Fortnite

Fortnite with this pairing is a perfectly matched oddity: a six-core, twelve-thread Zen 3 CPU (base 3.60 GHz, boost 4.20 GHz, 16 MB L3, 65 W TDP, PCIe Gen 3) paired with a flagship Blackwell GPU that carries 32 GB of GDDR7 memory, a 512-bit bus, and 1.79 TB/s of peak bandwidth. Ingame the combo lands at rank 158 out of 3700 measured combos, which places it in the upper tier of the database despite the CPU's modest standing. The processor itself sits in the 73rd percentile of all CPUs, while the GPU in the 92nd percentile, and the measured frame data shows exactly how each component's role changes as resolution and settings climb. This is the sort of data you'd want if you were planning a high-refresh competitive setup while questioning whether an older AM4 processor can still do the job.

How This Combo Ranks

The measured rank for this specific CPU+GPU pair in Fortnite is 158 out of 3700 tested combos. Given that the GPU is in the 92nd percentile of all GPUs and the CPU in the 63rd percentile, a large chunk of that ranking is clearly earned by the RTX 5090's raw rendering power. Still, the rank is a statement about the total system: it outperforms the large majority of tested pairs, including those using faster CPUs with slower GPUs.

The CPU itself is not a top-flight part, but it sits within a hair of its immediate rivals. Its average benchmark score is 19593. The AMD Ryzen AI 5 430 scores 19617, a −0.1% deviation; the Intel Core i5-12500 scores 19668, −0.4%; the Intel Core i7-10700F scores 19499 with a +0.5% deviation; and the Intel Core i5-11600KF scores 19723, −0.7%. That scatter is tiny – under a single percentage point in every direction – which makes the CPU an approximate neutral factor in the ranking compared to those alternatives.

On the GPU side, the field is similarly tight among the nearest scored rivals in the database. The RTX 5090's average score is 79842. The NVIDIA Tesla P100 PCIe 16 GB scores 79605, a mere 0.3% below that reference; the other Tesla P100 with 12 GB scores 79396 (0.6% low); an AMD Radeon RX 6850M XT scores 78940 (1.1% low); and a Radeon Pro Vega 64X hits 80254, only 1.4% higher than the RTX 5090's average. These differences are much smaller than the differences in real‑game frame rates you'll see below, which is the crucial takeaway: the GPU's raw compute average is far less meaningful than what it does in the actual game engine.

The fact that the combo sits at rank 158, instead of inside the top 50, is most likely a artifact of the CPU bottleneck in certain scenes. At 1080p the game can push past 350 frames per second, but those high numbers only confirm that the GPU is nowhere near its limit at that resolution. Once you move to 4K and enable Medium/High/Ultra, the CPU constraint is reduced, and the GPU has to actually do the heavy lifting. So the ranking reflects a balance: a CPU that can keep up with feeding frames up to around 170–300 fps, and a GPU that just wants more resolution and more settings among all the fastest options.

GPU Role

The GPU in this combo is the RTX 5090, presumably the head of the Blackwell 20 series in terms of raw specs: base clock 2017 MHz, boost 2407 MHz, memory clock 1750 MHz with 28 Gbps effective, and a 512‑bit bus feeding a 32 GB GDDR7 buffer. Peak bandwidth is 1.79 TB/s, and the card carries 21760 shade units, 680 texture mapping units, 176 ROPS, 170 RT cores, and 680 tensor cores. FP32 throughput is 104.8 TFLOPS; the pixel rate is 423.6 GPixels/s, the texture of rate is 1,636.8 GTixels/s. All of those are irrelevant on paper, but for Fortnite they matter in a very specific way.

In the game's low settings profile, the GPU becomes essentially free. The measured FPS at 1080p Low is 367 frames per second, 311 at 1440p Low, 124 at 4K Low. Those are all very high numbers, but the most information is hidden in the difference between Low and Medium: at 1080p, Low gives 367 fps and Medium gives 322 fps (minimum 274 – 370 maximum). That very small escalation between Low and Medium reveals that the GPU is barely under load; the CPU is doing the work, and the game's draw calls and physics still limit the framerate. The card's massive memory bandwidth – 1.79 TB/s – is not the killer feature here; the card can handle far more textured work, which is why at Medium 4K the framerate is 116, versus 110 for Low.

The card's 32 GB virtual frame buffer is far larger than anything this game needs at any measured setting. Because all the measured data is captured on actual hardware, the buffered memory values are not shown, but the fact that the data reservoir is 2–3 times the render size at 4K is not an issue. The card's 170 RT cores provide hard RT capability but Fortnite usually uses traditional rasterization; the measurements don't include RT toggles, so there is no data to distinguish RT settings. The reality is that the GPU is nowhere near capability to the cards in this title – the SCORES move because of geometry and shading, not because of memory size.

Settings Recommendations

Using the measured FPS rows, the sweet spot for this mixed setup is the Medium preset at 4K and High at 1440p. But that has to be examined resolution by resolution.

At 4K, the stats are: Low 164 fps, Medium 116 fps (98 min / 133 max), High 90 fps, Ultra 59 fps. For a 60–120 Hz panel, the Medium format offers 116 average, which holds above 60 even on frame-drop worst case (98 min). For a 120 Hz screen, Medium gets you only 116 average – borderline; High gives 90 and is not enough to sustain 120, but for 90 Hz content it's fine. Ultra at 59 fps exactly saturates a 60 Hz monitor but leaves no headroom. To maximize smoothness and still hit quality, Medium is the best for the measured set. Low at 164 fps is excellent for e‑sports players with 165 Hz displays, because it by part.

At 1440p, Medium is already 219 fps (186 min / 252 max), High is 217 fps. This is a pleasant spot: 170fps average with High and 220 fps with Medium. For a 144 Hz panel, High or even Ultra (112 fps) can be used, while the Low at 231 fps is there if you want absolute performance. High appears to be the best balance because it keeps average 170 and goes down to 186 in worst-case, which is above 144 Hz. If you'd rather max graphics, Ultra is a decent minus one (112).

At 1080p, the measurements are: Low 367, Medium 322 (min 274–max370), High 267, Ultra 176. For a 240 fps display, High at 267 and Ultra at 176 are both playable but High is the skill ceiling. Medium at 322 exceeds 240, and even the min value of 274 keeps over 240. So Medium is a good fit for 240Hz monitors. High's max is 267 – not enough to max a 240Hz behavior but the model provides an excellent smooth experience. Honestly, at this resolution, even Ultra gives 176 fps for a slow 144Hz, so the user has freedom.

The overall recommendation: for a high‑refresh 1440p screen, pick High. For a 4K monitor, choose between Medium (if you want 116 fps) or High (90) when you prefer quality over speed. 1080p has enough headroom that any setting is playable, but Medium is the best per measured fps per frame.

Measured FPS breakdown

Here are the exact numbers measured for each resolution and preset.

1920×1080 (1080p)

• Low: 367 avg
• Medium: 322 avg (274 min / 370 max)
• High: 267 avg
• Ultra: 176 avg

The range from Low to High is 367 → 322 → 267 → 176 init, with Ultra cutting More than a third of the frame from High. The intermediate min/max for Medium is a min of 274 and a max of 376, which is the hardest level for the CPU.

2560×1440 (1440p)

• Low: 311 avg
• Medium: 219 avg (186 min / 252 max)
• High: 170 avg
• Ultra: 112 avg

Notice the pattern: Low at 311, Medium at 219, High at 170, Ultra at 112. The drop from Low to Medium is 92 fps, from Medium to High is 49, and from High to Ultra is 58. The minimum for Medium in 1440p is 186, which still beats a 144 Hz refresh rate.

3840×2160 (4K)

• Low: 164 avg
• Medium: 116 avg (98 min / 133 max)
• High: 90 avg
• Ultra: 59 avg

At 4K, the scaling gets even more severe: Low 164, Medium 116, High 90, Ultra 59. The Medium worst case is 98 fps, so even a 100