Spent some time on different config settings, the difference is minimal if you are targeting 60 fps. In my case the limiting factor is CPU, because FSR 2.0 is a software upscaling solution thus it increases the CPU load by 10%-15%. When CPU load is above 70% it limits the performance of the graphics card, so the best my PC can do is 40-45 fps in hordes. With a better CPU and FSR 2.0 set to performance or balanced it's possible to maintain 50-60 fps in hordes with GTX 1650. It seems pointless to lower the graphics setting beyond certain level if the bottleneck is CPU, even with a potato looking config I still can't get a stable 60 fps in mission (CPU load hits 80% in hordes limits the GPU to about 60% producing ~35 fps).
Disabling FSR? Can't even get 60 fps to start with.
Edit: Tested with a very low setting config found out few things: 1, Dynamic shadows (sun shadows, local light shadows) has the most impact on performance, about 10% each. Static shadows have little performance impact, and disabling it completely break the lighting everything looks way too bright. Volumetric lighting settings do not have much performance impact if the volumetric data size is optimal. 2, Without the using of FSR the memory usage seems a bit higher, had 2 out of memory crashes with a potato config not using FSR with 16g RAM, but it does reach 60 fps though ( still drops to ~40 fps in hordes that's why I think it's kind of pointless to use a potato config if the CPU load is so high in hordes and cause the frame drop ). Even just loading into the mission the CPU load is already 50-60 percent there's really not much I can do. Essentially you have almost the same performance if lock to 60 fps, a potato looking config would be almost stable 60 fps out of hordes while the good looking config have few frame drops here and there but the overall experience is smooth enough to be playable. I personally do care about how the game looks and I don't enjoy playing with a potato config.
and also, is it just me? but whenever I try using Intel XeSS, the lag and stutter gets way worse than when I enabled any upscaling to begin with, this doesnt happen to neither of the FSR tho. I use an RX580 so maybe the fact that it is an old outdated GPU is the reason why? idk just putting this here
FSR 1.0 is much blurrier and more pixelated than FSR 2.0 in Darktide. FSR 2.0 even in Ultra Performance mode does not look that bad while FSR 1.0 performance mode already looks very bad, I didn't notice much FPS difference between those two but the visual difference is massive.
I tried Intel XeSS too, the result was very dissapointing, I don't see any reason to use XeSS over FSR 2.0. As far as I can tell XeSS seems not trying to achieve the same goal as DLSS and FSR.
6 comments
Disabling FSR? Can't even get 60 fps to start with.Edit:
Tested with a very low setting config found out few things:
1, Dynamic shadows (sun shadows, local light shadows) has the most impact on performance, about 10% each. Static shadows have little performance impact, and disabling it completely break the lighting everything looks way too bright. Volumetric lighting settings do not have much performance impact if the volumetric data size is optimal.
2, Without the using of FSR the memory usage seems a bit higher, had 2 out of memory crashes with a potato config not using FSR with 16g RAM, but it does reach 60 fps though ( still drops to ~40 fps in hordes that's why I think it's kind of pointless to use a potato config if the CPU load is so high in hordes and cause the frame drop ). Even just loading into the mission the CPU load is already 50-60 percent there's really not much I can do.
Essentially you have almost the same performance if lock to 60 fps, a potato looking config would be almost stable 60 fps out of hordes while the good looking config have few frame drops here and there but the overall experience is smooth enough to be playable. I personally do care about how the game looks and I don't enjoy playing with a potato config.
and also, is it just me? but whenever I try using Intel XeSS, the lag and stutter gets way worse than when I enabled any upscaling to begin with, this doesnt happen to neither of the FSR tho. I use an RX580 so maybe the fact that it is an old outdated GPU is the reason why? idk just putting this here
I tried Intel XeSS too, the result was very dissapointing, I don't see any reason to use XeSS over FSR 2.0. As far as I can tell XeSS seems not trying to achieve the same goal as DLSS and FSR.