Memory model v2

[nkaskov]

The current memory model version allowed us to implement multiple complex algorithms. Now, we need to upgrade it according to our needs and make it more suitable for usage in the following use-cases:

• zkEVM1
• Placeholder verifier
• Logic proof checker

Suggestions on the plan how to resolve this issue:

1. Define use-cases needs (which use-case from the list below requires which features to be added to the compiler):

• https://github.com/NilFoundation/zkLLVM/issues/328
• https://github.com/NilFoundation/zkLLVM/issues/357
• https://github.com/NilFoundation/zkLLVM/issues/337
• https://github.com/NilFoundation/zkLLVM/issues/138

2. Planning and brainstorming sessions with the research team. Once finished - could you add issues in the corresponding repositories?

• [ ] Define ways to do the required functionality on the compiler side.
• [ ] Which proof system features must be added to support this functionality.
• [ ] Which blueprint components must be implemented to support this functionality.

3. Implementing the issues defined in the previous step.

[aleasims]

During experiments with new memory model, I found out, that is will be very useful to know the non-native field type size at compile time. So this task is required:

• #517

[aleasims]

RFC: Assigner memory model v2

[nil-foundation-cicd-bot[bot]]

PRs, synced with the current issue:

• NilFoundation/zkLLVM#554 (SHA: c4d002f)
• NilFoundation/zkllvm-assigner#199 (SHA: 316a765)

Note: Please do not edit this comment; it's generated automatically.