Implement poly_use_hint on top of poly_decompose

[mkannwischer]

• Depends on Add native implementation of poly_decompose #411
• Resolves AVX2: Consider adding poly_use_hint assembly #400
• Resolves AArch64: Consider adding poly_use_hint assembly #398

This eliminates the helper function use_hint and instead inlines it
into poly_use_hint to be able to re-use the native poly_decompose code.
The disadvantage is that it requires an additional polynomial on the
stack.

[github-actions]

Mac Mini (M1, 2020) benchmarks (opt)

Benchmark suite	Current: 30d236c	Previous: 1736882	Ratio
ML-DSA-44 keypair	50500 cycles	50491 cycles	1.00
ML-DSA-44 sign	205382 cycles	222970 cycles	0.92
ML-DSA-44 verify	71838 cycles	72843 cycles	0.99
ML-DSA-65 keypair	87381 cycles	87390 cycles	1.00
ML-DSA-65 sign	330994 cycles	356076 cycles	0.93
ML-DSA-65 verify	111562 cycles	112665 cycles	0.99
ML-DSA-87 keypair	140123 cycles	140111 cycles	1.00
ML-DSA-87 sign	400435 cycles	425586 cycles	0.94
ML-DSA-87 verify	171789 cycles	173177 cycles	0.99

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

[github-actions]

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

Benchmark suite	Current: 30d236c	Previous: 1736882	Ratio
ML-DSA-44 keypair	115992 cycles	116006 cycles	1.00
ML-DSA-44 sign	455097 cycles	455029 cycles	1.00
ML-DSA-44 verify	139943 cycles	136872 cycles	1.02
ML-DSA-65 keypair	197959 cycles	197989 cycles	1.00
ML-DSA-65 sign	733264 cycles	733207 cycles	1.00
ML-DSA-65 verify	220850 cycles	216803 cycles	1.02
ML-DSA-87 keypair	335256 cycles	335091 cycles	1.00
ML-DSA-87 sign	915276 cycles	915023 cycles	1.00
ML-DSA-87 verify	358855 cycles	353226 cycles	1.02

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

[oqs-bot]

Intel Xeon 4th gen (c7i)

Benchmark suite	Current: 30d236c	Previous: 1736882	Ratio
ML-DSA-44 keypair	37849 cycles	37381 cycles	1.01
ML-DSA-44 sign	156256 cycles	169592 cycles	0.92
ML-DSA-44 verify	47881 cycles	50107 cycles	0.96
ML-DSA-65 keypair	66644 cycles	65875 cycles	1.01
ML-DSA-65 sign	262073 cycles	279226 cycles	0.94
ML-DSA-65 verify	75847 cycles	78656 cycles	0.96
ML-DSA-87 keypair	99769 cycles	100966 cycles	0.99
ML-DSA-87 sign	302956 cycles	326836 cycles	0.93
ML-DSA-87 verify	111504 cycles	117910 cycles	0.95

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

[oqs-bot]

AMD EPYC 4th gen (c7a)

Benchmark suite	Current: 30d236c	Previous: 1736882	Ratio
ML-DSA-44 keypair	44319 cycles	44805 cycles	0.99
ML-DSA-44 sign	176844 cycles	197282 cycles	0.90
ML-DSA-44 verify	57118 cycles	60605 cycles	0.94
ML-DSA-65 keypair	75944 cycles	85544 cycles	0.89
ML-DSA-65 sign	282633 cycles	329012 cycles	0.86
ML-DSA-65 verify	88237 cycles	101933 cycles	0.87
ML-DSA-87 keypair	116142 cycles	115872 cycles	1.00
ML-DSA-87 sign	335251 cycles	369088 cycles	0.91
ML-DSA-87 verify	130685 cycles	138380 cycles	0.94

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

[oqs-bot]

Intel Xeon 4th gen (c7i) (no-opt)

Benchmark suite	Current: 30d236c	Previous: 1736882	Ratio
ML-DSA-44 keypair	96173 cycles	96162 cycles	1.00
ML-DSA-44 sign	351427 cycles	351587 cycles	1.00
ML-DSA-44 verify	106204 cycles	105379 cycles	1.01
ML-DSA-65 keypair	163642 cycles	164036 cycles	1.00
ML-DSA-65 sign	580111 cycles	583065 cycles	0.99
ML-DSA-65 verify	171360 cycles	169132 cycles	1.01
ML-DSA-87 keypair	273379 cycles	273210 cycles	1.00
ML-DSA-87 sign	733534 cycles	733446 cycles	1.00
ML-DSA-87 verify	281281 cycles	280884 cycles	1.00

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

[oqs-bot]

Intel Xeon 3rd gen (c6i)

Benchmark suite	Current: 30d236c	Previous: 1736882	Ratio
ML-DSA-44 keypair	61006 cycles	61211 cycles	1.00
ML-DSA-44 sign	241339 cycles	264347 cycles	0.91
ML-DSA-44 verify	76803 cycles	80506 cycles	0.95
ML-DSA-65 keypair	107425 cycles	107017 cycles	1.00
ML-DSA-65 sign	400256 cycles	436791 cycles	0.92
ML-DSA-65 verify	122508 cycles	127415 cycles	0.96
ML-DSA-87 keypair	166031 cycles	165409 cycles	1.00
ML-DSA-87 sign	479942 cycles	516438 cycles	0.93
ML-DSA-87 verify	184026 cycles	190577 cycles	0.97

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

[oqs-bot]

AMD EPYC 3rd gen (c6a)

Benchmark suite	Current: 30d236c	Previous: 1736882	Ratio
ML-DSA-44 keypair	77797 cycles	74255 cycles	1.05
ML-DSA-44 sign	257720 cycles	269307 cycles	0.96
ML-DSA-44 verify	89884 cycles	89061 cycles	1.01
ML-DSA-65 keypair	126917 cycles	126527 cycles	1.00
ML-DSA-65 sign	400004 cycles	433751 cycles	0.92
ML-DSA-65 verify	137095 cycles	142344 cycles	0.96
ML-DSA-87 keypair	210105 cycles	210098 cycles	1.00
ML-DSA-87 sign	506180 cycles	543857 cycles	0.93
ML-DSA-87 verify	223266 cycles	230154 cycles	0.97

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: 30d236c	Previous: 1736882	Ratio
ML-DSA-44 keypair	826958 cycles	830867 cycles	1.00
ML-DSA-44 sign	3365565 cycles	3361137 cycles	1.00
ML-DSA-44 verify	934455 cycles	930654 cycles	1.00
ML-DSA-65 keypair	1393322 cycles	1394406 cycles	1.00
ML-DSA-65 sign	5487332 cycles	5490767 cycles	1.00
ML-DSA-65 verify	1484934 cycles	1482041 cycles	1.00
ML-DSA-87 keypair	2305366 cycles	2310541 cycles	1.00
ML-DSA-87 sign	6883188 cycles	6926078 cycles	0.99
ML-DSA-87 verify	2430628 cycles	2437273 cycles	1.00

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

[oqs-bot]

⚠️ Performance Alert ⚠️

Possible performance regression was detected for benchmark 'AMD EPYC 3rd gen (c6a)'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.03.

Benchmark suite	Current: 30d236c	Previous: 1736882	Ratio
ML-DSA-44 keypair	77797 cycles	74255 cycles	1.05

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

[oqs-bot]

Graviton4

Benchmark suite	Current: 30d236c	Previous: 1736882	Ratio
ML-DSA-44 keypair	73057 cycles	72955 cycles	1.00
ML-DSA-44 sign	262119 cycles	282599 cycles	0.93
ML-DSA-44 verify	85637 cycles	87098 cycles	0.98
ML-DSA-65 keypair	128897 cycles	128338 cycles	1.00
ML-DSA-65 sign	431334 cycles	461044 cycles	0.94
ML-DSA-65 verify	138078 cycles	139034 cycles	0.99
ML-DSA-87 keypair	208233 cycles	208000 cycles	1.00
ML-DSA-87 sign	536103 cycles	563267 cycles	0.95
ML-DSA-87 verify	220559 cycles	222676 cycles	0.99

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

[oqs-bot]

AMD EPYC 4th gen (c7a) (no-opt)

Benchmark suite	Current: 30d236c	Previous: 1736882	Ratio
ML-DSA-44 keypair	121614 cycles	121992 cycles	1.00
ML-DSA-44 sign	466651 cycles	467100 cycles	1.00
ML-DSA-44 verify	138107 cycles	137172 cycles	1.01
ML-DSA-65 keypair	205647 cycles	205879 cycles	1.00
ML-DSA-65 sign	752379 cycles	752229 cycles	1.00
ML-DSA-65 verify	218566 cycles	216901 cycles	1.01
ML-DSA-87 keypair	340764 cycles	341355 cycles	1.00
ML-DSA-87 sign	955524 cycles	953098 cycles	1.00
ML-DSA-87 verify	360409 cycles	357806 cycles	1.01

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

[oqs-bot]

Intel Xeon 3rd gen (c6i) (no-opt)

Benchmark suite	Current: 30d236c	Previous: 1736882	Ratio
ML-DSA-44 keypair	159081 cycles	159435 cycles	1.00
ML-DSA-44 sign	574012 cycles	574575 cycles	1.00
ML-DSA-44 verify	176683 cycles	175326 cycles	1.01
ML-DSA-65 keypair	271339 cycles	272743 cycles	0.99
ML-DSA-65 sign	950325 cycles	948311 cycles	1.00
ML-DSA-65 verify	286775 cycles	284273 cycles	1.01
ML-DSA-87 keypair	452772 cycles	452203 cycles	1.00
ML-DSA-87 sign	1193487 cycles	1194217 cycles	1.00
ML-DSA-87 verify	471296 cycles	469663 cycles	1.00

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

[oqs-bot]

Graviton3

Benchmark suite	Current: 30d236c	Previous: 1736882	Ratio
ML-DSA-44 keypair	78057 cycles	78015 cycles	1.00
ML-DSA-44 sign	282099 cycles	304192 cycles	0.93
ML-DSA-44 verify	92841 cycles	95531 cycles	0.97
ML-DSA-65 keypair	135025 cycles	134970 cycles	1.00
ML-DSA-65 sign	464808 cycles	496694 cycles	0.94
ML-DSA-65 verify	149724 cycles	151189 cycles	0.99
ML-DSA-87 keypair	217783 cycles	217547 cycles	1.00
ML-DSA-87 sign	573887 cycles	606490 cycles	0.95
ML-DSA-87 verify	237677 cycles	239637 cycles	0.99

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

[oqs-bot]

AMD EPYC 3rd gen (c6a) (no-opt)

Benchmark suite	Current: 30d236c	Previous: 1736882	Ratio
ML-DSA-44 keypair	136457 cycles	136784 cycles	1.00
ML-DSA-44 sign	554395 cycles	554088 cycles	1.00
ML-DSA-44 verify	156645 cycles	155245 cycles	1.01
ML-DSA-65 keypair	228109 cycles	227723 cycles	1.00
ML-DSA-65 sign	893154 cycles	894394 cycles	1.00
ML-DSA-65 verify	245855 cycles	244248 cycles	1.01
ML-DSA-87 keypair	375905 cycles	377140 cycles	1.00
ML-DSA-87 sign	1112907 cycles	1115587 cycles	1.00
ML-DSA-87 verify	400199 cycles	398881 cycles	1.00

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

[oqs-bot]

Graviton2

Benchmark suite	Current: 30d236c	Previous: 1736882	Ratio
ML-DSA-44 keypair	120487 cycles	120805 cycles	1.00
ML-DSA-44 sign	453632 cycles	489107 cycles	0.93
ML-DSA-44 verify	142359 cycles	146029 cycles	0.97
ML-DSA-65 keypair	207537 cycles	207549 cycles	1.00
ML-DSA-65 sign	749222 cycles	802620 cycles	0.93
ML-DSA-65 verify	229852 cycles	231722 cycles	0.99
ML-DSA-87 keypair	336755 cycles	336647 cycles	1.00
ML-DSA-87 sign	930353 cycles	986391 cycles	0.94
ML-DSA-87 verify	365481 cycles	370216 cycles	0.99

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

[oqs-bot]

Graviton4 (no-opt)

Benchmark suite	Current: 30d236c	Previous: 1736882	Ratio
ML-DSA-44 keypair	134640 cycles	134737 cycles	1.00
ML-DSA-44 sign	520155 cycles	508680 cycles	1.02
ML-DSA-44 verify	152799 cycles	149623 cycles	1.02
ML-DSA-65 keypair	228468 cycles	228452 cycles	1.00
ML-DSA-65 sign	825053 cycles	824132 cycles	1.00
ML-DSA-65 verify	240390 cycles	237291 cycles	1.01
ML-DSA-87 keypair	376971 cycles	377483 cycles	1.00
ML-DSA-87 sign	1029692 cycles	1030717 cycles	1.00
ML-DSA-87 verify	395171 cycles	391020 cycles	1.01

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

[oqs-bot]

Graviton3 (no-opt)

Benchmark suite	Current: 30d236c	Previous: 1736882	Ratio
ML-DSA-44 keypair	139780 cycles	139749 cycles	1.00
ML-DSA-44 sign	512932 cycles	505375 cycles	1.01
ML-DSA-44 verify	156718 cycles	154405 cycles	1.01
ML-DSA-65 keypair	245178 cycles	244944 cycles	1.00
ML-DSA-65 sign	824538 cycles	824048 cycles	1.00
ML-DSA-65 verify	252977 cycles	248646 cycles	1.02
ML-DSA-87 keypair	397778 cycles	397303 cycles	1.00
ML-DSA-87 sign	1044062 cycles	1043765 cycles	1.00
ML-DSA-87 verify	417220 cycles	411283 cycles	1.01

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

[github-actions]

Arm Cortex-A76 (Raspberry Pi 5) benchmarks (opt)

Benchmark suite	Current: 30d236c	Previous: 1736882	Ratio
ML-DSA-44 keypair	120417 cycles	120271 cycles	1.00
ML-DSA-44 sign	453396 cycles	488352 cycles	0.93
ML-DSA-44 verify	142184 cycles	145585 cycles	0.98
ML-DSA-65 keypair	207481 cycles	207309 cycles	1.00
ML-DSA-65 sign	748396 cycles	802690 cycles	0.93
ML-DSA-65 verify	229905 cycles	231615 cycles	0.99
ML-DSA-87 keypair	336582 cycles	336114 cycles	1.00
ML-DSA-87 sign	929423 cycles	985677 cycles	0.94
ML-DSA-87 verify	365760 cycles	370072 cycles	0.99

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

[oqs-bot]

Graviton2 (no-opt)

Benchmark suite	Current: 30d236c	Previous: 1736882	Ratio
ML-DSA-44 keypair	215251 cycles	215510 cycles	1.00
ML-DSA-44 sign	799024 cycles	809958 cycles	0.99
ML-DSA-44 verify	244130 cycles	239793 cycles	1.02
ML-DSA-65 keypair	383492 cycles	383922 cycles	1.00
ML-DSA-65 sign	1310912 cycles	1313723 cycles	1.00
ML-DSA-65 verify	391832 cycles	385228 cycles	1.02
ML-DSA-87 keypair	611660 cycles	611830 cycles	1.00
ML-DSA-87 sign	1665709 cycles	1666283 cycles	1.00
ML-DSA-87 verify	648641 cycles	637481 cycles	1.02

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: 30d236c	Previous: 1736882	Ratio
ML-DSA-44 keypair	214965 cycles	214904 cycles	1.00
ML-DSA-44 sign	797783 cycles	797156 cycles	1.00
ML-DSA-44 verify	243977 cycles	239869 cycles	1.02
ML-DSA-65 keypair	383157 cycles	383600 cycles	1.00
ML-DSA-65 sign	1318623 cycles	1321544 cycles	1.00
ML-DSA-65 verify	391639 cycles	384834 cycles	1.02
ML-DSA-87 keypair	610983 cycles	611160 cycles	1.00
ML-DSA-87 sign	1663256 cycles	1663574 cycles	1.00
ML-DSA-87 verify	648213 cycles	637385 cycles	1.02

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

[mkannwischer]

Hmmmm, seems to work okay on some Intel platforms, but not great on AArch64.
Will need to compare this to a fully native implementation.

[github-actions]

Arm Cortex-A55 (Snapdragon 888) benchmarks (opt)

Benchmark suite	Current: 30d236c	Previous: 1736882	Ratio
ML-DSA-44 keypair	295501 cycles	299821 cycles	0.99
ML-DSA-44 sign	1136287 cycles	1229073 cycles	0.92
ML-DSA-44 verify	342899 cycles	361928 cycles	0.95
ML-DSA-65 keypair	502192 cycles	503702 cycles	1.00
ML-DSA-65 sign	1858520 cycles	1996981 cycles	0.93
ML-DSA-65 verify	540919 cycles	559218 cycles	0.97
ML-DSA-87 keypair	853068 cycles	850138 cycles	1.00
ML-DSA-87 sign	2426368 cycles	2557081 cycles	0.95
ML-DSA-87 verify	905590 cycles	924273 cycles	0.98

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

[github-actions]

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

Benchmark suite	Current: 30d236c	Previous: 1736882	Ratio
ML-DSA-44 keypair	470070 cycles	469100 cycles	1.00
ML-DSA-44 sign	2254019 cycles	2237875 cycles	1.01
ML-DSA-44 verify	571416 cycles	564986 cycles	1.01
ML-DSA-65 keypair	787327 cycles	789950 cycles	1.00
ML-DSA-65 sign	3674530 cycles	3700139 cycles	0.99
ML-DSA-65 verify	878677 cycles	872036 cycles	1.01
ML-DSA-87 keypair	1262444 cycles	1268690 cycles	1.00
ML-DSA-87 sign	4489950 cycles	4543181 cycles	0.99
ML-DSA-87 verify	1403237 cycles	1397372 cycles	1.00

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

[github-actions]

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

Benchmark suite	Current: 30d236c	Previous: 1736882	Ratio
ML-DSA-44 keypair	229696 cycles	236241 cycles	0.97
ML-DSA-44 sign	777476 cycles	809081 cycles	0.96
ML-DSA-44 verify	255039 cycles	267657 cycles	0.95
ML-DSA-65 keypair	417139 cycles	404121 cycles	1.03
ML-DSA-65 sign	1309627 cycles	1315060 cycles	1.00
ML-DSA-65 verify	436970 cycles	416690 cycles	1.05
ML-DSA-87 keypair	670883 cycles	659871 cycles	1.02
ML-DSA-87 sign	1598694 cycles	1679421 cycles	0.95
ML-DSA-87 verify	687899 cycles	677911 cycles	1.01

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 (opt)'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.03.

Benchmark suite	Current: 30d236c	Previous: 1736882	Ratio
ML-DSA-65 keypair	417139 cycles	404121 cycles	1.03
ML-DSA-65 verify	436970 cycles	416690 cycles	1.05

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: 30d236c	Previous: 1736882	Ratio
ML-DSA-44 keypair	314316 cycles	320637 cycles	0.98
ML-DSA-44 sign	1232665 cycles	1259345 cycles	0.98
ML-DSA-44 verify	363930 cycles	352392 cycles	1.03
ML-DSA-65 keypair	577868 cycles	586954 cycles	0.98
ML-DSA-65 sign	1982825 cycles	2044076 cycles	0.97
ML-DSA-65 verify	563056 cycles	569356 cycles	0.99
ML-DSA-87 keypair	911438 cycles	889842 cycles	1.02
ML-DSA-87 sign	2621391 cycles	2589004 cycles	1.01
ML-DSA-87 verify	953245 cycles	923943 cycles	1.03

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: 30d236c	Previous: 1736882	Ratio
ML-DSA-44 verify	363930 cycles	352392 cycles	1.03
ML-DSA-87 verify	953245 cycles	923943 cycles	1.03

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