feat(testing): Add SSA interpreter for testing SSA

[jfecher]

Description

Problem*

Resolves #6826

Summary*

Implements an SSA interpreter purely for debugging purposes.

Additional Context

Documentation*

Check one:

• No documentation needed.
• Documentation included in this PR.
• [For Experimental Features] Documentation to be submitted in a separate PR.

PR Checklist*

• I have tested the changes locally.
• I have formatted the changes with Prettier and/or cargo fmt on default settings.

[jfecher]

The interpreter is currently in a very incomplete yet semi-working state so I thought I'd put this up to see if anyone had any comments or suggestions.

Going forward I think most of the remaining work will be:

• Implementing each builtin
• Threading errors throughout the interpreter instead of panics
• Continuing to write more tests
• Ways to invoke the interpreter from the CLI instead of just via unit tests.
  • Run each test_program against the ssa interpreter automatically and ensure values match?
  • CLI option to run the interpreter after each SSA pass to catch if a pass ever changes the result of the program.
  • Debugger-like options for the interpreter? E.g. "give me the value of v203." This one at least is out of scope for this PR.

[github-actions]

⚠️ Performance Alert ⚠️

Possible performance regression was detected for benchmark 'Execution Time'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.20.

Benchmark suite	Current: 2872cec	Previous: 934126b	Ratio
rollup-merge	0.004 s	0.003 s	1.33

This comment was automatically generated by workflow using github-action-benchmark.

CC: @TomAFrench

[jfecher]

This PR should be ready for review now. Some things to note:

• ~2.3k lines of this PR are just tests in ssa/interpreter/tests/mod.rs and ssa/interpreter/tests/instructions.rs.
• The interpreter currently still panics instead of returning error variants on error. We should update this later but I'm considering this ok for an MVP for testing.
• A few intrinsics are still not implemented but again I'd rather leave them out of the MVP. These are:
  • Intrinsic::ApplyRangeConstraint we could just apply a range constraint but there's no documentation on any of the intrinsics so I'm not sure argument ordering, etc.
  • Intrinsic::BlackBox(_) all black box functions are unimplemented
  • Intrinsic::Hint(_) unsure what to do here
  • DerivePedersenGenerators this one was just a lot of code
  • We should implement these in follow-ups. I'll make an issue.
• There are some necessary differences between this and SSA. Notably assert_constant is always true. I had it failing before on non-constant ValueIds but this meant we'd fail the same SSA on different optimization levels which is undesirable. static_assert is similar but just functions exactly like assert now.
• The interpreter can still be used for unit testing and can be easily fed inputs and outputs. Constrains in the code are still asserted, etc.
• I removed the CLI code from this PR, we can add this in a later PR so that it can be invoked via e.g. nargo interpret-ssa. Embedding it in other commands like nargo execute isn't ideal since we'd need to pass the program arguments down to noirc_evaluator.

[github-actions]

⚠️ Performance Alert ⚠️

Possible performance regression was detected for benchmark 'Test Suite Duration'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.20.

Benchmark suite	Current: 0029fef	Previous: 4bc636a	Ratio
test_report_zkpassport_noir_rsa_	30 s	22 s	1.36

This comment was automatically generated by workflow using github-action-benchmark.

CC: @TomAFrench

[aakoshh]

Nice, looking forward to finding ways to use this with the fuzzer as an additional pillar of oracles!

[aakoshh]

Awesome :)

This is going to be be a great reference in itself for what the behaviour should be 👍