Add Neon NTT

[mkannwischer]

• Resolves Neon backend: Add NTT #197

I have not run it through SLOTHY yet. I'll do that in a follow-up PR.

[github-actions]

Arm Cortex-A55 (Snapdragon 888) benchmarks (no-opt)

Benchmark suite	Current: 8fda85e	Previous: 3443561	Ratio
ML-DSA-44 keypair	490685 cycles	490045 cycles	1.00
ML-DSA-44 sign	1854022 cycles	1854795 cycles	1.00
ML-DSA-44 verify	563360 cycles	563551 cycles	1.00
ML-DSA-65 keypair	829537 cycles	829847 cycles	1.00
ML-DSA-65 sign	2970020 cycles	2971899 cycles	1.00
ML-DSA-65 verify	880555 cycles	880975 cycles	1.00
ML-DSA-87 keypair	1345789 cycles	1343166 cycles	1.00
ML-DSA-87 sign	3774144 cycles	3772895 cycles	1.00
ML-DSA-87 verify	1433636 cycles	1433664 cycles	1.00

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

[github-actions]

Mac Mini (M1, 2020) benchmarks (no-opt)

Benchmark suite	Current: 8fda85e	Previous: 3443561	Ratio
ML-DSA-44 keypair	132460 cycles	132396 cycles	1.00
ML-DSA-44 sign	411876 cycles	411635 cycles	1.00
ML-DSA-44 verify	142608 cycles	142566 cycles	1.00
ML-DSA-65 keypair	231387 cycles	231322 cycles	1.00
ML-DSA-65 sign	679524 cycles	679473 cycles	1.00
ML-DSA-65 verify	231551 cycles	231676 cycles	1.00
ML-DSA-87 keypair	381567 cycles	381535 cycles	1.00
ML-DSA-87 sign	877936 cycles	878007 cycles	1.00
ML-DSA-87 verify	391407 cycles	391350 cycles	1.00

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

[github-actions]

Arm Cortex-A76 (Raspberry Pi 5) benchmarks (no-opt)

Benchmark suite	Current: 8fda85e	Previous: 3443561	Ratio
ML-DSA-44 keypair	224150 cycles	224243 cycles	1.00
ML-DSA-44 sign	654841 cycles	654969 cycles	1.00
ML-DSA-44 verify	240715 cycles	240707 cycles	1.00
ML-DSA-65 keypair	400314 cycles	400195 cycles	1.00
ML-DSA-65 sign	1074694 cycles	1073293 cycles	1.00
ML-DSA-65 verify	394290 cycles	394266 cycles	1.00
ML-DSA-87 keypair	647216 cycles	646992 cycles	1.00
ML-DSA-87 sign	1394731 cycles	1395196 cycles	1.00
ML-DSA-87 verify	661660 cycles	661462 cycles	1.00

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

[github-actions]

SpacemiT K1 8 (Banana Pi F3) benchmarks (no-opt)

Benchmark suite	Current: 8fda85e	Previous: 3443561	Ratio
ML-DSA-44 keypair	1026839 cycles	1025279 cycles	1.00
ML-DSA-44 sign	3900331 cycles	3899447 cycles	1.00
ML-DSA-44 verify	1164942 cycles	1164811 cycles	1.00
ML-DSA-65 keypair	1727119 cycles	1729203 cycles	1.00
ML-DSA-65 sign	6377417 cycles	6384540 cycles	1.00
ML-DSA-65 verify	1864556 cycles	1863531 cycles	1.00
ML-DSA-87 keypair	2848223 cycles	2852215 cycles	1.00
ML-DSA-87 sign	7952701 cycles	7949346 cycles	1.00
ML-DSA-87 verify	3034841 cycles	3034846 cycles	1.00

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

[github-actions]

Arm Cortex-A72 (Raspberry Pi 4) benchmarks (no-opt)

Benchmark suite	Current: 8fda85e	Previous: 3443561	Ratio
ML-DSA-44 keypair	310005 cycles	309866 cycles	1.00
ML-DSA-44 sign	976974 cycles	1091537 cycles	0.90
ML-DSA-44 verify	334029 cycles	334798 cycles	1.00
ML-DSA-65 keypair	566219 cycles	568405 cycles	1.00
ML-DSA-65 sign	1584266 cycles	1636741 cycles	0.97
ML-DSA-65 verify	543612 cycles	542674 cycles	1.00
ML-DSA-87 keypair	875101 cycles	875559 cycles	1.00
ML-DSA-87 sign	2057254 cycles	2044313 cycles	1.01
ML-DSA-87 verify	902044 cycles	901713 cycles	1.00

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

[github-actions]

⚠️ Performance Alert ⚠️

Possible performance regression was detected for benchmark 'Arm Cortex-A72 (Raspberry Pi 4) benchmarks (no-opt)'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.03.

Benchmark suite	Current: 6141304	Previous: 3443561	Ratio
ML-DSA-44 sign	1130374 cycles	1091537 cycles	1.04

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

[github-actions]

Arm Cortex-A72 (Raspberry Pi 4) benchmarks (opt)

Benchmark suite	Current: 8fda85e	Previous: 535dc05	Ratio
ML-DSA-44 keypair	295530 cycles	304742 cycles	0.97
ML-DSA-44 sign	968878 cycles	991066 cycles	0.98
ML-DSA-44 verify	299510 cycles	329368 cycles	0.91
ML-DSA-65 keypair	548558 cycles	561239 cycles	0.98
ML-DSA-65 sign	1428807 cycles	1544501 cycles	0.93
ML-DSA-65 verify	499443 cycles	537517 cycles	0.93
ML-DSA-87 keypair	850185 cycles	866585 cycles	0.98
ML-DSA-87 sign	1910316 cycles	1995595 cycles	0.96
ML-DSA-87 verify	845528 cycles	894378 cycles	0.95

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

[mkannwischer]

@hanno-becker - could you please take a look at this one?
I kept the changes to the assembly (from SLOTHY) as separate commits so that it's easier to review. Before we merge, I'd like to squash all commits changing the asm.

[hanno-becker]

Are you sure this is going to work? I always get confused what is a string and what is a boolean working with GH actions.

[mkannwischer]

It wasn't working before - this step was never run - that's why I changed it.