Docs: Nexus Proof Compression

[danielmarinq]

We need a full description of the Nexus Proof Compression sequence:

• Architecture
• Theoretical prover costs
• Concrete prover costs
• Links to the sections in the code with the implementations
• Relevant diagrams that showcase the architecture

[dorebell]

there are two sets of documentation for the compression feature at this point:

the blog post and the CLI docs.

the former gives important context for the problem and addresses the various technical challenges and our solutions to them. the latter gives a practical guide for using the system, including some discussion of its performance.

right now, the compression features are only running successfully on machines with greater than 64GB of RAM (128GB works, 64 doesn't: the reported "maximum resident set size" is 58GB), and this is when k = 1. The memory usage and SRS size scale piecewise-constant and asymptotically linearly in k: they double whenever the total nova circuit size, measured in terms of the number of non-zero entries in the r1cs matrices, crosses a power of 2. these numbers should improve with further development, by an expected factor of ~4.

i'm not sure what else we want to explain or where it should live.