Core i7-14700F, which sits 0.6% below in score.

Q: Are the measured FPS numbers real-world or synthetic?

A: The "measuredFPS" field indicates that the data is coming from actual game testing, not from a synthetic read, as noted by "dataIsMeasured: true".

Q: What are the minimum and maximum FPS recorded?

A: Only the Medium preset has recorded min and max values. At 1080p Min/ Max: 142/192; at 1440p Min/ Max: 97/131; at 4K Min/ Max: 57/77.

Measured FPS Breakdown

The database reports 12 separate measured scenarios for this exact CPU/GPU combination. At 3840×2160 (4K), Thief runs at an average of 51 FPS on High, 89 FPS on Low, 67 FPS on Medium (with min 57, max 77), and 49 FPS on Ultra. At 2560×1440, the averages increase substantially: High 87, Low 151, Medium 114 (min 97, max 131), Ultra 83. At 1920×1080, the averages are High 127, Low 208, Medium 167 (min 142, max 192), Ultra 121.

The scaling pattern is uniform: Ultra is consistently lower than High by a few FPS (e.g., 121 vs 127 at 1080p, 83 vs 87 at 1440p, 49 vs 51 at 4K). Low settings provide a 1.6× or more boost over Ultra at every resolution: at 4K it's 89 vs 49 (1.8×), at 1440p 151 vs 83 (1.8×), at 1080p 208 vs 121 (1. 7×). The Medium values remain between High and Low, as expected.

For 1440p, the 87 FPS High average is nearly equal to the 1440p Ultra's 83 FPS, and the 114 FPS Medium suggests a smooth experience. The only visibility of min/max is on the Medium preset, where the ranges keep the frame times under 20 ms at 1080p (142–192) and 4K (57–77) – the latter confirming the frame pacing remains consistent even when the average dips near to 25 ms at 4K.

CPU Role

The Ryzen 9 7900X is a 12-core, 24-thread Zen 4 processor running on the AM5 platform with a base clock of 4.70 GHz and a boost of 5.60 GHz. It contains 64 MB of shared L3 cache and per-core 64 KB L1 and 1 MB L2. The CPU's single-thread performance is reflected in Geekbench 6 (2617) and 3DMark single-thread score of 1093. These metrics matter for Thief because the game's engine is largely single-thread bound in the overall frame composition; the GPU is more often the bottleneck, as seen in the scaling.

The CPU's multi-thread advantage is evident in Cinebench R23 multi-core with a score of 29,300, supporting background tasks and thread-heavy sections, but not directly pushing frame Thief. The 7900X's nearest competitive chips in the database are the AMD EPYC 7310P (avg score 53.206, delta +0.2%) and the Intel Xeon Phi 7290 (delta –0.3%), indicating that the Ryzen is a top-tier desktop part, but not an extreme outlier. In the blend of scores, with a 91st percentile rank, it ensures that the CPU does not limit frame rates at any resolution, as shown by the fact that at 1080p Low the 208 FPS remains above the GPU's 1:1 capability.

Settings Recommendations

Given the measured data, the best experience at each resolution depends on the target frame rate. For 1920×1080, the High preset at 127 fps is the highest quality that exceeds 100 fps, while Ultra at 121 fps is also above 120 fps, but only 6 fps faster at the cost of a nominal quality drop. If the user wants the largest margin, Medium at 167 fps offers nearly 50 more fps while still being visually acceptable. For 2560×1440, Medium at 114 fps keeps the frame window, while High at 87 fps is still smooth. Low provides 151 fps, but with lower visual fidelity. For 4K (3840×2160), High is the least aggressive setting that stays above 50 fps (51 fps), while Ultra 189 fps drops to 49 fps. Medium at 67 fps is still above 60; however, if 60+ fps is the desired target, Medium ist the only preset above 60 at 4K, but it has a minimal variance (min 57) that can drop below 60 in occasional frames.

Thus, if the goal is 60/144, the recommended compiler would be: 1080p Ultra (121 fps), 1440p High (83 fps), and 4K Medium (67 fps). These all have positive averages and no reported stutter issues based on the delivery.

However, the biggest frame boost per setting is seen when going from Ultra to Medium at 4K—the difference is 18 fps (49 to 67) under a significant reduction in some effects. At 1440p, dropping to Low from Ultra and an 68 fps gain (83 to 151) shows how the GPU is used.

Resolution Scaling

The scaling by resolution is the classical 4x pixel difference between 1080p and 4K. The data shows a drop from 1 P0 Ultra 121 fps to 4K Ultra 49 fps, a reduction by 60 fps, while 1080p Low 2.8 fool to 4K Low 89 fps (-119 fps). This is a semilinear behavior with resolution. At 1080p, GPU utilization is likely capped by CPU and API overhead; the CPU can push 208 fps at Low, while at 1440p it gets 18 faster than the 1080p High (127 fps). The combination of a 24-thread CPU that is 91st percentile among all CPUs and a GPU with 21.75 TFLOPS and 608.3 GB/s memory ensures a gradual fall rather than a cliff. The median stage shows that the GPU is the bottlenecks for the 4K resolution, as the CPU still has headroom (as shown by the fact that 4K Low only 89 fps, while 1080p High gives 127 fps – the CPU could have pushed a higher 4K Low, but the GPU doesn't deliver.)

In this pair, the measured results indicate that there is no stage where the spec is extremely mismatched: from Medium 4K at 67 fps to 1080p Low 208 fps, the scaling is within ~2.5× for a 4× pixel increase. The GPU is the limiting factor at all tested resolutions above 1080p Low, as evidenced by the fact that at 1080p Low the CPU delivers 208 fps, but 4K Low drops to 89 fps, a fraction of 2.3×, consistent with the ~4 million pixel counts. For 1440p Medium, the CPU still has Time for the game to operate at 114 fps, showing that the bottleneck shifts.