Dragon’s Dogma 2 has gotten its latest title update with tons of fixes and a lower CPU overhead in NPC-rich towns and cities. The previous patches improved CPU performance and added DLSS 3 Frame Generation to the game (broken at the time). If you’re still struggling with the game’s performance, our optimization guide is here to help. It includes an analysis of every graphics setting, including performance impact, VRAM usage, and CPU bottlenecks.
The latest update improves the game’s performance by 10-15%, primarily in densely populated settlements like Vermund, where these tests were conducted. On our GeForce RTX 4090 + Core i9-13900KF test bench, we observed gains of 7-10% in rasterization and up to 15% in ray-tracing at 1440p and 4K, respectively.
Dragon’s Dogma 2 now performs nearly 20% faster than at launch, both with and without ray tracing at all resolutions. The lows, in particular, are higher, with uplifts of up to 36% and 53% at 4K and 1440p, respectively. These are critical numbers as they can make the game playable or unplayable irrespective of the averages. They’re also an indication of CPU bottlenecks. More FPS drops imply a more severe bottleneck.
At 1440p “Ultra” (RT Off), we see a similar GPU busy deviation as before (20-30%).
4K “Max” with ray-tracing enabled gets the largest performance uplift, with the lows improving by 66% (30.7 FPS to 51 FPS) versus at launch.
The GPU busy deviation has also reduced from nearly 50% at launch to just 15% after the latest patch. Interestingly, it isn’t much lower with DLSS or FSR enabled. We’re still seeing figures of 26-30% at 4K.
Dragon’s Dogma 2 is now more performant with upscalers and frame generation. DLSS yields 50% higher minimum framerates than at launch, going from 36 FPS to 54 FPS at 4K “Max.”
Frame generation is now properly rendered without artifacts, granting a ~2x performance gain versus 4K native (at launch). The 1% lows are 64% higher than native, going from mildly stuttery to silky smooth.
Enabling frame generation with DLSS also reduces the CPU overhead by 7% as the GPU busy deviation drops from 24% to 17%.
