TL;DR a RTX 2080 yields about 30% performance improvement over a GTX1070 at ~2K (2560x1600) resolution. On average 55FPS from a GTX1070 and 72FPS from an RTX 2080. I overclocked the 2080 and wrung out another 4FPS for about 76FPS average, or a 38% performance improvement over a GTX1070. One thing I know about Star Citizen is that it will use whatever you throw at it. More CPU? Yes, please. Oh, more GPU? Sure, I'll use that too. RAM? Yes, and in increasing amounts up to your whole amount and beyond. Until last night, I've been testing with a 8700K and GTX1070 combo pushing 2560x1600, 2560x1440, and 1920x1200 to a Dell U3011 16:10 monitor. Not ideal because of the sedate refresh rate and the UI doesn't work well with 16:10 but the U3011's IPS is gorgeous. I use "Fast Vsynch" set in the nVidia control panel 3d Settings to throw frames over 60 and eliminate screen tearing. The difficulty I was having, was that the GTX1070 was just not quite able to push 60FPS at 2560x1600, no matter the quality settings--Low was not setting me free. At 1920x1200, I could achieve 60 FPS pretty regularly but with obvious compromise to visual quality--the effect was like being a bit blurry-eyed, or farsighted. Text was smudged at distances. So, I got a RTX 2080. I suppose I would have gotten a GTX1080Ti--the obvious advantage of having 11GB of VRAM was not lost on me and from what I have read its performance seems quite comparable to the RTX2080. Plus, there's been some hype and hokum from nVidia about RTX that I didn't like and some egregious shilling about it from some quarters. But plain market economics don't lie--the 1080 Ti is more costly. So, I got an RTX 2080. Before I installed it, I decided to take the GTX1070 on one last tour of 3.3.5, because I wanted to record benchmarks. I took a Super Hornet to Delamar, landed at Levski, went to Grand Barter, then skipped to Hurston, landed at Lorville, and looked out over the hangars from the window near the ship deck at Teasa Spaceport. At various points I would just look up at the FPS counter and write down the figure I saw, then change resolutions and see what the effect was. What I think I've discovered, is that with a GTX1070, I was limited by my GPU--the 8700K had more to give, but the 1070 just couldn't keep up. With an RTX2080, I am limited by my CPU--the GPU isn't fully utilized and the CPU is giving it all it has. Here's my work: