Microsoft Flight Simulator

The Intel Core Ultra X9 388H paired with the NVIDIA GeForce RTX 5090 lands at rank 586 out of 2996 tested combos in Microsoft Flight Simulator, placing it in the 80th percentile of all configurations measured for this title. That rank reflects a system with substantial headroom for high-refresh monitors at 1080p and 1440p, while still delivering a playable experience at 4K with the right settings. The combination of a 16-core mobile processor and a flagship desktop GPU is unusual, and the benchmark results show exactly where each component carries the load.

How This Combo Ranks

Out of 2996 tested combos, this pairing sits at 586, which means roughly 80% of all other CPU-GPU combinations tested in Microsoft Flight Simulator produce lower average frame rates. The rank is heavily influenced by the GPU, since this game scales strongly with graphics hardware, but the CPU still matters for simulation logic and draw calls. The RTX 5090 is a top-tier card, and its presence alone pushes this combo well above the median. The 388H, while not the fastest mobile CPU available, provides enough single-thread performance to avoid bottlenecking the GPU at most settings.

Comparing the CPU to its nearest rivals, the 388H’s average benchmark score of 44466 is nearly identical to the AMD Ryzen 5 7500X3D, which scores 44573, a delta of -0.2%. The Intel Core i9-13950HX is just 0.3% behind at 44342, and the AMD Ryzen AI Max 385 is 0.4% behind at 44309. The Intel Core i5-13600 trails by 0.5% with a score of 44240. These deltas are within noise, meaning the 388H is statistically tied with those four processors in general compute benchmarks. For Microsoft Flight Simulator, that means the CPU is not a differentiator among this peer group; the GPU will dominate the frame rate differences.

The GPU’s average benchmark score of 79842 places it at the 92nd percentile of all GPUs. Its nearest rivals include the NVIDIA Tesla P100 PCIe 16 GB at 79605 (0.3% slower), the Tesla P100 PCIe 12 GB at 79396 (0.6% slower), and the AMD Radeon RX 6850M XT at 78940 (1.1% slower). The AMD Radeon Pro Vega 64X is 1.4% faster at 80959. These are older or workstation-oriented cards, which underscores how far ahead the RTX 5090 is in raw compute. In gaming, the RTX 5090’s architecture and driver support will matter more than these synthetic scores, but the percentile rank confirms it is a top-8% GPU.

CPU Role

The Intel Core Ultra X9 388H is a 16-core, 16-thread processor from the Core Ultra Series 3, built on the Panther Lake architecture using a 3 nm process. Its base clock is 2.10 GHz, boosting to 5.10 GHz. That boost clock is critical for Microsoft Flight Simulator, which relies heavily on single-thread performance for the main simulation loop, including flight physics, AI traffic, and scenery loading. The 388H’s single-thread score in Cinebench R23 is 2200.5, which is modest compared to desktop chips but respectable for a mobile part with a 25 W TDP.

The L3 cache is 18 MB shared, with 3 MB L2 per core and 192 KB L1 per core. For a flight simulator that streams terrain and aircraft data, cache size can reduce memory latency, but the 388H’s 18 MB L3 is on the lower side for modern CPUs. The memory support is LPDDR5X over a dual-channel bus, providing 153.6 GB/s of bandwidth. That bandwidth is sufficient for feeding the RTX 5090, but it is a fraction of what the GPU’s own 1.79 TB/s memory bandwidth can handle. In practice, the CPU’s memory bandwidth is rarely the limiting factor in this game; the bottleneck is more likely the CPU’s single-thread speed.

The 388H’s 16 threads are all physical cores, with no hyper-threading. That is a deliberate choice for efficiency, but it means the CPU cannot rely on extra logical threads to handle background tasks. In Microsoft Flight Simulator, which can use multiple cores for terrain generation and weather simulation, the 16 threads are adequate, but not overkill. The Cinebench R23 multicore score of 18911 is about 8.6 times the single-core score, showing decent scaling, but the game’s frame rate will track the single-core score more closely.

The CPU’s 88th percentile ranking among all CPUs indicates it is faster than most processors on the market, but among its nearest rivals it is essentially tied. The delta of -0.2% to the Ryzen 5 7500X3D means the 388H and that AMD chip are interchangeable in compute workloads. In Microsoft Flight Simulator, the difference would be imperceptible. The 388H’s integrated graphics, Arc B390, are not used in this combo, so they have no impact on the measured FPS.

GPU Role

The NVIDIA GeForce RTX 5090 is the primary driver of frame rates in this combo. It features the GB202 chip on the Blackwell 2.0 architecture, built on a 5 nm process at TSMC. The GPU has 32 GB of GDDR7 memory on a 512-bit bus, delivering 1.79 TB/s of bandwidth. That memory capacity is enormous for a flight simulator, which can load large photogrammetry areas and high-resolution textures. With 32 GB, the card will never run out of VRAM at 4K Ultra, even with the most demanding add-ons.

The RTX 5090’s base clock is 2017 MHz, boosting to 2407 MHz. It has 21760 shading units, 680 TMUs, and 176 ROPs. The pixel rate is 423.6 GPixel/s, and the texture rate is 1,636.8 GTexel/s. FP32 performance is 104.8 TFLOPS, which is massive for any workload, including the physics and rendering in Microsoft Flight Simulator. The card has 170 RT cores and 680 tensor cores, meaning ray-traced shadows and AI-based upscaling are fully supported, though the measured FPS data does not indicate whether these features were enabled.

The TDP is 575 W, requiring a 950 W suggested PSU and a single 16-pin power connector. This is a power-hungry card, but the benchmark results show it delivers. The GPU’s 92nd percentile ranking among all GPUs is expected for a flagship. Its nearest rivals in synthetic benchmarks are Tesla P100 variants, which have far less memory bandwidth and no modern gaming features, so the RTX 5090’s real-world gaming lead is larger than the raw scores suggest.

The RTX 5090’s launch MSRP is 1,999 USD, which positions it as a premium product. The measured FPS data shows it is capable of high refresh rates even at 1440p, and it can handle 4K with the right settings. The GPU’s 32 GB VRAM also provides future-proofing for upcoming flight simulators and heavy modding.

Measured FPS Breakdown

At 1080p with Low settings, the combo averages 173 FPS. That is the highest frame rate measured across all resolutions and settings. Moving to Medium at 1080p, the average drops to 135 FPS, with a minimum of 114 and a maximum of 155. High settings at 1080p yield 101 FPS, and Ultra drops further to 96 FPS. The gap between Low and Ultra at 1080p is 77 FPS, or about 44% lower, which shows the GPU is being pushed harder as settings increase, but the CPU still has headroom to deliver over 100 FPS on High.

At 1440p, the pattern continues. Low settings average 139 FPS, Medium averages 109 FPS (minimum 92, maximum 125), High averages 82 FPS, and Ultra averages 78 FPS. The drop from Low to Ultra at 1440p is 61 FPS, or about 44% lower, similar to the 1080p scaling. The minimum FPS at 1440p Medium is 92, which is above the threshold for a smooth experience on a 60 Hz monitor, but below what a 144 Hz display would need for a locked frame rate.

At 4K, the frame rates fall significantly. Low settings average 103 FPS, Medium averages 80 FPS (minimum 68, maximum 92), High averages 60 FPS, and Ultra averages 57 FPS. The drop from Low to Ultra at 4K is 46 FPS, or about 45% lower. Notably, the 4K High and Ultra settings both hover near 60 FPS, which is the target for many console-like gaming experiences, but not ideal for competitive or high-refresh play.

The data shows that the RTX 5090 is the dominant factor in frame rates. At 1080p Low, the CPU may become a bottleneck, but the 173 FPS result is still excellent. At 4K Ultra, the GPU is clearly the limiting component, as the CPU has more than enough single-thread performance to feed the card at lower resolutions.

Resolution Scaling

The FPS drops from 1080p to 4K are steep, indicating that the GPU is the primary constraint at higher resolutions. At Low settings, the average FPS goes from 173 at 1080p to 139 at 1440p (a 20% drop) and then to 103 at 4K (a 40% drop from 1080p). That is a larger relative drop than what a purely CPU-bound scenario would show, confirming the GPU is doing most of the heavy lifting.

At Medium settings, the progression is 135 FPS at 1080p, 109 FPS at 1440p (a 19% drop), and 80 FPS at 4K (a 41% drop from 1080p). The minimum FPS at 1440p Medium is 92, and at 4K Medium it is 68, which means even the low end stays playable on a 60 Hz display. At High settings, 1080p yields 101 FPS, 1440p yields 82 FPS (a 19% drop), and 4K yields 60 FPS (a 41% drop from 1080p). The 4K High result is exactly 60 FPS, which is the standard for console-like smoothness.

At Ultra settings, the drop is similar: 96 FPS at 1080p, 78 FPS at 1440p (a 19% drop), and 57 FPS at 4K (a 41% drop from 1080p). The consistent 19-20% drop from 1080p to 1440p and the 41-45% drop from 1080p to 4K across all settings indicate that the GPU is scaling predictably with pixel count. The RTX 5090’s memory bandwidth of 1.79 TB/s is sufficient to feed the 4K resolution, but the raw shader throughput becomes the limiting factor at Ultra.

The CPU’s role in resolution scaling is minimal. If the 388H were the bottleneck, we would see smaller FPS drops at higher resolutions, because the CPU would cap the frame rate at 1080p as well. Instead, the 4K results are roughly half of the 1080p results, which is typical for a GPU-bound scenario. The 388H’s 5.10 GHz boost clock ensures that it can keep up with the GPU’s frame output at 1080p, but at 4K, the GPU is clearly the limiting component.

Settings Recommendations

For 4K displays, the data shows that Medium settings provide the best balance of visual quality and performance. The average of 80 FPS at 4K Medium is well above the 60 FPS target, and the minimum of 68 FPS ensures that even demanding scenes stay smooth. High settings at 4K drop to 60 FPS average, which is playable but leaves no headroom for dips. Ultra at 4K averages 57 FPS, which is below 60 and may cause noticeable stutters in dense urban areas or during weather changes. For a locked 60 FPS experience at 4K, Medium is the recommended preset.

At 1440p, High settings are the sweet spot. The average of 82 FPS at 1440p High is sufficient for a 75 Hz or 100 Hz monitor, and the GPU has enough headroom to handle complex scenery. Ultra at 1440p averages 78 FPS, which is only 5% lower than High, so if the visual fidelity of Ultra is important, it is still a viable option. Medium at 1440p averages 109 FPS, which is ideal for a 120 Hz display, but the visual quality drop from High to Medium is noticeable in a flight simulator where ground detail and cloud shadows matter. For most 1440p users, High offers the best compromise.

At 1080p, the recommendation depends on the refresh rate of the monitor. For a 144 Hz display, High settings at 101 FPS will not saturate the refresh rate, but Medium at 135 FPS will get closer, and Low at 173 FPS will exceed it. The difference between Medium and High at 1080p is 34 FPS, which is significant. If the monitor is 120 Hz, Medium is the right choice. If it is 60 Hz, Ultra at 96 FPS is overkill, and High or Medium would allow for lower GPU temperatures and power draw. The RTX 5090’s 575 W TDP means that running at 1080p Low could reduce power consumption, but the measured data does not include power figures.

The data indicates that this combo is overkill for 1080p gaming. Even at Ultra, the average of 96 FPS is more than enough for most monitors. The GPU’s 32 GB VRAM is also unnecessary at 1080p, but it provides headroom for texture mods and future titles. For best experience per the measured rows, 1440p High or 4K Medium are the recommended settings, depending on the display resolution. At 4K, avoid Ultra unless a 57 FPS average with potential dips below 50 is acceptable. At 1440p, Ultra is close to High in performance but offers slightly better visuals; the choice comes down to personal preference for frame rate stability versus image quality.