Thief
This combination offers playable performance with an average of 39 FPS, though you may want to lower settings for smoother gameplay.
Average FPS: resolution vs quality settings
| Ultra | High | Medium | Low | |
|---|---|---|---|---|
| 4K Ultra HD | 17 | 18 | 23 | 31 |
| 1440p QHD | 29 | 30 | 40 | 53 |
| 1080p Full HD | 42 | 44 | 58 | 77 |
Cell color: green is 120+ FPS, teal is 60+, amber is 30+, red is below 30.
Thief with Intel Core Ultra X9 388H + AMD Radeon 890M, In-Depth Analysis
Resolution Scaling
The measured FPS data for Thief shows a steep, predictable drop as resolution climbs from 1920x1080 to 3840x2160. At Low settings, the average frame rate falls from 77 FPS at 1080p to 53 FPS at 1440p, then down to 31 FPS at 4K. That is a 59.7% reduction from 1080p to 4K, which clearly indicates the integrated AMD Radeon 890M is the primary bottleneck at higher pixel counts. The scaling is not linear — the gap between 1440p and 4K (a 41.5% drop) is steeper than the gap between 1080p and 1440p (a 31.2% drop), suggesting that the GPU's shading and memory bandwidth resources are exhausted more quickly as pixel count grows.
At High settings, the pattern repeats but with lower absolute numbers: 44 FPS at 1080p, 30 FPS at 1440p, and 18 FPS at 4K. The percentage drop from 1080p to 4K here is 59.1%, nearly identical to Low settings. This consistency across presets implies that resolution scaling is dominated by the same limiting factor — the Radeon 890M's raw throughput — regardless of how much graphical detail is enabled. When you halve the resolution from 4K to 1080p at High settings, the frame rate more than doubles (from 18 to 44 FPS), which is a classic sign of a fill-rate or pixel-throughput limit rather than a CPU or driver overhead limit.
The Medium preset tells a similar story, with averages of 58 FPS at 1080p, 40 FPS at 1440p, and 23 FPS at 4K. Notably, Medium is the only preset with recorded minimum and maximum FPS values: at 1080p you get a min of 49 and a max of 67, at 1440p a min of 34 and a max of 46, and at 4K a min of 20 and a max of 27. These tight ranges indicate that frame pacing remains stable even when the GPU is under heavy load — there are no dramatic spikes or crashes, just a consistent degradation as resolution rises. The min-to-max spread at each resolution is roughly 18 FPS, which suggests the engine's workload is fairly uniform.
Ultra settings are the most demanding, producing 42 FPS at 1080p, 29 FPS at 1440p, and only 17 FPS at 4K. The 1080p-to-4K drop is 59.5%, again matching the other presets. This uniformity across all four quality levels is the key takeaway: resolution scaling in Thief is remarkably consistent, with the Radeon 890M losing roughly 60% of its performance when moving from 1080p to 4K, independent of settings. For practical purposes, 1440p is the realistic ceiling for playable frame rates on this iGPU, and even then only at Low or Medium settings. 4K should be considered a non-starter for anything beyond static screenshots, since every preset falls below 31 FPS.
FAQ
Q: Can this system run Thief at 4K with any playable frame rate?
A: No. The data shows 4K averages of 18 FPS on High, 23 FPS on Medium, 31 FPS on Low, and 17 FPS on Ultra. Even the best case (Low) sits at 31 FPS, which is below the typical smoothness threshold for a stealth game where quick reactions matter. The 4K Medium preset has a minimum of 20 FPS, indicating frequent dips below 30.
Q: What is the best resolution for a consistent 60 FPS experience?
A: At 1920x1080 with Low settings, the average FPS is 77, which gives headroom above 60 even during demanding scenes. The 1080p Medium preset averages 58 FPS with a minimum of 49, so it will occasionally drop below 60. No other resolution or settings combination in the measured data reaches a 60 FPS average — the next closest is 1440p Low at 53 FPS.
Q: How does the CPU compare to its nearest rivals in overall benchmark scores?
A: The Intel Core Ultra X9 388H has an average benchmark score of 44466, which places it in the 88th percentile of all CPUs. Its closest rival, the AMD Ryzen 5 7500X3D, scores 44573, a mere 0.2% difference. The Intel Core i9-13950HX trails by 0.3% at 44342, and the AMD Ryzen AI Max 385 is 0.4% behind at 44309. This puts the Ultra X9 388H in a tight cluster where performance parity is the norm.
Q: Is the integrated GPU comparable to any dedicated graphics cards?
A: The AMD Radeon 890M's average benchmark score is 9210, placing it in the 45th percentile of all GPUs. Its nearest rival is the AMD Radeon Vega 8 at 9221 (0.1% difference), followed by the NVIDIA GeForce GTX 960 at 9273 (0.7% ahead) and the NVIDIA GeForce GTX 465 at 9294 (0.9% ahead). The NVIDIA GeForce GTX 850M is 1% ahead at 9302. These are all older discrete GPUs, which gives context for the 890M's performance class.
Q: What settings preset offers the best balance of visual quality and frame rate at 1080p?
A: The Medium preset at 1920x1080 delivers 58 FPS average with a minimum of 49 FPS, which keeps gameplay fluid while providing better visuals than Low (77 FPS). High settings drop to 44 FPS, which is still playable but leaves less margin for dips. For a stealth game where you need to track moving guards and environmental details, Medium strikes the best compromise in the measured data.
Q: How does the GPU's memory configuration affect performance in this game?
A: The Radeon 890M uses System Shared memory, with bandwidth described as "System Dependent." This means the GPU relies on the same LPDDR5X memory as the CPU, over a dual-channel bus providing 153.6 GB/s. In Thief, this shared bandwidth becomes a limiting factor at higher resolutions, which explains why the FPS drop from 1080p to 4K is so consistent — the memory subsystem cannot feed the GPU enough data at high pixel counts regardless of settings.
CPU Role
The Intel Core Ultra X9 388H is a 16-core, 16-thread processor based on the Panther Lake architecture, built on Intel's 3 nm process node. Its base clock is 2.10 GHz with a boost clock of 5.10 GHz, which provides substantial single-thread headroom for game logic. The chip draws a 25 W TDP, making it a low-power mobile part, yet its multi-core benchmark results are impressive: 2955 in Cinebench R15 multi-core, 13101 in R20, and 18911 in R23. Single-core scores are equally strong for a mobile chip, hitting 309.5 in R15, 1849 in R20, and 2200.5 in R23.
In Thief, the CPU's role is less about raw multi-core throughput and more about maintaining consistent frame pacing. The game, released in 2014, was designed for the multi-core CPUs of that era, and a modern 16-core processor with a 5.10 GHz boost clock has far more headroom than the game can utilize. The Passmark multi-thread score of 36811 and single-thread score of 4280 indicate that the CPU is not the bottleneck in any of the measured scenarios. Even at 1080p Low, where the GPU is least stressed, the average FPS of 77 is well below what this CPU could theoretically push — the limit is clearly the integrated graphics.
The cache hierarchy is generous: 192 KB of L1 per core, 3 MB of L2 per core, and 18 MB of shared L3. This large L3 cache helps keep frequently accessed game data close to the cores, reducing memory latency that could otherwise stall frame rendering. The dual-channel LPDDR5X memory configuration, with 153.6 GB/s bandwidth, is shared with the GPU, but for CPU-bound tasks like AI pathfinding or physics calculations in Thief, this bandwidth is more than sufficient. The CPU's 88th percentile ranking among all CPUs, with an average benchmark score of 44466, places it in the top tier of mobile processors, and its nearest rivals — all within 0.5% of its score — confirm that it is not holding back gaming performance.
The 16 threads (one per core, no hyperthreading) are an interesting choice, as most modern CPUs use two threads per core. However, Thief's engine was built for an era when 4-8 threads were typical, so the physical cores provide more than enough parallel execution units. The Passmark physics score of 3226 and extended instructions score of 29943 suggest the CPU can handle complex game logic without strain. In practice, the data shows that the Ultra X9 388H is a "good enough" partner for the Radeon 890M — it never becomes the limiting factor, and the measured FPS curves are purely a function of GPU load.
Measured FPS Breakdown
At 1920x1080, the full range of settings is playable, with the Low preset delivering 77 FPS average — the highest number in the entire dataset. Medium drops to 58 FPS, with a minimum of 49 and maximum of 67, indicating that even in busy scenes with multiple guards or complex lighting, the frame rate stays above 45 FPS. High settings yield 44 FPS, which is borderline for a stealth game where smooth camera panning matters, but still usable. Ultra settings produce 42 FPS, only 2 FPS below High, suggesting that the Ultra preset's additional effects (likely shadow quality and post-processing) have diminishing returns on this hardware.
At 2560x1440, the playability drops significantly. Low settings average 53 FPS, which is fine for most players, but Medium drops to 40 FPS with a minimum of 34 — noticeably less smooth. High settings fall to 30 FPS, which is the bare minimum for console-like playability, and Ultra drops to 29 FPS. The gap between Low and Medium at 1440p is 13 FPS, while the gap between Medium and High is 10 FPS, showing that each quality step costs roughly the same amount of performance. The 1440p Medium minimum of 34 FPS is concerning for action sequences, as sudden drops to the low 30s can cause noticeable stutter.
At 3840x2160, the picture is bleak. Low settings average 31 FPS, which is technically playable but far from smooth. Medium averages 23 FPS with a minimum of 20, High averages 18 FPS, and Ultra averages 17 FPS. The 4K results are uniformly below 30 FPS, and even the best case (Low) has no headroom for dips. The Medium preset's max FPS of 27 is lower than the Low preset's average of 31, confirming that raising quality at 4K has an outsized impact. The measured data shows a clear pattern: 1080p is the native resolution for this iGPU, 1440p requires low-to-medium settings, and 4K is only suitable for screenshot purposes or very patient players.
The combination rank of 2692 out of 2851 combos puts this system in the bottom 6% of all tested configurations for Thief. This is not surprising given that the Radeon 890M is an integrated GPU, but it contextualizes the performance: this is a machine that can run Thief at 1080p with decent settings, but it is far from a high-end gaming rig. The lack of minimum FPS data for most settings (only Medium has min/max recorded) means the averages may hide occasional stutters, but the Medium data suggests frame pacing is stable.
GPU Role
The AMD Radeon 890M is an integrated graphics processor based on the RDNA 3.5 architecture, fabricated on TSMC's 4 nm process. It packs 1024 shading units, 64 texture mapping units, and 32 ROPs, along with 16 ray tracing cores. The boost clock reaches 2900 MHz, with a base clock of 400 MHz, and the chip has a 15 W TDP. These specifications produce 5.939 TFLOPS of FP32 performance, a texture rate of 185.6 GTexel/s, and a pixel rate of 92.80 GPixel/s. For an iGPU, these are respectable numbers, but they place the 890M in the 45th percentile of all GPUs, with an average benchmark score of 9210.
The GPU's nearest rivals are all older discrete graphics cards: the AMD Radeon Vega 8 (9221, 0.1% faster), NVIDIA GeForce GTX 960 (9273, 0.7% faster), NVIDIA GeForce GTX 465 (9294, 0.9% faster), and NVIDIA GeForce GTX 850M (9302, 1% faster). This means the 890M performs like a mid-range discrete GPU from roughly a decade ago, which is exactly what the Thief FPS data confirms. At 1080p Low, the 77 FPS average is respectable, but the 4K results show the GPU running out of steam — the 17-31 FPS range is typical of a chip that cannot fill enough pixels per second.
The memory configuration is the 890M's Achilles' heel for high-resolution gaming. It uses System Shared memory, meaning it draws from the same LPDDR5X pool as the CPU, with bandwidth rated as "System Dependent." In this system, the dual-channel memory provides 153.6 GB/s, which is shared between CPU and GPU. At 4K, the GPU needs to access texture data and write frame buffers at a much higher rate than at 1080p, and the shared memory bus becomes a bottleneck. This explains why the FPS drop from 1080p to 4K is so consistent across all settings — it is a memory bandwidth limit, not a compute or shading limit.
The GPU's DirectX 12 Ultimate support (12_2) and Vulkan 1.4 API compatibility mean it can run modern rendering techniques, but Thief is a 2014 game that uses DirectX 11 primarily. The Passmark DirectX 11 score of 73 is the highest among the DirectX tests (DirectX 9 scores 97, DirectX 12 scores 37, DirectX 10 scores 33), which aligns with the game's era. The 3DMark Steel Nomad DX12 score of 590 and Geekbench Vulkan score of 40808 are more relevant for modern titles, but for Thief, the older DirectX 11 path is what matters. The GPU's 16 ray tracing cores are unused in this game, so they do not contribute to the measured performance.
The Radeon 890M's 34,000 million transistors on a 233 mm² die, with a transistor density of 145.9M per mm², show that AMD has packed a lot of hardware into a small space. However, the 15 W TDP limits sustained clocks, and the boost clock of 2900 MHz is likely only achievable in short bursts. In Thief, where the GPU is under continuous load, thermal throttling may reduce clocks below the boost figure, contributing to the lower-than-expected 4K performance. The pixel rate of 92.80 GPixel/s sounds high, but at 4K (8.3 million pixels), the theoretical maximum fill rate is only about 11 frames per second — and the measured 17-31 FPS suggests the game is not fill-rate limited but rather shader or texture-bound.
Settings Recommendations
For a smooth, consistent experience in Thief on this hardware, the data points clearly to 1920x1080 with Medium settings as the sweet spot. The 58 FPS average with a minimum of 49 FPS provides enough headroom to avoid noticeable stutter during quick camera movements or crowded scenes, while the Medium preset offers significantly better visuals than Low. The max FPS of 67 means there are moments of higher performance, but the 49 FPS minimum ensures the experience never becomes choppy. If you prefer maximum smoothness over visual fidelity, 1080p Low at 77 FPS is the best choice — it is the only configuration that stays comfortably above 60 FPS at all times.
At 1440p, the recommendation shifts to Low settings, which average 53 FPS. This is playable, but the lack of minimum FPS data for this preset means you should expect some dips below 50 FPS in heavy scenes. The 1440p Medium preset, at 40 FPS with a 34 FPS minimum, is risky for a stealth game where you need to track enemies — the low 30s are noticeably less fluid. Avoid 1440p High (30 FPS) and Ultra (29 FPS) unless you are willing to accept a borderline cinematic frame rate.
4K is not recommended for actual gameplay on this system. The best 4K result is Low at 31 FPS, which feels sluggish for a game that requires precise aiming and quick reactions. If you must run at 4K, use Low settings and expect occasional drops below 30 FPS. The Medium 4K preset's minimum of 20 FPS makes it effectively unplayable for action sequences. For the best experience, treat 4K as a screenshot-only resolution or drop to 1440p for a playable compromise.
The Ultra preset offers no meaningful advantage over High at any resolution — at 1080p, Ultra is 42 FPS versus High's 44 FPS, a negligible 2 FPS difference. The same pattern holds at 1440p (29 vs 30 FPS) and 4K (17 vs 18 FPS). This suggests the Ultra preset's extra effects are either not rendering on this hardware or have a minimal visual impact relative to their cost. Stick with Medium for the best balance, or High if you want better shadows and textures and can tolerate 44 FPS at 1080p. The measured data does not support using Ultra presets on the Radeon 890M, as the performance cost is not justified by any measurable gain in the FPS numbers.
Hardware Specifications
Intel Core Ultra X9 388H
AMD Radeon 890M
FPS Benchmarks by Resolution & Settings
Performance data for Intel Core Ultra X9 388H + AMD Radeon 890M in Thief
| Resolution | Settings | Avg FPS |
|---|---|---|
| 3840x2160 | Low | 31.0 |
| 3840x2160 | Medium | 23.0 |
| 3840x2160 | High | 18.0 |
| 3840x2160 | Ultra | 17.0 |
| 2560x1440 | Low | 53.0 |
| 2560x1440 | Medium | 40.0 |
| 2560x1440 | High | 30.0 |
| 2560x1440 | Ultra | 29.0 |
| 1920x1080 | Low | 77.0 |
| 1920x1080 | Medium | 58.0 |
| 1920x1080 | High | 44.0 |
| 1920x1080 | Ultra | 42.0 |
Thief with iUltra X9 388H + Other GPUs
See how Thief performs with the same CPU but different graphics cards.
Other Games with iUltra X9 388H + 890M
Explore FPS benchmarks for other games using this same hardware combination.