Doom

The Intel Core Ultra 5 125H paired with the NVIDIA GeForce RTX 5090 is an extreme pairing of a mobile processor with a flagship desktop GPU, and in Doom (2016), the benchmark results show this combo ranks at #788 out of 3712 tested combinations. This places it in the top 21% of all tested combos in the database, which is a strong showing, but the data suggests this is not purely a GPU-dominated victory. The rank reflects the 125H’s ability to feed the RTX 5090 at high frame rates, though the CPU’s mobile pedigree becomes evident in the scaling patterns across resolutions and settings.

The RTX 5090 is a massive GPU, with 21,760 shading units and a 512-bit memory bus delivering 1.79 TB/s of bandwidth. The Ultra 5 125H, by contrast, is a 28 W mobile part with 14 cores (18 threads) and a boost clock of 4.50 GHz. In Doom, which is a well-optimized Vulkan/DirectX 12 title, the combo’s rank is respectable but not elite; the data shows that the GPU is often waiting on the CPU, particularly at lower resolutions. The percentile ranking for the GPU alone is 92nd, while the CPU sits at 79th, so the bottleneck is clearly the processor in many scenarios.

How This Combo Ranks

At #788 out of 3712 combos, this pairing lands in the 78.8th percentile of all tested systems. For a game like Doom, which is known for its scalability, this rank is solid but not spectacular. The RTX 5090 is capable of far more, but the Ultra 5 125H’s mobile architecture (Meteor Lake, 7 nm process) limits how much of that potential is realized. The combo’s rank is roughly 21% from the top, meaning there are 787 combos that outperform it; these are almost certainly pairings with desktop-class CPUs that can keep up with the 5090’s raw throughput.

Benchmark data for the CPU shows an average score of 27,507, with nearest rivals including the AMD Ryzen 7 5800 (delta -0.1%) and Intel Core i5-12600K (delta -0.3%). This means the 125H is essentially on par with those desktop chips in synthetic tests, which is surprising for a mobile part. However, in Doom specifically, the FPS results reveal that the CPU’s 18 threads and 14 cores are adequate, but not exceptional, for feeding the GPU. The combo’s rank improves at higher resolutions where the GPU workload dominates, but at 1080p, the CPU becomes the limiting factor, pulling the overall rank down.

The GPU’s own percentile is 92nd, and its nearest rivals include the Tesla P100 and Radeon RX 6850M XT, but those are not relevant here. What matters is that the 5090 is in the top 8% of GPUs, yet the combo only reaches the top 21% of combos. The gap between the GPU’s percentile and the combo’s rank is the CPU’s contribution—or lack thereof. The 125H’s PassMark single-thread score of 3,399 is decent, but its multi-thread score of 21,100 is modest compared to what the 5090 can demand. In a game like Doom, which can scale to many threads, the 14-core/18-thread configuration helps, but the mobile power envelope (28 W TDP) means sustained clocks are lower than desktop parts.

FAQ

*Q: Is this combo overkill for Doom at 1080p?*

A: Yes, the data shows 1080p Ultra still hits 198 FPS average, and Low settings reach 330 FPS. The GPU is rarely stressed at this resolution; the CPU becomes the bottleneck, as evidenced by the scaling from Low to Ultra.

Q: What is the best resolution to use with this combo?

A: 2560x1440 appears to be the sweet spot. High settings deliver 157 FPS average, which is well above the 144 Hz threshold for most monitors, and even Ultra holds 128 FPS. The GPU is more utilized here than at 1080p, but the CPU still keeps up.

Q: How does the Ultra 5 125H compare to its nearest CPU rivals in synthetic tests?

A: The 125H’s average benchmark score is 27,507, which is within 0.3% of the Ryzen 7 5800 (27,535) and the Core i5-12600K (27,578). It is 0.5% faster than the Ryzen 3 PRO 5355G (27,382). These deltas are negligible in real-world gaming.

*Q: Does the RTX 5090’s 32 GB of GDDR7 memory matter for Doom?*

A: Doom is not a VRAM-heavy game at these settings; the 32 GB frame buffer is irrelevant for this title. The 1.79 TB/s bandwidth is more relevant, as it helps at 4K Ultra where the game still demands 70 FPS.

Q: Why is the 4K Ultra FPS so much lower than 4K High?

A: The jump from High (85 FPS) to Ultra (70 FPS) is an 18% drop, while the jump from High to Medium (101 FPS) is a 19% increase. Ultra settings impose a significant GPU load, and the CPU’s 18 threads are not enough to fully mitigate the frame time spikes at that preset.

Q: Is the CPU or GPU the limiting factor at 4K?

A: At 4K, the GPU is the primary limiter. The FPS drops from 116 (Low) to 70 (Ultra), a 40% reduction, which is typical of GPU-bound scenarios. The CPU can still deliver 116 FPS at 4K Low, indicating it is not the bottleneck there.

Settings Recommendations

For the best experience with this combo, the measured data points to 2560x1440 with High settings as the optimal balance. At this configuration, the average FPS is 157, which is smooth for high-refresh-rate monitors, and the minimum FPS (not recorded for High, but Medium shows 158 min) suggests stability. Moving up to Ultra at 1440p drops the average to 128 FPS, which is still playable but begins to approach the threshold where frame time variance might be noticeable. The 1440p Medium preset offers 186 FPS, which is excellent for competitive play, but High is the sweet spot for visual fidelity without sacrificing smoothness.

At 1080p, the data shows that High settings (243 FPS) are already above the refresh rate of most monitors, and Ultra (198 FPS) is still extremely high. There is no reason to use Low (330 FPS) or Medium (287 FPS) at this resolution unless you are chasing absolute maximum frame rates for esports. However, the CPU’s limitations become apparent here: the gap between 1080p Low (330 FPS) and 1080p Ultra (198 FPS) is 40%, which is larger than the gap at 1440p (214 to 128, a 40% drop as well). This suggests the CPU is struggling to maintain the highest FPS figures.

For 4K, the recommendation is High settings (85 FPS) over Ultra (70 FPS) or Medium (101 FPS). The jump from High to Ultra is a 17.6% FPS loss for marginal visual gains, while Medium offers a 18.8% improvement over High but with noticeably reduced detail. The 4K Low preset (116 FPS) is viable if you prioritize frame rate over visuals, but the game’s art style benefits from higher settings. The data indicates that 4K Medium (101 FPS) is the best compromise if you must play at 4K, as it stays above 100 FPS while maintaining acceptable detail.

Measured FPS Breakdown

At 1920x1080, the combo delivers exceptional performance across all presets. Low settings produce an average of 330 FPS, which is the highest recorded for this combo. Medium settings drop to 287 FPS (with a minimum of 244 and maximum of 331), High settings achieve 243 FPS, and Ultra still holds 198 FPS. The scaling from Low to Ultra is a 40% reduction, indicating that the CPU is not able to fully saturate the GPU at the lowest settings, but the GPU is still capable of massive frame rates.

At 2560x1440, the numbers are still impressive but show a clearer GPU influence. Low settings average 214 FPS, Medium drops to 186 FPS (min 158, max 214), High achieves 157 FPS, and Ultra falls to 128 FPS. The drop from Low to Ultra is 40.2%, nearly identical to the 1080p scaling. This consistency suggests that the CPU is still a factor, but the GPU is beginning to take on more work as resolution increases.

At 3840x2160, the GPU becomes the dominant limiter. Low settings average 116 FPS, Medium drops to 101 FPS (min 86, max 116), High achieves 85 FPS, and Ultra falls to 70 FPS. The drop from Low to Ultra is 39.7%, which is similar to the lower resolutions, but the absolute numbers are lower. The minimum FPS recorded at 4K Medium (86 FPS) is still above 60, but the 4K Ultra figure of 70 FPS average suggests that frame pacing could be an issue in intense scenes.

The data reveals a pattern: the percentage drop from Low to Ultra is consistent (~40%) across all resolutions, which is unusual. Typically, higher resolutions show a larger drop due to GPU load, but here it is stable. This indicates that the CPU is limiting the maximum FPS at Low settings, and the GPU is limiting the minimum FPS at Ultra settings, creating a balanced bottleneck that scales linearly with resolution.

Resolution Scaling

The FPS drops from 1080p to 4K are substantial but predictable. At Low settings, the average FPS goes from 330 (1080p) to 116 (4K), a 64.8% reduction. At Medium, it goes from 287 to 101, a 64.8% reduction. At High, it goes from 243 to 85, a 65.0% reduction. At Ultra, it goes from 198 to 70, a 64.6% reduction. The consistency of these percentages—all around 65%—is a clear sign that the scaling is resolution-bound, not settings-bound.

This uniform scaling suggests that the limiting component is the GPU, but not in the way one might expect. The RTX 5090 has 176 ROPs and a pixel rate of 423.6 GPixel/s, which should handle 4K easily. The fact that the FPS drops by a fixed percentage across all settings implies that the CPU is setting a ceiling at 1080p (330 FPS at Low) that is not achievable at 4K, but the GPU is still the primary constraint at 4K. The 125H’s 18 threads are enough to maintain a high frame rate at 1080p, but at 4K, the GPU’s fill rate and memory bandwidth (1.79 TB/s) become the deciding factors.

The data shows that the combo loses about 65% of its 1080p performance at 4K, which is typical for a high-end GPU paired with a mid-range CPU. If the CPU were the sole bottleneck, the drop would be less severe at 4K, as the GPU would be more utilized. Instead, the consistent drop indicates that both components are working near their limits, but the GPU is the one that cannot keep up with the increased pixel count. This is a healthy balance for a gaming system, as it means neither component is completely wasted.

CPU Role

The Intel Core Ultra 5 125H is a 14-core, 18-thread mobile processor with a base clock of 3.60 GHz and a boost clock of 4.50 GHz. Its L3 cache is 18 MB shared, and it has 2 MB L2 per core. In Doom, which can utilize multiple threads, this configuration is adequate but not ideal. The CPU’s PassMark multi-thread score is 21,100, and its single-thread score is 3,399. These numbers place it in the 79th percentile of all CPUs, which is good for a mobile part but not exceptional for gaming.

The data shows that the CPU is not the primary bottleneck at 4K, but it does limit the maximum FPS at lower resolutions. At 1080p Low, the combo hits 330 FPS, which is likely close to the CPU’s limit for this game. The fact that increasing resolution to 1440p Low only drops to 214 FPS (a 35% reduction) while increasing to 4K Low drops to 116 FPS (a 65% reduction from 1080p) suggests that the CPU can sustain around 200+ FPS, but the GPU cannot deliver at 4K.

The 125H’s 18 threads are sufficient for Doom’s engine, which was designed to scale well on multi-core CPUs. However, the mobile power envelope (28 W TDP) means that sustained boost clocks are likely lower than the 4.50 GHz maximum, especially in a system with a 575 W GPU. This thermal and power constraint is the likely reason why the combo does not rank higher. The CPU’s nearest rivals, such as the Ryzen 7 5800 and i5-12600K, have similar synthetic scores (within 0.3%), but those are desktop parts with higher sustained clocks, which would likely perform better in Doom at 1080p. The data shows the 125H is a capable gaming CPU, but it is not a match for the RTX 5090’s full potential in a CPU-bound scenario.