feat(ssa): Deduplicate intrinsics with predicates

[aakoshh]

Description

Problem*

Resolves #6533

Summary*

Allows Instruction::Intrinsic types to be deduplicated in constant_folding if the Context is in use_constraint_info mode, and the Intrinsic is one that doesn't have a side effect other than the possibility of violating an implicit constraint on its inputs, which should be preserved in the first occurrence.

Additional Context

I found it a bit confusing why we're using use_constraint_info for these whenrequires_acir_gen_predicate returns false for all these instructions, which will cause cached_instruction_results to contain None for the predicate regardless of use_constraint_info. For some, but not all intrinsics, the predicate that tells whether the side effect is enabled is attached to their inputs during Ssa::flatten_cfg, and they don't depend any more on the enable_side_effect instruction. But it seems like this flag is the one that communicates whether we want such instructions to be deduplicated, so I used it even though I'm not sure what adverse effect we would see if we just returned true for these, other than the implicit coupling by knowing that flatten_cfg has done things in a certain way.

Another method that seems to have some of the truth is remove_enable_side_effects::Context::responds_to_side_effects_var

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]

Changes to number of Brillig opcodes executed

Generated at commit: d2c6232bdbb19ab72036f695e3884ca49cabb682, compared to commit: 1e965bc8b9c4222c7b2ad7502df415781308de7f

🧾 Summary (10% most significant diffs)

Program	Brillig opcodes (+/-)	%
slice_regex	+44 ❌	+1.30%
merkle_insert	-33 ✅	-0.87%

---

Full diff report 👇

Program	Brillig opcodes (+/-)	%
slice_regex	3,434 (+44)	+1.30%
sha256_regression	116,113 (-6)	-0.01%
merkle_insert	3,757 (-33)	-0.87%

[github-actions]

Changes to Brillig bytecode sizes

Generated at commit: d2c6232bdbb19ab72036f695e3884ca49cabb682, compared to commit: 1e965bc8b9c4222c7b2ad7502df415781308de7f

🧾 Summary (10% most significant diffs)

Program	Brillig opcodes (+/-)	%
merkle_insert	-11 ✅	-1.42%
to_le_bytes	-10 ✅	-7.58%

---

Full diff report 👇

Program	Brillig opcodes (+/-)	%
sha256_regression	6,523 (-6)	-0.09%
merkle_insert	764 (-11)	-1.42%
to_le_bytes	122 (-10)	-7.58%

[github-actions]

Changes to circuit sizes

Generated at commit: d2c6232bdbb19ab72036f695e3884ca49cabb682, compared to commit: 1e965bc8b9c4222c7b2ad7502df415781308de7f

🧾 Summary (10% most significant diffs)

Program	ACIR opcodes (+/-)	%	Circuit size (+/-)	%
u128	-44 ✅	-6.72%	-61 ✅	-1.31%
bit_shifts_runtime	-118 ✅	-13.32%	-141 ✅	-3.00%

---

Full diff report 👇

Program	ACIR opcodes (+/-)	%	Circuit size (+/-)	%
merkle_insert	46 (-8)	-14.81%	3,677 (-7)	-0.19%
u128	611 (-44)	-6.72%	4,604 (-61)	-1.31%
bit_shifts_runtime	768 (-118)	-13.32%	4,557 (-141)	-3.00%