feat: Deduplicate to common ancestor

[jfecher]

Description

Problem*

This implements the alternate version mentioned in #6517 where we only hoist if there are multiple ValueIds. This means we do not hoist the loop example given there since there is no usage of x + 1 before the loop.

Summary*

Improves the instruction deduplication optimization by moving the instruction to a common ancestor of both blocks if neither dominate the other.

Additional Context

Testing this to see how much of an improvement it gives. The hoisting doesn't occur recursively (so multiple instructions depending on eachother will only have the first hoisted) and relies on another constant folding pass to change the earlier instructions to the hoisted version so performance can still be improved.

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 Brillig bytecode sizes

Generated at commit: d756449fa167486e5f7f0989f39fa64ce6e7ae6c, compared to commit: 852c87ae9ecdd441ee4c2ab3e78e86b2da07d8a4

🧾 Summary (10% most significant diffs)

Program	Brillig opcodes (+/-)	%
binary_operator_overloading	-9 ✅	-2.28%
conditional_regression_661	-3 ✅	-2.31%
break_and_continue	-5 ✅	-8.20%
regression_3607	-5 ✅	-10.20%

---

Full diff report 👇

Program	Brillig opcodes (+/-)	%
slice_dynamic_index	2,744 (+8)	+0.29%
hashmap	26,648 (-8)	-0.03%
uhashmap	16,302 (-8)	-0.05%
eddsa	10,876 (-18)	-0.17%
brillig_slices	531 (-1)	-0.19%
poseidon_bn254_hash_width_3	5,680 (-12)	-0.21%
poseidon_bn254_hash	5,680 (-12)	-0.21%
slices	2,258 (-5)	-0.22%
regression_5252	4,867 (-12)	-0.25%
poseidonsponge_x5_254	4,490 (-12)	-0.27%
sha2_byte	3,373 (-15)	-0.44%
tuple_inputs	375 (-2)	-0.53%
regression	724 (-5)	-0.69%
nested_array_in_slice	1,192 (-10)	-0.83%
aes128_encrypt	522 (-5)	-0.95%
references	286 (-4)	-1.38%
nested_array_dynamic	2,145 (-36)	-1.65%
bigint	2,045 (-45)	-2.15%
binary_operator_overloading	386 (-9)	-2.28%
conditional_regression_661	127 (-3)	-2.31%
break_and_continue	56 (-5)	-8.20%
regression_3607	44 (-5)	-10.20%

[github-actions]

Changes to number of Brillig opcodes executed

Generated at commit: d756449fa167486e5f7f0989f39fa64ce6e7ae6c, compared to commit: 852c87ae9ecdd441ee4c2ab3e78e86b2da07d8a4

🧾 Summary (10% most significant diffs)

Program	Brillig opcodes (+/-)	%
break_and_continue	+3 ❌	+2.70%
slice_dynamic_index	+53 ❌	+1.22%
references	-20 ✅	-4.87%

---

Full diff report 👇

Program	Brillig opcodes (+/-)	%
break_and_continue	114 (+3)	+2.70%
slice_dynamic_index	4,408 (+53)	+1.22%
conditional_1	5,896 (+36)	+0.61%
regression_5252	955,406 (+4,200)	+0.44%
poseidonsponge_x5_254	191,757 (+840)	+0.44%
poseidon_bn254_hash	170,131 (+579)	+0.34%
poseidon_bn254_hash_width_3	170,131 (+579)	+0.34%
eddsa	739,151 (+1,872)	+0.25%
brillig_slices	712 (+1)	+0.14%
hashmap	58,469 (+48)	+0.08%
slices	3,832 (+2)	+0.05%
nested_array_dynamic	3,117 (+1)	+0.03%
uhashmap	153,666 (+48)	+0.03%
tuple_inputs	636 (-1)	-0.16%
references	391 (-20)	-4.87%