chore(ssa): Run purity analysis before preprocessing of functions

[vezenovm]

Description

Problem*

No issue, I just noticed that during preprocessing we run LICM however we do not have any purity information. We run purity analysis before LICM in our standard SSA pass flow for this reason.

Summary*

Add a purity analysis pass before preprocessing.

In order for this to be safe we also had to unify recursive ACIR functions with PureWithPredicates. Similarly to Brillig, if we marked recursive functions as pure we could accidentally remove an infinite loop. This only came up after this change as purity is now being run before inlining and as we inline everything in ACIR, indefinite recursion (or in this case a recursion limit) is handled during inlining.

If we have this program:

acir(inline) fn pure main f0 {
  b0():
    call f1()
    return
}
acir(inline) fn pure main f1 {
  b0():
    call f1()
    return
}

Running DIE before inlining (which is what occurs during preprocess_fns) would remove the infinite loop.

The benchmarks don't show many improvements. As there are not any regressions either though I think this is fine to merge as LICM will now have purity information it did not before.

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]

ACVM Benchmarks

Details

Benchmark suite	Current: 9c20e98	Previous: 8460767	Ratio
purely_sequential_opcodes	249441 ns/iter (± 1537)	257200 ns/iter (± 726)	0.97
perfectly_parallel_opcodes	219301 ns/iter (± 4161)	234232 ns/iter (± 1108)	0.94
perfectly_parallel_batch_inversion_opcodes	2782702 ns/iter (± 21870)	2819468 ns/iter (± 1954)	0.99

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

[github-actions]

Artifact Size

Details

Benchmark suite	Current: 9c20e98	Previous: 8460767	Ratio
private-kernel-inner	712.1 KB	712.1 KB	1
private-kernel-reset	1983.6 KB	1983.6 KB	1
private-kernel-tail	506.6 KB	506.6 KB	1
rollup-base-private	4361.8 KB	4361.8 KB	1
rollup-base-public	3365.6 KB	3365.6 KB	1
rollup-block-root-empty	3851.2 KB	3851.2 KB	1
rollup-block-root-single-tx	30879.7 KB	30879.7 KB	1
rollup-block-root	30908.3 KB	30908.3 KB	1
rollup-merge	189.1 KB	189.1 KB	1
rollup-root	391.2 KB	391.2 KB	1
semaphore-depth-10	634 KB	634 KB	1
sha512-100-bytes	527.6 KB	527.6 KB	1

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

[github-actions]

Compilation Time

Details

Benchmark suite	Current: 9c20e98	Previous: 8460767	Ratio
private-kernel-inner	1.874 s	1.83 s	1.02
private-kernel-reset	8.494 s	8.21 s	1.03
private-kernel-tail	1.35 s	1.414 s	0.95
rollup-base-private	19 s	20 s	0.95
rollup-base-public	17.16 s	17.52 s	0.98
rollup-block-root-empty	18.02 s	18.12 s	0.99
rollup-block-root-single-tx	200 s	199 s	1.01
rollup-block-root	195 s	191 s	1.02
rollup-merge	1.38 s	1.422 s	0.97
rollup-root	1.602 s	1.612 s	0.99
semaphore-depth-10	0.792 s	0.79 s	1.00
sha512-100-bytes	1.901 s	1.791 s	1.06

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

[github-actions]

Opcode count

Details

Benchmark suite	Current: 9c20e98	Previous: 8460767	Ratio
private-kernel-inner	14792 opcodes	14792 opcodes	1
private-kernel-reset	68868 opcodes	68868 opcodes	1
private-kernel-tail	11177 opcodes	11177 opcodes	1
rollup-base-private	221346 opcodes	221346 opcodes	1
rollup-base-public	159937 opcodes	159937 opcodes	1
rollup-block-root-empty	67868 opcodes	67868 opcodes	1
rollup-block-root-single-tx	963687 opcodes	963687 opcodes	1
rollup-block-root	964973 opcodes	964973 opcodes	1
rollup-merge	1409 opcodes	1409 opcodes	1
rollup-root	2619 opcodes	2619 opcodes	1
semaphore-depth-10	5700 opcodes	5700 opcodes	1
sha512-100-bytes	13173 opcodes	13173 opcodes	1

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

[github-actions]

Execution Time

Details

Benchmark suite	Current: 9c20e98	Previous: 8460767	Ratio
private-kernel-inner	0.016 s	0.015 s	1.07
private-kernel-reset	0.162 s	0.16 s	1.01
private-kernel-tail	0.009 s	0.01 s	0.90
rollup-base-private	0.268 s	0.267 s	1.00
rollup-base-public	0.164 s	0.159 s	1.03
rollup-block-root	14.5 s	14.3 s	1.01
rollup-merge	0.002 s	0.002 s	1
rollup-root	0.005 s	0.004 s	1.25
semaphore-depth-10	0.019 s	0.019 s	1
sha512-100-bytes	0.103 s	0.103 s	1

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

[github-actions]

Test Suite Duration

Details

Benchmark suite	Current: 9c20e98	Previous: 8460767	Ratio
test_report_AztecProtocol_aztec-packages_noir-projects_aztec-nr	103 s	103 s	1
test_report_AztecProtocol_aztec-packages_noir-projects_noir-contracts	132 s	133 s	0.99
test_report_AztecProtocol_aztec-packages_noir-projects_noir-protocol-circuits_crates_blob	211 s	195 s	1.08
test_report_AztecProtocol_aztec-packages_noir-projects_noir-protocol-circuits_crates_private-kernel-lib	218 s	226 s	0.96
test_report_AztecProtocol_aztec-packages_noir-projects_noir-protocol-circuits_crates_reset-kernel-lib	34 s	35 s	0.97
test_report_AztecProtocol_aztec-packages_noir-projects_noir-protocol-circuits_crates_rollup-lib	642 s	647 s	0.99
test_report_AztecProtocol_aztec-packages_noir-projects_noir-protocol-circuits_crates_types	102 s	101 s	1.01
test_report_noir-lang_noir-bignum_	136 s	130 s	1.05
test_report_noir-lang_noir_bigcurve_	348 s	343 s	1.01
test_report_noir-lang_sha256_	16 s	15 s	1.07
test_report_noir-lang_sha512_	13 s	13 s	1
test_report_zkpassport_noir-ecdsa_	2 s	1 s	2
test_report_zkpassport_noir_rsa_	1 s	1 s	1

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

[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: 9c20e98	Previous: 8460767	Ratio
test_report_zkpassport_noir-ecdsa_	2 s	1 s	2

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

CC: @TomAFrench

[github-actions]

Execution Memory

Details

Benchmark suite	Current: 9c20e98	Previous: 8460767	Ratio
private-kernel-inner	258.74 MB	258.74 MB	1
private-kernel-reset	291.13 MB	291.13 MB	1
private-kernel-tail	242.63 MB	242.63 MB	1
rollup-base-private	506.71 MB	506.71 MB	1
rollup-base-public	440.02 MB	440.02 MB	1
rollup-block-root	1530 MB	1530 MB	1
rollup-merge	331.06 MB	331.06 MB	1
rollup-root	333.57 MB	333.57 MB	1
semaphore_depth_10	73 MB	73 MB	1
sha512_100_bytes	69.57 MB	69.57 MB	1

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

[github-actions]

Compilation Memory

Details

Benchmark suite	Current: 9c20e98	Previous: 8460767	Ratio
private-kernel-inner	285.1 MB	285.13 MB	1.00
private-kernel-reset	593.15 MB	593.15 MB	1
private-kernel-tail	258.97 MB	258.97 MB	1
rollup-base-private	1360 MB	1360 MB	1
rollup-base-public	1420 MB	1420 MB	1
rollup-block-root-empty	1020 MB	1020 MB	1
rollup-block-root-single-tx	9680 MB	9680 MB	1
rollup-block-root	9690 MB	9690 MB	1
rollup-merge	333.28 MB	333.28 MB	1
rollup-root	343.99 MB	343.99 MB	1
semaphore_depth_10	107.77 MB	107.77 MB	1
sha512_100_bytes	265.66 MB	265.73 MB	1.00

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

[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: 9c20e98	Previous: 8460767	Ratio
rollup-root	0.005 s	0.004 s	1.25

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

CC: @TomAFrench