Just a quick heads up:
Never do performance-comparisons “in editor”, or in development builds, since this is not representative of the actual performance. (Create a shipping build, and compare actual fps.)
If Epic added some debugging stuff in the background etc. this can and will have an impact on in-editor performance.
For example, my project currently performs almost twice as good in shipping builds than development, and editor isnt even something to consider.
From my experience, 5.2 performs pretty much the same as 5.0 - at the same settings. (but some settings changed, which is why it looks different etc.)
EDIT:
Actually, I can show that your comparison doesnt work.
I marked green what is faster than 5.1 and red what is slower:
All of this lies within “margin of error” within editor (and the results you get can change a lot from frame to frame too…), especially if the frame is that different.
Here is the exact same scene, without moving the camera or anything, in 5.2, but I asked the Engine twice to give me the performance values.
As you can see, there are large differences between each request, despite nothing changing:
Even IF everything would perform equally between 5.1 and 5.2, you still would get random differences each time you tell the engine to make that list.
The fact that lumen runs as good as it does is kind of a “miracle”… my game runs at (barely) 30fps on a (fast) 1060 with it enabled (but some other settings tuned down), native, 1080p… and your 3060 is roughly 2x as fast as a 1060…
CPU: Ryzen 3700X
GPU: RX 6900 XT
Keep in mind, I am running 1440p in my screenshots, my numbers are not directly comparable to yours - but your card can run my game at roughly 60 fps, with lumen enabled, probably more. (1080p, native)
But my games have a “fallback mode” (aka: when Lumen is off), which makes them look like they would have in UE4 - so I dont really care that much if I get 55 or 65 fps on a 3060, or about older hardware.