chore(acir): Test whether the predicate has an effect on slice intrinsics

[vezenovm]

Description

Problem*

Resolves #8355 (comment)

Summary*

So although this resolves the comment on #8355, this functionality can be tested in isolation without considering that PR at all. The following was done to determine whether a predicate affected a certain instruction:

1. Setup an SSA where there is only instruction after EnableSideEffectIf instruction on a dynamic predicate
2. Check the ACIR and final witness index for an SSA where that EnableSideEffectIf is in place and an SSA where there is no EnableSideEffectIf.
3. If the instruction is meant to be unaffected by the predicate, check for equality. If the predicate is meant to have an affect, check for inequality.

For example we may have this SSA:

acir(inline) predicate_pure fn main f0 {
  b0(v0: u32, v1: u1):
    v4 = make_array [Field 2, Field 3] : [Field; 2]
    v6 = make_array [Field 1, v4] : [(Field, [Field; 2])]
    v8 = array_set v6, index v0, value Field 4
    enable_side_effects v1
    v11, v12 = call slice_push_back(u32 1, v8, Field 1, v4) -> (u32, [(Field, [Field; 2])])
    return
}

and we want to check that the ACIR generated is equivalent to this SSA:

acir(inline) predicate_pure fn main f0 {
  b0(v0: u32, v1: u1):
    v4 = make_array [Field 2, Field 3] : [Field; 2]
    v6 = make_array [Field 1, v4] : [(Field, [Field; 2])]
    v8 = array_set v6, index v0, value Field 4
    v11, v12 = call slice_push_back(u32 1, v8, Field 1, v4) -> (u32, [(Field, [Field; 2])])
    return
}

Composite types were used for setting up the slice tests. This was necessary to guarantee going that slice_insert and slice_remove were affected by the predicate.

Additional Context

In general, I think this strategy of testing could be an effective simple way to test requires_acir_gen_predicate is in fact marked appropriately. We could expand it to the remaining intrinsic calls and instructions. If after a review we decide to go forward with these kinds of tests, I am happy to capture an issue for adding more ACIR gen predicate tests for the remaining instructions.

The helper in this PR is specific to setting up slice intrinsic calls, however, it could definitely be expand to support all intrinsic calls, and even all instructions (perhaps we will want a new helper at that point).

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.

[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: d6847a3	Previous: 8ab59da	Ratio
test_report_noir-lang_noir_bigcurve_	309 s	230 s	1.34

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

CC: @TomAFrench

[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: a3b8d24	Previous: 9dc9cf6	Ratio
rollup-merge	0.004 s	0.003 s	1.33

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

CC: @TomAFrench