Any significant limitation from RTX 30xx series? nvidia compute capability

r/StableDiffusion
AI Hardware

According to nvidia the RTX 30xx series have 8.6 compute capability. I just wanted to know if there are any hardware limitations that impact model inference and