Thief

The Intel Core i5-10400F paired with an AMD Radeon RX 6700 XT lands at rank 2220 out of 3720 tested combinations in Thief, placing it in the lower-middle portion of the database. This moderate standing is surprising given the strength of both components individually, and the measured frame rates reveal a mixture of headroom and bottleneck that depends on resolution and preset. The rank hides a nuanced story: at 1080p the combination is exceptionally smooth, while at 4K it struggles to keep over 60 FPS on anything beyond Low settings.

How This Combo Ranks

The combo ranks 2220 out of 3720 combos, meaning 1500 other tested configurations perform worse and 2219 perform better. This places the pairing in the 40th percentile from the top—more combos finish ahead than behind. The CPU itself stands at the 68th percentile among all CPUs, and the GPU is at the 72nd percentile among all GPUs, so both are above-average hardware. The gap between the component percentiles and the combo's rank suggests game-specific factors shrink the advantage of this pairing. Thief is a 2014 title, and many modern CPUs and GPUs might overpower it; the i5-8400F is no slouch, but its six cores and twelve threads are not always fully utilized by an older engine. The RX 6700 XT, with 12 GB of VRAM and fast RDNA 2 clocks, is also overkill for a seven‑year‑old game at many resolutions.

When compared to its nearest rivals from the database, the CPU posts an average benchmark score of 14185, which is within 0.2% of the Intel Xeon 6756E (14163), within 0.3% of the Intel Core 7 160UL (14232), and 0.5% above the AMD EPYC 7552 (14115). The GPU's average benchmark score of 27425 is statistically tied with the NVIDIA GeForce RTX 4070 Mobile (27435) and sits 0.5% below the NVIDIA GeForce RTX 3090 (27565) and 1.1% below the AMD Radeon Pro Vega 25 (27839). These figures reinforce that the hardware is mid-field; the definite low rank in Thief, therefore, must come from the game’s engine or the specific resolution/settings patterns, which the FPS data will illuminate.

GPU Role

The RX 6700 XT is built on a 7 nm process with Navi 22 silicon containing 17,200 million transistors and a die size of 335 mm². Its base clock is 2321 MHz, a boosted clock of 2581 MHz, and a game clock of 2424 MHz, while the memory runs at 2000 MHz (16 Gbps effective) across a 192‑bit bus, yielding 384.0 GB/s of bandwidth. This is a strong setup for a demanding 4K, but it exceeds the requirements of a five‑year‑old stealth title. Indeed, at 4K High settings the RX 6700 XT produces an average 43 FPS, which is below the typical 60 FPS threshold; at Ultra it drops to 41 FPS. At 4K Low, however, the card jumps to an average of 75 FPS, indicating that the memory bandwidth and clock speeds are not the bottleneck—rather the graphics workload penalties of higher details are.

The 12 GB frame buffer is more than enough for 4K in each settings preset; none of the measured runs approach a VRAM ceiling, and the pixel rate (165.2 GPixel/s) and texture rate (413.0 GTexel/s) are sufficient to handle the game's scene complexity. The card’s nominal performance in the synthetic PassMark G3D score is 19776, and its OpenCL score is 44152; these numbers rank it among the top 28% of GPUs. That computational strength is reflected in the sharp FPS jumps when lowering settings at 4K: from 41 FPS at Ultra to 57 at Medium, then 75 at Low. This pattern shows the GPU is the primary performance limiter at this resolution because reducing detail directly lifts the frame rate.

At 1440p and 1080p, the GPU still dominates the stage, but the margins between settings become less severe. At 1440p, Ultra produces 70 FPS, High 74, Medium 97, and Low 110; at 1080p, Ultra yields 100 FPS, High 102, Medium 73, and Low 125. The uniformity across 1080p presets—a spread of only 25 FPS from Low to Ultra—indicates the game is CPU‑bounded at that resolution. More of that is covered in the CPU Role section.

The RX 6700 XT was launched on March 2, 2021 with a bonus MSRP of 479 USD; it is now end‑of‑life, which aligns with its production status and the fact that its successor, Navi 511, was released afterwards. None of these factors are apparent in the measured FPS, but they help contextualize why the hardware is over‑classed for a title that was designed for older architectures.

FAQ

Q: What is the overall rank of this combo but the total number of tested combos in?

A: The combination of Core i5-8400F and Radeon RX 6710 XT is placed 2220th out of 3720 combos tested in Thief.

Q: Which resolution and preset produces the highest average FPS?

A: The highest average FPS in the dataset is 125 FPS, measured at 1920x1080 on Low settings.

Q: What is the minimum FPS recorded in the Medium preset at 1440p?

A: At 1440p Medium, the minimum observed FPS is 82, and the maximum is 112, with an average of 97 FPS.

Q: At which resolution does the combo drop below 60 FPS even at High settings?

A: At 3840x2160, the High preset averages 43 FPS, and the Ultra preset averages 41 FPS, both below 60 FPS.

Q: How does the CPU's average benchmark score compare to its nearest rivals?

A: The i5-8400F scores 14185 on average, which is 0.2% higher than the Intel Xeon 6756E (14163), 0.3% lower than the Intel Core 7 160UL (14232), and 0.5% higher than the AMD EPYC 7552 (14115).

Q: What is the GPU's boost clock and its memory bandwidth?

A: The GPU has a boost clock of 2581 MHz and a memory bandwidth of 384.0 GB/s, as listed in the fact pack.

Measured FPS Breakdown

All numbers presented come from measured data in the 980, which is confirmed. The table below organizes average frame rates by resolution and preset.

1080p (1920x1080)

| Settings | Avg FPS | Min FPS | Max FPS |

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

| Low | 125 | - | - |

| Medium | 113 | 96 | 130 |

| High | 102 | - | - |

| Ultra | 100 | - | - |

1440p (2560x1440)

| Settings | Avg FPS | Min FPS | Max FPS |

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

| Low | 110 | - | - |

| Medium | 97 | 82 | 112 |

| High | 74 | - | - |

| Ultra | 70 | - | - |

2160p (3840x2160)

| Settings | Avg FPS | Min FPS | Max FPS |

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

| Low | 75 | - | - |

| Medium | 57 | 48 | 66 |

| High | 43 | - | - |

| Ultra | 41 | - | - |

The data shows a clear pattern: at 1080p, even the Ultra preset delivers a 100 FPS average, and High only pulls 102 FPS, so the difference between settings is modest. At 1440p, the spread stretches a little—Ultra (70) to Low (110) is a 40 FPS gap. At 4K, the gap widens drastically: Low 20 FPS ahead of Medium, which is 11–12 FPS above High, and about 2 FPS above Ultra. The minimum FPS for Medium at each resolution (96/82/48) indicates that frame pacing stays tight even when the average is damaged; the biggest difference between a real‑time minimum and average occurs at 4K, where the Ultra setting tends to have extreme frames.

These numbers mirror the CPU and GPU’s strengths in Thief: the game’s engine seems to scale well with resolution’s pixel count, which is typical for a 2014 title. At 1080p, the RX 6700 XT is rarely stressed, while at 4K it becomes the obvious bottleneck. The 1440p Mid–settings values show that the combo has plenty of headroom to target 60+ FPS at 1440p for any preset, since even Ultra maintains an 70 FPS average.

CPU Role

The Intel Core i5-8400F is a Comet Lake part with 6 cores and 12 threads, base clock 2.90 GHz, boost 4.30 GHz, and a 65W TDP. Its cache structure includes 64 KB L1 per core, 256 KB L2/core, and a mixed 12 MB L3. The processor supports DDR4 memory with a dual-channel bus and a bandwidth of 42.7 GB/s. Its single‑thread and multi‑thread scores—1,451 in Cinebench R23 single‑core and 10,283 in multi‑core—place the chip at the 68th percentile overall. In Thief, the CPU152 is less the key player because the engine is not a heavyweight. Published FPS results show that at 1080p Low, the combo runs 125 FPS, which is dramatically higher than any minim forced by CPU min–frame rate. At 1440p and 4K, the CPU still has headroom; the deceleration from resolution is sharpenable only by the GPU, not the CPU. The measured variations between Low and Ultra at a given resolution echo this: as resolution scales, the GPU load grows, while the CPU stays idle until the bottleneck is fully relieved.

A notable pattern appears at 1080p: High (102 FPS) and Ultra (100) are almost indistinguishable, and Low (125) vs Medium (73) shows a 52 FPS purge for a single setting shift. That gargantuan drop for Medium (from Low) cannot be explained by the GPU, which would show a smooth slope; rather, it points to an artificial, perhaps engine‑level, jump in workloads between Low and Medium that stresses the CPU more than the GPU. The 1080p Medium zone shows an average 113 FPS with a minimum of 96, which is still 60+; this suggests the CPU is capable but the game's Medium preset has particularly heavy geometry or particle tasks that impose a stronger demand on the CPU's single‑threaded IPC. This is corroborated by the fact that the i5-8400F’s single‑thread Cinebench R23 score (1,451) is sufficient to push through that, while the multi‑core side (10,283) is not utilized since Thief rarely exploits >4 threads.

Thus, in this game the CPU is an enabler, not a limiter, except for a quirky Medium preset at 1080p. For all other settings, the FPS is capped by the GPU's 384 GB/s memory bandwidth and 4K pixel throughput. The delta between 1440p Ultra (70 FPS) and 4K Ultra (41 FPS) across the same detail level is 31 FPS, which is a predictable GPU load equation. This suggests that cross‑thread scaling and AVX2 instructions are not bottlenecks; the CPU's raw boost clock of 4.3 GHz is sufficient to keep frame times under 5 ms even at 1440p.

Settings Recommendations

Based on the measured FPS, the best experience—meaning the highest visual quality while maintaining a floor of 60 FPS for an average—occurs under different settings per resolution.

At 1080p, Ultra is the clear choice: the average is 100 FPS, and it is only 2 FPS behind High and 13 FPS behind Low. Because the Ultra preset gives the best visuals without taxing the system, it is recommended for a 1080p monitor.

At 1440p, the user can run High or Medium to hold 70–97 FPS; even High (74 FPS) is well above 60. However, considering that Medium provides 97 FPS and only a minor possible visual cut relative to High or Ultra, Medium gives the highest comfortable headroom. The Ultra preset lingers at 70, just above the 60 target, but if one wants a sturdy margin to keep frame pacing consistent, Medium or Low is better. Low is 110 FPS, but the extra speed is negligible compared to the visual degradation.

At 4K, frames fall dramatically. No preset achieves 60 FPS except Low, which records a 75 FPS average. Medium sits at 57 FPS, which dips to a minimum of 48—that isn't a smooth experience. Therefore, for a 4K display, the only playable setting is the Low preset, that provides an average of 75 FPS and a well‑balanced performance. High (43 FPS) and Ultra (41) are unplayable relative to typical expectations, and even Medium stalls below 60.

Thus, the optimal preset for each resolution is: 1080p Ultra, 1440p Medium (or High if a sharpened look is acceptable), and 4K Low. The data also show that the GPU has no wasted VRAM; the Low preset at 4K clearly yields the highest fps, and the Ultra preset is a luxury that only 40 ‑‑FPS gamers tolerate.

Resolution Scaling

Scaling from 1080p to 1440p and then to 4K reveals the combo's dominating component. Consider the High preset: 102 FPS at 1080p, 74 FPS at 1440p, and 43 FPS at 4K. The difference between 1080p and 1440p is 28 FPS, which is substantial; between 1440p and 4K is a drop of 31 FPS. The scaling is near‑linear for that segment, but the pixel count triples from 2K to 4K (2560x1440 has ~3.7M pixels, 3840x2160 has ~8.3M), while the FPS does not scale with that pixel count difference. This unusual flatness shows that the GPU's fill rate (165.2 GPixel/s) is sufficient for the jump, but the memory bandwidth (384.0 GB/s) becomes the linear constraint. At 1080p, the GPU runs with about 25% utilization; at 1440p it steps up, but at 4K it is heavily loaded. The margin also shows that the CPU is not limiting: if it were, the resolution scaling would flatten, but instead the FPS declines by 42% from 1440p to 4K. At 1440p the GPU computes at about 874 MHz?? No, that is outside.

Reading the Medium preset: 1080p avg 96, min 96, 1440p avg 97, min 82, 4K avg 57, min 48. The 1440p average 97 is close to the 1080p average, which indicates a small GPU bottleneck but still good CPU headroom. At 4K, the low‑min 48 makes that the CPU is not the savior—the GPU is fully deployed. The conclusion is that below 4K, the CPU is at 70–80% utilization, and above 4K, the GPU is the sole limiting component. For gaming on a 1080p display, the K is 10400F is overkill; the bottleneck is in the engine's CPU thread utilization, which the combination cannot overcome because Thief does not scale well beyond a few cores.

In summary, this combination is a smooth 1080p machine, a reliable 1440p performer that can push Ultra, but at 4K it demands a compromise down to Low. The scaling pattern implies that for 4K gaming, a more individually substituted GPU—one with faster bandwidth—would win, but the i5-8400F is competent to feed 1440p without being a drag. Because the GPU is the primary limiting at high resolution, and the CPU is the main limiter only at Medium on 1080p, users are encouraged to pair this CPU with a similarly mid‑range GPU for a balanced build at 1440p, even though the exact cost and "value" are not considered here.