Just wanted to share with you the results from a non scientific benchmark I did yesterday (althrough I tried to control the parameters as much as possible)
I have upgraded from a 2080 TI MSI Gaming X Trio to an Asus TUF OC RTX 3090.
Ran some comparisons with both GPUs OCed. (MSI +162 clock / +1000 memory stock fan curve // ASUS +180 clock / + 1100 memory agressive fan curve). No tweak to voltage. Open case and room temperatura around 23C.
Tested at 4 airports cockpit view, external view, with broken clouds, clear skies day and night (only internal cockpit view for night). The aircraft was a cj4 and positioned on the runway.
Ultra settings @4k 100% Render Scaling.
The average fps increase was 26% but at 3 airports it was 30%. LAX lowered the average to 26% probably due to CPU bottleneck.
My system:
10900k stock
32gb hyper X @ 3733mhz
Firecuda 520 2tb
Corsair H115 watercooler
Asus rog strix z490 gaming-f motherboard
2080 ti overclocks better than the 3090 and that is why the difference isn’t bigger.
Timespy graphic scores are around 16000 for the 2080 TI and around 20400 for the RTX 3090.
More or less what I was expecting… the higher end 2080 TIs are really good as can be seen.
It probably won’t change, MSFS never actually uses more than around 8GBs at 4K in my experience. It’ll allocate more most likely, but if you look in dev mode, it won’t really use more
Flight sim happily eats 13GB of VRAM, running @ 4k with a 3090, various, mix of high/medium/ultra.
Also coming from a 2080ti, main thing ive noticed is instead of running 80% scale most of the time I can now run native 4k and still have a better frame rate.
For now around 8,6GBs VRAM measured using developer mode. Keep in mind that I didn’t fly. The benchmarks were done with the aircraft halted on the runway.
