feat: Blake3 utility circuit

[storojs72]

This PR includes several utility chips for Blake3 hashing stuff:

• u32_from_le_bytes (for correct hash value outputting)
• u32_xor (for flags computing)

[arthurpaulino]

Thanks for the implementations!
The thing that's bugging me at the moment is the replication of data across preprocessed traces from different circuits.
For example, the pair of columns accounting for all ~65k combinations of all bytes seems to be a repeating theme. We'll need to figure out a better way to structure these circuits so that we can avoid this issue.

[storojs72]

Yeah, agree. But this is a bit out of scope of multi-stark. It is in scope of Aiur. The idea is actually having multi-stark chips at first, which contain implementation of lookups, so we don't need to keep in mind this, when implementing particular Aiur operation, since every operation involves Rust, FFI and Lean components at once.

[storojs72]

Also added u64 to u32 conversion chips (using 32 bits right / left shifting). This is necessary for computing counter_low / counter_high in Blake3 context

[storojs72]

Added 1-bit shift right of u64 value chip. It is used in chunks processing of Blake3