chore: revert #6685

[TomAFrench]

Description

Problem*

Resolves

Summary*

This PR reverts the commit which contains #6685 due to the correctness issues identified in #6763

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.

[github-actions]

Changes to Brillig bytecode sizes

Generated at commit: d2816639721b972fde9772cc08e9b8fb5788bd17, compared to commit: e97339733000d19163372c76b299333c20db0299

🧾 Summary (10% most significant diffs)

Program	Brillig opcodes (+/-)	%
bench_2_to_17	+62 ❌	+19.08%
poseidon2	+59 ❌	+17.77%
fold_2_to_17	+98 ❌	+16.87%
brillig_rc_regression_6123	+26 ❌	+14.53%
no_predicates_numeric_generic_poseidon	+102 ❌	+13.47%
fold_numeric_generic_poseidon	+102 ❌	+13.47%
keccak256	-368 ✅	-16.75%
slices	-541 ✅	-22.81%
reference_counts	-170 ✅	-32.08%

---

Full diff report 👇

Program	Brillig opcodes (+/-)	%
bench_2_to_17	387 (+62)	+19.08%
poseidon2	391 (+59)	+17.77%
fold_2_to_17	679 (+98)	+16.87%
brillig_rc_regression_6123	205 (+26)	+14.53%
no_predicates_numeric_generic_poseidon	859 (+102)	+13.47%
fold_numeric_generic_poseidon	859 (+102)	+13.47%
regression_6674_3	671 (+68)	+11.28%
wildcard_type	354 (+26)	+7.93%
regression_6674_1	272 (+16)	+6.25%
regression_6674_2	278 (+16)	+6.11%
array_sort	310 (+14)	+4.73%
nested_array_dynamic_simple	140 (+6)	+4.48%
brillig_constant_reference_regression	90 (+3)	+3.45%
conditional_2	119 (+3)	+2.59%
regression_4202	125 (+3)	+2.46%
conditional_regression_661	130 (+3)	+2.36%
regression_struct_array_conditional	620 (+14)	+2.31%
loop_invariant_regression	140 (+3)	+2.19%
references	157 (+3)	+1.95%
fold_call_witness_condition	114 (+2)	+1.79%
fold_complex_outputs	639 (+11)	+1.75%
brillig_cow_assign	140 (+2)	+1.45%
uhashmap	14,378 (+200)	+1.41%
pedersen_hash	375 (+5)	+1.35%
array_dynamic_nested_blackbox_input	939 (+11)	+1.19%
nested_dyn_array_regression_5782	177 (+2)	+1.14%
slice_loop	281 (+3)	+1.08%
array_dynamic_blackbox_input	1,055 (+11)	+1.05%
pedersen_commitment	211 (+2)	+0.96%
databus_two_calldata	213 (+2)	+0.95%
to_be_bytes	214 (+2)	+0.94%
conditional_1	1,217 (+11)	+0.91%
ecdsa_secp256k1	934 (+8)	+0.86%
brillig_nested_arrays	383 (+3)	+0.79%
array_dynamic	321 (+2)	+0.63%
regression_4709	134,511 (+771)	+0.58%
regression	959 (+5)	+0.52%
modulus	1,774 (+6)	+0.34%
slice_dynamic_index	2,524 (+2)	+0.08%
sha2_byte	2,786 (+1)	+0.04%
brillig_cow_regression	2,204 (-2)	-0.09%
pedersen_check	573 (-1)	-0.17%
brillig_pedersen	573 (-1)	-0.17%
nested_array_in_slice	1,418 (-3)	-0.21%
nested_array_dynamic	2,771 (-7)	-0.25%
brillig_cow	386 (-1)	-0.26%
u128	2,807 (-10)	-0.35%
regression_5252	4,708 (-18)	-0.38%
poseidon_bn254_hash	5,483 (-22)	-0.40%
poseidon_bn254_hash_width_3	5,483 (-22)	-0.40%
higher_order_functions	692 (-3)	-0.43%
hashmap	21,743 (-156)	-0.71%
array_if_cond_simple	131 (-1)	-0.76%
sha256_brillig_performance_regression	1,708 (-14)	-0.81%
brillig_arrays	114 (-1)	-0.87%
regression_mem_op_predicate	109 (-1)	-0.91%
struct_inputs	276 (-3)	-1.08%
array_dynamic_main_output	91 (-1)	-1.09%
debug_logs	5,085 (-75)	-1.45%
poseidonsponge_x5_254	4,283 (-70)	-1.61%
slice_regex	2,566 (-63)	-2.40%
simple_shield	892 (-22)	-2.41%
brillig_oracle	361 (-9)	-2.43%
conditional_regression_short_circuit	1,268 (-32)	-2.46%
6	1,192 (-35)	-2.85%
array_to_slice	712 (-21)	-2.86%
strings	914 (-27)	-2.87%
sha256_var_witness_const_regression	1,295 (-43)	-3.21%
aes128_encrypt	535 (-20)	-3.60%
sha256_var_size_regression	1,800 (-68)	-3.64%
merkle_insert	776 (-32)	-3.96%
sha256_regression	6,733 (-280)	-3.99%
sha256	2,348 (-132)	-5.32%
sha256_var_padding_regression	4,927 (-283)	-5.43%
7_function	546 (-36)	-6.19%
bigint	2,104 (-163)	-7.19%
keccak256	1,829 (-368)	-16.75%
slices	1,831 (-541)	-22.81%
reference_counts	360 (-170)	-32.08%

[github-actions]

Changes to number of Brillig opcodes executed

Generated at commit: d2816639721b972fde9772cc08e9b8fb5788bd17, compared to commit: e97339733000d19163372c76b299333c20db0299

🧾 Summary (10% most significant diffs)

Program	Brillig opcodes (+/-)	%
pedersen_hash	+141 ❌	+26.65%
brillig_pedersen	+178 ❌	+22.25%
pedersen_check	+178 ❌	+22.25%
simple_shield	+453 ❌	+18.88%
pedersen_commitment	+46 ❌	+18.47%
merkle_insert	+586 ❌	+18.34%
fold_2_to_17	+198,851 ❌	+18.18%
bench_2_to_17	+104,559 ❌	+17.73%
reference_counts	-133 ✅	-29.10%

---

Full diff report 👇

Program	Brillig opcodes (+/-)	%
pedersen_hash	670 (+141)	+26.65%
brillig_pedersen	978 (+178)	+22.25%
pedersen_check	978 (+178)	+22.25%
simple_shield	2,853 (+453)	+18.88%
pedersen_commitment	295 (+46)	+18.47%
merkle_insert	3,781 (+586)	+18.34%
fold_2_to_17	1,292,773 (+198,851)	+18.18%
bench_2_to_17	694,397 (+104,559)	+17.73%
fold_numeric_generic_poseidon	5,862 (+855)	+17.08%
no_predicates_numeric_generic_poseidon	5,862 (+855)	+17.08%
poseidon2	803 (+108)	+15.54%
array_sort	607 (+81)	+15.40%
wildcard_type	592 (+70)	+13.41%
regression_6674_3	1,812 (+200)	+12.41%
brillig_rc_regression_6123	335 (+36)	+12.04%
databus_two_calldata	445 (+46)	+11.53%
regression_struct_array_conditional	1,710 (+165)	+10.68%
to_be_bytes	2,454 (+235)	+10.59%
conditional_1	5,909 (+484)	+8.92%
brillig_cow_regression	520,383 (+40,946)	+8.54%
ram_blowup_regression	780,775 (+60,552)	+8.41%
regression_6674_1	896 (+68)	+8.21%
brillig_oracle	2,624 (+185)	+7.59%
sha256_var_padding_regression	220,289 (+15,438)	+7.54%
regression_6674_2	884 (+50)	+6.00%
7_function	2,476 (+131)	+5.59%
array_dynamic_nested_blackbox_input	4,753 (+246)	+5.46%
nested_array_dynamic_simple	126 (+6)	+5.00%
fold_call_witness_condition	70 (+3)	+4.48%
aes128_encrypt	4,514 (+176)	+4.06%
array_dynamic_blackbox_input	18,587 (+675)	+3.77%
ecdsa_secp256k1	6,982 (+245)	+3.64%
brillig_constant_reference_regression	87 (+3)	+3.57%
sha2_byte	46,959 (+1,423)	+3.13%
uhashmap	151,372 (+4,424)	+3.01%
sha256_var_witness_const_regression	7,017 (+183)	+2.68%
higher_order_functions	1,269 (+33)	+2.67%
conditional_2	122 (+3)	+2.52%
conditional_regression_661	123 (+3)	+2.50%
6	7,502 (+156)	+2.12%
conditional_regression_short_circuit	7,580 (+156)	+2.10%
keccak256	33,164 (+579)	+1.78%
sha256_brillig_performance_regression	23,107 (+361)	+1.59%
regression_4202	214 (+3)	+1.42%
references	239 (+3)	+1.27%
nested_dyn_array_regression_5782	176 (+2)	+1.15%
fold_complex_outputs	922 (+10)	+1.10%
brillig_nested_arrays	356 (+3)	+0.85%
nested_array_dynamic	3,746 (+26)	+0.70%
u128	25,012 (+118)	+0.47%
array_dynamic	500 (+2)	+0.40%
slice_loop	960 (+3)	+0.31%
sha256_var_size_regression	16,658 (+44)	+0.26%
loop_invariant_regression	1,167 (+3)	+0.26%
regression	2,831 (+6)	+0.21%
brillig_cow_assign	534 (+1)	+0.19%
hashmap	55,604 (+64)	+0.12%
slice_dynamic_index	4,684 (+2)	+0.04%
modulus	19,184 (+6)	+0.03%
sha256	14,534 (+2)	+0.01%
poseidonsponge_x5_254	184,085 (-61)	-0.03%
brillig_cow	1,153 (-2)	-0.17%
sha256_regression	116,973 (-207)	-0.18%
array_if_cond_simple	537 (-1)	-0.19%
regression_5252	917,714 (-2,300)	-0.25%
array_dynamic_main_output	296 (-1)	-0.34%
nested_array_in_slice	1,735 (-6)	-0.34%
struct_inputs	584 (-3)	-0.51%
brillig_arrays	139 (-1)	-0.71%
array_to_slice	1,664 (-21)	-1.25%
debug_logs	5,097 (-75)	-1.45%
strings	1,767 (-27)	-1.51%
slice_regex	3,798 (-63)	-1.63%
poseidon_bn254_hash	162,947 (-3,593)	-2.16%
poseidon_bn254_hash_width_3	162,947 (-3,593)	-2.16%
regression_4449	214,164 (-7,128)	-3.22%
slices	3,136 (-358)	-10.25%
reference_counts	324 (-133)	-29.10%

[github-actions]

Changes to circuit sizes

Generated at commit: d2816639721b972fde9772cc08e9b8fb5788bd17, compared to commit: e97339733000d19163372c76b299333c20db0299

🧾 Summary (10% most significant diffs)

Program	ACIR opcodes (+/-)	%	Circuit size (+/-)	%
reference_counts	-4 ✅	-57.14%	0 ➖	0.00%
hashmap	-18,650 ✅	-37.87%	-43,024 ✅	-38.07%

---

Full diff report 👇

Program	ACIR opcodes (+/-)	%	Circuit size (+/-)	%
reference_counts	3 (-4)	-57.14%	16 (0)	0.00%
hashmap	30,599 (-18,650)	-37.87%	69,989 (-43,024)	-38.07%

[github-actions]

Peak Memory Sample

Program	Peak Memory	%
keccak256	82.02M	1%
workspace	121.87M	-1%
regression_4709	327.82M	0%
ram_blowup_regression	2.54G	0%
private-kernel-tail	231.79M	0%
private-kernel-reset	1.37G	0%
private-kernel-inner	338.30M	0%
parity-root	199.34M	0%

[github-actions]

Compilation Sample

Program	Compilation Time	%
sha256_regression	0m1.616s	10%
regression_4709	0m1.535s	-4%
ram_blowup_regression	0m17.064s	-1%
private-kernel-tail	0m1.224s	0%
private-kernel-reset	0m10.015s	4%
private-kernel-inner	0m2.474s	-7%
parity-root	0m1.196s	35%
noir-contracts	2m49.183s	0%

[jfecher]

Closing this in favor of #6778