SLOTHY: Superoptimize `poly_pointwise_montgomery` #776

mkannwischer · 2025-12-08T10:18:31Z

Resolves SLOTHY: Apply to pointwise_montgomery #758

Platform	Before	After	Speedup
Apple M1	114	114	1.00x
Graviton4 (Neoverse V2)	227	192	1.18x
Graviton3 (Neoverse V1)	252	205	1.23x
Cortex-A76 (RPi5)	407	401	1.01x
Graviton2 (Neoverse N1)	407	401	1.01x
Cortex-A72 (RPi4)	599	423	1.42x
Cortex-A55 (Snapdragon 888)	1292	792	1.63x

github-actions

Mac Mini (M1, 2020) benchmarks (opt)

Details

Benchmark suite	Current: `b5e75c6`	Previous: `734e659`	Ratio
`ML-DSA-44 keypair`	`46411` cycles	`46405` cycles	`1.00`
`ML-DSA-44 sign`	`131850` cycles	`131836` cycles	`1.00`
`ML-DSA-44 verify`	`47783` cycles	`47781` cycles	`1.00`
`ML-DSA-65 keypair`	`81330` cycles	`81324` cycles	`1.00`
`ML-DSA-65 sign`	`218016` cycles	`218008` cycles	`1.00`
`ML-DSA-65 verify`	`80041` cycles	`80045` cycles	`1.00`
`ML-DSA-87 keypair`	`132430` cycles	`132426` cycles	`1.00`
`ML-DSA-87 sign`	`279453` cycles	`279510` cycles	`1.00`
`ML-DSA-87 verify`	`130428` cycles	`130426` cycles	`1.00`

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

github-actions

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

Details

Benchmark suite	Current: `b5e75c6`	Previous: `734e659`	Ratio
`ML-DSA-44 keypair`	`113902` cycles	`114057` cycles	`1.00`
`ML-DSA-44 sign`	`360807` cycles	`361097` cycles	`1.00`
`ML-DSA-44 verify`	`118231` cycles	`118126` cycles	`1.00`
`ML-DSA-65 keypair`	`197903` cycles	`197782` cycles	`1.00`
`ML-DSA-65 sign`	`595056` cycles	`595721` cycles	`1.00`
`ML-DSA-65 verify`	`194425` cycles	`194600` cycles	`1.00`
`ML-DSA-87 keypair`	`323823` cycles	`324160` cycles	`1.00`
`ML-DSA-87 sign`	`760988` cycles	`761901` cycles	`1.00`
`ML-DSA-87 verify`	`320283` cycles	`320590` cycles	`1.00`

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

github-actions

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

Details

Benchmark suite	Current: `b5e75c6`	Previous: `734e659`	Ratio
`ML-DSA-44 keypair`	`114357` cycles	`114343` cycles	`1.00`
`ML-DSA-44 sign`	`428819` cycles	`428765` cycles	`1.00`
`ML-DSA-44 verify`	`121526` cycles	`121525` cycles	`1.00`
`ML-DSA-65 keypair`	`195860` cycles	`195902` cycles	`1.00`
`ML-DSA-65 sign`	`697653` cycles	`697579` cycles	`1.00`
`ML-DSA-65 verify`	`196423` cycles	`196436` cycles	`1.00`
`ML-DSA-87 keypair`	`322701` cycles	`322728` cycles	`1.00`
`ML-DSA-87 sign`	`879883` cycles	`880174` cycles	`1.00`
`ML-DSA-87 verify`	`326619` cycles	`326631` 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)

Details

Benchmark suite	Current: `b5e75c6`	Previous: `734e659`	Ratio
`ML-DSA-44 keypair`	`213998` cycles	`213790` cycles	`1.00`
`ML-DSA-44 sign`	`781509` cycles	`782853` cycles	`1.00`
`ML-DSA-44 verify`	`230072` cycles	`229992` cycles	`1.00`
`ML-DSA-65 keypair`	`384966` cycles	`384750` cycles	`1.00`
`ML-DSA-65 sign`	`1307979` cycles	`1314587` cycles	`0.99`
`ML-DSA-65 verify`	`375904` cycles	`375582` cycles	`1.00`
`ML-DSA-87 keypair`	`607170` cycles	`606698` cycles	`1.00`
`ML-DSA-87 sign`	`1621914` cycles	`1624428` cycles	`1.00`
`ML-DSA-87 verify`	`618098` cycles	`617520` cycles	`1.00`

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

oqs-bot

Intel Xeon 4th gen (c7i)

Details

Benchmark suite	Current: `b5e75c6`	Previous: `734e659`	Ratio
`ML-DSA-44 keypair`	`35109` cycles	`35055` cycles	`1.00`
`ML-DSA-44 sign`	`120885` cycles	`120076` cycles	`1.01`
`ML-DSA-44 verify`	`38302` cycles	`38175` cycles	`1.00`
`ML-DSA-65 keypair`	`61762` cycles	`60985` cycles	`1.01`
`ML-DSA-65 sign`	`201531` cycles	`199714` cycles	`1.01`
`ML-DSA-65 verify`	`63258` cycles	`63081` cycles	`1.00`
`ML-DSA-87 keypair`	`93583` cycles	`93419` cycles	`1.00`
`ML-DSA-87 sign`	`230820` cycles	`229965` cycles	`1.00`
`ML-DSA-87 verify`	`94268` cycles	`94960` cycles	`0.99`

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

oqs-bot

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

Details

Benchmark suite	Current: `b5e75c6`	Previous: `734e659`	Ratio
`ML-DSA-44 keypair`	`96057` cycles	`96412` cycles	`1.00`
`ML-DSA-44 sign`	`348605` cycles	`348592` cycles	`1.00`
`ML-DSA-44 verify`	`101574` cycles	`101686` cycles	`1.00`
`ML-DSA-65 keypair`	`163167` cycles	`163862` cycles	`1.00`
`ML-DSA-65 sign`	`563855` cycles	`564967` cycles	`1.00`
`ML-DSA-65 verify`	`165717` cycles	`166173` cycles	`1.00`
`ML-DSA-87 keypair`	`267263` cycles	`267167` cycles	`1.00`
`ML-DSA-87 sign`	`723063` cycles	`722495` cycles	`1.00`
`ML-DSA-87 verify`	`272737` cycles	`272915` cycles	`1.00`

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

oqs-bot

Intel Xeon 3rd gen (c6i)

Details

Benchmark suite	Current: `b5e75c6`	Previous: `734e659`	Ratio
`ML-DSA-44 keypair`	`56775` cycles	`56653` cycles	`1.00`
`ML-DSA-44 sign`	`179913` cycles	`179739` cycles	`1.00`
`ML-DSA-44 verify`	`61248` cycles	`61065` cycles	`1.00`
`ML-DSA-65 keypair`	`99300` cycles	`99496` cycles	`1.00`
`ML-DSA-65 sign`	`295966` cycles	`295892` cycles	`1.00`
`ML-DSA-65 verify`	`100984` cycles	`100885` cycles	`1.00`
`ML-DSA-87 keypair`	`153137` cycles	`153188` cycles	`1.00`
`ML-DSA-87 sign`	`352895` cycles	`353164` cycles	`1.00`
`ML-DSA-87 verify`	`153238` cycles	`154165` cycles	`0.99`

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

oqs-bot

Graviton4

Details

Benchmark suite	Current: `b5e75c6`	Previous: `734e659`	Ratio
`ML-DSA-44 keypair`	`68708` cycles	`68720` cycles	`1.00`
`ML-DSA-44 sign`	`202506` cycles	`202847` cycles	`1.00`
`ML-DSA-44 verify`	`70810` cycles	`70716` cycles	`1.00`
`ML-DSA-65 keypair`	`121830` cycles	`121547` cycles	`1.00`
`ML-DSA-65 sign`	`332532` cycles	`332849` cycles	`1.00`
`ML-DSA-65 verify`	`117902` cycles	`117943` cycles	`1.00`
`ML-DSA-87 keypair`	`198843` cycles	`199110` cycles	`1.00`
`ML-DSA-87 sign`	`427849` cycles	`429525` cycles	`1.00`
`ML-DSA-87 verify`	`194435` cycles	`195097` cycles	`1.00`

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

oqs-bot

AMD EPYC 3rd gen (c6a)

Details

Benchmark suite	Current: `b5e75c6`	Previous: `734e659`	Ratio
`ML-DSA-44 keypair`	`69528` cycles	`69766` cycles	`1.00`
`ML-DSA-44 sign`	`185664` cycles	`186080` cycles	`1.00`
`ML-DSA-44 verify`	`69225` cycles	`69350` cycles	`1.00`
`ML-DSA-65 keypair`	`119326` cycles	`119237` cycles	`1.00`
`ML-DSA-65 sign`	`295611` cycles	`296114` cycles	`1.00`
`ML-DSA-65 verify`	`115395` cycles	`115290` cycles	`1.00`
`ML-DSA-87 keypair`	`204016` cycles	`204149` cycles	`1.00`
`ML-DSA-87 sign`	`387327` cycles	`388103` cycles	`1.00`
`ML-DSA-87 verify`	`195382` cycles	`195578` cycles	`1.00`

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

oqs-bot

Graviton2

Details

Benchmark suite	Current: `b5e75c6`	Previous: `734e659`	Ratio
`ML-DSA-44 keypair`	`114192` cycles	`114200` cycles	`1.00`
`ML-DSA-44 sign`	`361340` cycles	`361473` cycles	`1.00`
`ML-DSA-44 verify`	`118590` cycles	`118219` cycles	`1.00`
`ML-DSA-65 keypair`	`198088` cycles	`197626` cycles	`1.00`
`ML-DSA-65 sign`	`595916` cycles	`596017` cycles	`1.00`
`ML-DSA-65 verify`	`194643` cycles	`194649` cycles	`1.00`
`ML-DSA-87 keypair`	`324229` cycles	`324677` cycles	`1.00`
`ML-DSA-87 sign`	`761648` cycles	`762322` cycles	`1.00`
`ML-DSA-87 verify`	`320253` cycles	`320892` cycles	`1.00`

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

oqs-bot

Graviton4 (no-opt)

Details

Benchmark suite	Current: `b5e75c6`	Previous: `734e659`	Ratio
`ML-DSA-44 keypair`	`128188` cycles	`128130` cycles	`1.00`
`ML-DSA-44 sign`	`457053` cycles	`456746` cycles	`1.00`
`ML-DSA-44 verify`	`136365` cycles	`136304` cycles	`1.00`
`ML-DSA-65 keypair`	`220489` cycles	`220417` cycles	`1.00`
`ML-DSA-65 sign`	`746830` cycles	`746762` cycles	`1.00`
`ML-DSA-65 verify`	`220381` cycles	`220665` cycles	`1.00`
`ML-DSA-87 keypair`	`364957` cycles	`365265` cycles	`1.00`
`ML-DSA-87 sign`	`943843` cycles	`943612` cycles	`1.00`
`ML-DSA-87 verify`	`368948` cycles	`369289` cycles	`1.00`

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

oqs-bot

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

Details

Benchmark suite	Current: `b5e75c6`	Previous: `734e659`	Ratio
`ML-DSA-44 keypair`	`158538` cycles	`159895` cycles	`0.99`
`ML-DSA-44 sign`	`566634` cycles	`569072` cycles	`1.00`
`ML-DSA-44 verify`	`170913` cycles	`171381` cycles	`1.00`
`ML-DSA-65 keypair`	`269707` cycles	`269919` cycles	`1.00`
`ML-DSA-65 sign`	`927910` cycles	`926921` cycles	`1.00`
`ML-DSA-65 verify`	`276456` cycles	`276464` cycles	`1.00`
`ML-DSA-87 keypair`	`451154` cycles	`451080` cycles	`1.00`
`ML-DSA-87 sign`	`1185537` cycles	`1184065` cycles	`1.00`
`ML-DSA-87 verify`	`460381` cycles	`461453` cycles	`1.00`

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

oqs-bot

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

Details

Benchmark suite	Current: `b5e75c6`	Previous: `734e659`	Ratio
`ML-DSA-44 keypair`	`135702` cycles	`136159` cycles	`1.00`
`ML-DSA-44 sign`	`541234` cycles	`540737` cycles	`1.00`
`ML-DSA-44 verify`	`149583` cycles	`148997` cycles	`1.00`
`ML-DSA-65 keypair`	`228441` cycles	`230145` cycles	`0.99`
`ML-DSA-65 sign`	`891398` cycles	`891208` cycles	`1.00`
`ML-DSA-65 verify`	`237766` cycles	`238434` cycles	`1.00`
`ML-DSA-87 keypair`	`373239` cycles	`373790` cycles	`1.00`
`ML-DSA-87 sign`	`1105668` cycles	`1107432` cycles	`1.00`
`ML-DSA-87 verify`	`387043` cycles	`387838` cycles	`1.00`

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

oqs-bot

Graviton3

Details

Benchmark suite	Current: `b5e75c6`	Previous: `734e659`	Ratio
`ML-DSA-44 keypair`	`72561` cycles	`72683` cycles	`1.00`
`ML-DSA-44 sign`	`212290` cycles	`213170` cycles	`1.00`
`ML-DSA-44 verify`	`75404` cycles	`75550` cycles	`1.00`
`ML-DSA-65 keypair`	`128326` cycles	`128392` cycles	`1.00`
`ML-DSA-65 sign`	`352468` cycles	`351759` cycles	`1.00`
`ML-DSA-65 verify`	`125272` cycles	`125537` cycles	`1.00`
`ML-DSA-87 keypair`	`206542` cycles	`209460` cycles	`0.99`
`ML-DSA-87 sign`	`443930` cycles	`451569` cycles	`0.98`
`ML-DSA-87 verify`	`204094` cycles	`205475` cycles	`0.99`

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

github-actions

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

Details

Benchmark suite	Current: `b5e75c6`	Previous: `734e659`	Ratio
`ML-DSA-44 keypair`	`230184` cycles	`224135` cycles	`1.03`
`ML-DSA-44 sign`	`632725` cycles	`635328` cycles	`1.00`
`ML-DSA-44 verify`	`218782` cycles	`220287` cycles	`0.99`
`ML-DSA-65 keypair`	`394325` cycles	`420149` cycles	`0.94`
`ML-DSA-65 sign`	`1030974` cycles	`1100994` cycles	`0.94`
`ML-DSA-65 verify`	`371518` cycles	`405803` cycles	`0.92`
`ML-DSA-87 keypair`	`657576` cycles	`676985` cycles	`0.97`
`ML-DSA-87 sign`	`1387619` cycles	`1424962` cycles	`0.97`
`ML-DSA-87 verify`	`637173` cycles	`655852` cycles	`0.97`

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

oqs-bot

AMD EPYC 4th gen (c7a)

Details

Benchmark suite	Current: `b5e75c6`	Previous: `734e659`	Ratio
`ML-DSA-44 keypair`	`48131` cycles	`40859` cycles	`1.18`
`ML-DSA-44 sign`	`152637` cycles	`129191` cycles	`1.18`
`ML-DSA-44 verify`	`51075` cycles	`43362` cycles	`1.18`
`ML-DSA-65 keypair`	`86661` cycles	`72744` cycles	`1.19`
`ML-DSA-65 sign`	`254228` cycles	`210841` cycles	`1.21`
`ML-DSA-65 verify`	`86937` cycles	`73036` cycles	`1.19`
`ML-DSA-87 keypair`	`127897` cycles	`108742` cycles	`1.18`
`ML-DSA-87 sign`	`293770` cycles	`254164` cycles	`1.16`
`ML-DSA-87 verify`	`129788` cycles	`110933` cycles	`1.17`

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

oqs-bot

⚠️ Performance Alert ⚠️

Possible performance regression was detected for benchmark 'AMD EPYC 4th gen (c7a)'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.03.

Benchmark suite	Current: `b5e75c6`	Previous: `734e659`	Ratio
`ML-DSA-44 keypair`	`48131` cycles	`40859` cycles	`1.18`
`ML-DSA-44 sign`	`152637` cycles	`129191` cycles	`1.18`
`ML-DSA-44 verify`	`51075` cycles	`43362` cycles	`1.18`
`ML-DSA-65 keypair`	`86661` cycles	`72744` cycles	`1.19`
`ML-DSA-65 sign`	`254228` cycles	`210841` cycles	`1.21`
`ML-DSA-65 verify`	`86937` cycles	`73036` cycles	`1.19`
`ML-DSA-87 keypair`	`127897` cycles	`108742` cycles	`1.18`
`ML-DSA-87 sign`	`293770` cycles	`254164` cycles	`1.16`
`ML-DSA-87 verify`	`129788` cycles	`110933` cycles	`1.17`

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

oqs-bot

Graviton2 (no-opt)

Details

Benchmark suite	Current: `b5e75c6`	Previous: `734e659`	Ratio
`ML-DSA-44 keypair`	`214300` cycles	`214113` cycles	`1.00`
`ML-DSA-44 sign`	`782173` cycles	`782562` cycles	`1.00`
`ML-DSA-44 verify`	`240269` cycles	`230218` cycles	`1.04`
`ML-DSA-65 keypair`	`385639` cycles	`384979` cycles	`1.00`
`ML-DSA-65 sign`	`1307966` cycles	`1310083` cycles	`1.00`
`ML-DSA-65 verify`	`376365` cycles	`375686` cycles	`1.00`
`ML-DSA-87 keypair`	`608590` cycles	`607564` cycles	`1.00`
`ML-DSA-87 sign`	`1625165` cycles	`1627404` cycles	`1.00`
`ML-DSA-87 verify`	`618856` cycles	`618004` 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 'Graviton2 (no-opt)'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.03.

Benchmark suite	Current: `b5e75c6`	Previous: `734e659`	Ratio
`ML-DSA-44 verify`	`240269` cycles	`230218` cycles	`1.04`

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

oqs-bot

Graviton3 (no-opt)

Details

Benchmark suite	Current: `b5e75c6`	Previous: `734e659`	Ratio
`ML-DSA-44 keypair`	`138884` cycles	`138812` cycles	`1.00`
`ML-DSA-44 sign`	`493574` cycles	`493965` cycles	`1.00`
`ML-DSA-44 verify`	`148553` cycles	`148433` cycles	`1.00`
`ML-DSA-65 keypair`	`242227` cycles	`242326` cycles	`1.00`
`ML-DSA-65 sign`	`809171` cycles	`808958` cycles	`1.00`
`ML-DSA-65 verify`	`240494` cycles	`240933` cycles	`1.00`
`ML-DSA-87 keypair`	`396232` cycles	`396993` cycles	`1.00`
`ML-DSA-87 sign`	`1026953` cycles	`1027054` cycles	`1.00`
`ML-DSA-87 verify`	`401247` cycles	`401855` cycles	`1.00`

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

oqs-bot

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

Details

Benchmark suite	Current: `b5e75c6`	Previous: `734e659`	Ratio
`ML-DSA-44 keypair`	`142330` cycles	`120621` cycles	`1.18`
`ML-DSA-44 sign`	`537540` cycles	`453749` cycles	`1.18`
`ML-DSA-44 verify`	`154245` cycles	`130830` cycles	`1.18`
`ML-DSA-65 keypair`	`244100` cycles	`204679` cycles	`1.19`
`ML-DSA-65 sign`	`876924` cycles	`737165` cycles	`1.19`
`ML-DSA-65 verify`	`251506` cycles	`210269` cycles	`1.20`
`ML-DSA-87 keypair`	`403721` cycles	`336824` cycles	`1.20`
`ML-DSA-87 sign`	`1105353` cycles	`928462` cycles	`1.19`
`ML-DSA-87 verify`	`412145` cycles	`345933` cycles	`1.19`

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

oqs-bot

⚠️ Performance Alert ⚠️

Possible performance regression was detected for benchmark 'AMD EPYC 4th gen (c7a) (no-opt)'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.03.

Benchmark suite	Current: `b5e75c6`	Previous: `734e659`	Ratio
`ML-DSA-44 keypair`	`142330` cycles	`120621` cycles	`1.18`
`ML-DSA-44 sign`	`537540` cycles	`453749` cycles	`1.18`
`ML-DSA-44 verify`	`154245` cycles	`130830` cycles	`1.18`
`ML-DSA-65 keypair`	`244100` cycles	`204679` cycles	`1.19`
`ML-DSA-65 sign`	`876924` cycles	`737165` cycles	`1.19`
`ML-DSA-65 verify`	`251506` cycles	`210269` cycles	`1.20`
`ML-DSA-87 keypair`	`403721` cycles	`336824` cycles	`1.20`
`ML-DSA-87 sign`	`1105353` cycles	`928462` cycles	`1.19`
`ML-DSA-87 verify`	`412145` cycles	`345933` cycles	`1.19`

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

github-actions

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

Details

Benchmark suite	Current: `b5e75c6`	Previous: `734e659`	Ratio
`ML-DSA-44 keypair`	`828442` cycles	`830982` cycles	`1.00`
`ML-DSA-44 sign`	`3341456` cycles	`3353667` cycles	`1.00`
`ML-DSA-44 verify`	`919643` cycles	`923303` cycles	`1.00`
`ML-DSA-65 keypair`	`1409261` cycles	`1409619` cycles	`1.00`
`ML-DSA-65 sign`	`5467305` cycles	`5485141` cycles	`1.00`
`ML-DSA-65 verify`	`1473629` cycles	`1474684` cycles	`1.00`
`ML-DSA-87 keypair`	`2310003` cycles	`2320095` cycles	`1.00`
`ML-DSA-87 sign`	`6818820` cycles	`6860135` cycles	`0.99`
`ML-DSA-87 verify`	`2408373` cycles	`2416540` 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)

Details

Benchmark suite	Current: `b5e75c6`	Previous: `734e659`	Ratio
`ML-DSA-44 keypair`	`326735` cycles	`321112` cycles	`1.02`
`ML-DSA-44 sign`	`1220151` cycles	`1243311` cycles	`0.98`
`ML-DSA-44 verify`	`346125` cycles	`339358` cycles	`1.02`
`ML-DSA-65 keypair`	`581130` cycles	`583644` cycles	`1.00`
`ML-DSA-65 sign`	`2021227` cycles	`1981964` cycles	`1.02`
`ML-DSA-65 verify`	`556290` cycles	`549389` cycles	`1.01`
`ML-DSA-87 keypair`	`929764` cycles	`907443` cycles	`1.02`
`ML-DSA-87 sign`	`2652043` cycles	`2537151` cycles	`1.05`
`ML-DSA-87 verify`	`945364` cycles	`907064` cycles	`1.04`

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

- poly_pointwise_montgomery - poly_caddq - poly_chknorm - poly_decompose - poly_use_hint - polyz_unpack Signed-off-by: Matthias J. Kannwischer <[email protected]>

github-actions

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

Details

Benchmark suite	Current: `b5e75c6`	Previous: `734e659`	Ratio
`ML-DSA-44 keypair`	`278567` cycles	`275602` cycles	`1.01`
`ML-DSA-44 sign`	`827923` cycles	`830077` cycles	`1.00`
`ML-DSA-44 verify`	`273872` cycles	`273817` cycles	`1.00`
`ML-DSA-65 keypair`	`481193` cycles	`472620` cycles	`1.02`
`ML-DSA-65 sign`	`1357802` cycles	`1358716` cycles	`1.00`
`ML-DSA-65 verify`	`452375` cycles	`451935` cycles	`1.00`
`ML-DSA-87 keypair`	`819880` cycles	`802150` cycles	`1.02`
`ML-DSA-87 sign`	`1878166` cycles	`1824003` cycles	`1.03`
`ML-DSA-87 verify`	`783426` cycles	`772590` cycles	`1.01`

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

Resolves #758 Signed-off-by: Matthias J. Kannwischer <[email protected]>

github-actions

⚠️ Performance Alert ⚠️

Possible performance regression was detected for benchmark 'Arm Cortex-A55 (Snapdragon 888) benchmarks (opt)'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.03.

Benchmark suite	Current: `b5e75c6`	Previous: `734e659`	Ratio
`ML-DSA-87 keypair`	`832780` cycles	`802150` cycles	`1.04`
`ML-DSA-87 sign`	`1918455` cycles	`1824003` cycles	`1.05`

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

github-actions

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

Details

Benchmark suite	Current: `b5e75c6`	Previous: `734e659`	Ratio
`ML-DSA-44 keypair`	`467239` cycles	`464519` cycles	`1.01`
`ML-DSA-44 sign`	`2212586` cycles	`2207891` cycles	`1.00`
`ML-DSA-44 verify`	`550919` cycles	`545586` cycles	`1.01`
`ML-DSA-65 keypair`	`783515` cycles	`778580` cycles	`1.01`
`ML-DSA-65 sign`	`3627885` cycles	`3615551` cycles	`1.00`
`ML-DSA-65 verify`	`852365` cycles	`847909` cycles	`1.01`
`ML-DSA-87 keypair`	`1274282` cycles	`1254986` cycles	`1.02`
`ML-DSA-87 sign`	`4520871` cycles	`4453037` cycles	`1.02`
`ML-DSA-87 verify`	`1374939` cycles	`1361356` 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 (no-opt)'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.03.

Benchmark suite	Current: `b5e75c6`	Previous: `734e659`	Ratio
`ML-DSA-87 sign`	`2652043` cycles	`2537151` cycles	`1.05`
`ML-DSA-87 verify`	`945364` cycles	`907064` cycles	`1.04`

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

mkannwischer added the benchmark label Dec 8, 2025