To a first approximation Snowmass and Redcloud use the hardware identically. With each release of the FPGA software I try to generate less power supply pollution noise in the FPGA. Tho those changes aren’t in general measurable, the blacker back ground they bring is obvious to most listeners.
The https://forum.psaudio.com/t/the-next-directstream-update/5273 thread contains very precise descriptions about what’s changed and why. In a nutshell Snowmass is a rewrite of the PCM upsampling code. But because it also lowers noise in the FPGA everything sounds better. The lower the input sampling rate the “more better.”
I’m always looking for technically better answers to questions - I don’t pay attention to “sonic ‘flavor signature’”, etc. My experience has always been that a better implementation gives better sound quality.
[Edit - I just saw your edit, but I’ll post this anyway 
]