Constant-time hardening for `caddq`, `poly_chknorm`, `decompose` by mkannwischer · Pull Request #371 · pq-code-package/mldsa-native

mkannwischer · 2025-07-16T11:55:15Z

Resolves Introduce constant-time barriers for caddq, decompose, and power2round #364
Resolves Introduce constant-time primitive for absolute value computation #152

This PR adds constant-time value barriers to 3 functions: caddq, poly_chknorm, decompose.
The same approach as in mlkem-native is followed: If available, by default value barriers are implemeted using an empty inline asm block which marks the target value as clobbered. If inline assembly is not available (e.g., on Windows), we XOR with the value of a volatile global that's set to 0 (as proposed by Bernstein here).

CBMC proofs are added for the added auxilary functions and the proofs of caddq, poly_chknorm, decompose. A config variation test is added to CI to test the volatile global optblocker config option.

Performance impact is considerable (up to 20% when using link-time optimization).

github-actions

Mac Mini (M1, 2020) benchmarks (opt)

Details

Benchmark suite	Current: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 keypair`	`50488` cycles	`48802` cycles	`1.03`
`ML-DSA-44 sign`	`215075` cycles	`182512` cycles	`1.18`
`ML-DSA-44 verify`	`72446` cycles	`67786` cycles	`1.07`
`ML-DSA-65 keypair`	`87392` cycles	`84879` cycles	`1.03`
`ML-DSA-65 sign`	`343568` cycles	`290742` cycles	`1.18`
`ML-DSA-65 verify`	`112046` cycles	`105851` cycles	`1.06`
`ML-DSA-87 keypair`	`140356` cycles	`136419` cycles	`1.03`
`ML-DSA-87 sign`	`412988` cycles	`355398` cycles	`1.16`
`ML-DSA-87 verify`	`172917` cycles	`163912` cycles	`1.05`

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

github-actions

⚠️ Performance Alert ⚠️

Possible performance regression was detected for benchmark 'Mac Mini (M1, 2020) benchmarks (opt)'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.03.

Benchmark suite	Current: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 keypair`	`50488` cycles	`48802` cycles	`1.03`
`ML-DSA-44 sign`	`215075` cycles	`182512` cycles	`1.18`
`ML-DSA-44 verify`	`72446` cycles	`67786` cycles	`1.07`
`ML-DSA-65 sign`	`343568` cycles	`290742` cycles	`1.18`
`ML-DSA-65 verify`	`112046` cycles	`105851` cycles	`1.06`
`ML-DSA-87 sign`	`412988` cycles	`355398` cycles	`1.16`
`ML-DSA-87 verify`	`172917` cycles	`163912` cycles	`1.05`

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

github-actions

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

Details

Benchmark suite	Current: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 keypair`	`116010` cycles	`114258` cycles	`1.02`
`ML-DSA-44 sign`	`447658` cycles	`414438` cycles	`1.08`
`ML-DSA-44 verify`	`136472` cycles	`131651` cycles	`1.04`
`ML-DSA-65 keypair`	`197953` cycles	`195527` cycles	`1.01`
`ML-DSA-65 sign`	`721048` cycles	`668417` cycles	`1.08`
`ML-DSA-65 verify`	`216225` cycles	`210000` cycles	`1.03`
`ML-DSA-87 keypair`	`325280` cycles	`321618` cycles	`1.01`
`ML-DSA-87 sign`	`902001` cycles	`844770` cycles	`1.07`
`ML-DSA-87 verify`	`352663` cycles	`343995` 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 'Mac Mini (M1, 2020) benchmarks (no-opt)'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.03.

Benchmark suite	Current: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 sign`	`447658` cycles	`414438` cycles	`1.08`
`ML-DSA-44 verify`	`136472` cycles	`131651` cycles	`1.04`
`ML-DSA-65 sign`	`721048` cycles	`668417` cycles	`1.08`
`ML-DSA-87 sign`	`902001` cycles	`844770` cycles	`1.07`

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

github-actions

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

Details

Benchmark suite	Current: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 keypair`	`292987` cycles	`285837` cycles	`1.03`
`ML-DSA-44 sign`	`1182571` cycles	`1017744` cycles	`1.16`
`ML-DSA-44 verify`	`352257` cycles	`327753` cycles	`1.07`
`ML-DSA-65 keypair`	`496703` cycles	`487572` cycles	`1.02`
`ML-DSA-65 sign`	`1906890` cycles	`1663334` cycles	`1.15`
`ML-DSA-65 verify`	`548650` cycles	`513217` cycles	`1.07`
`ML-DSA-87 keypair`	`836385` cycles	`825275` cycles	`1.01`
`ML-DSA-87 sign`	`2452000` cycles	`2186302` cycles	`1.12`
`ML-DSA-87 verify`	`906276` cycles	`856528` cycles	`1.06`

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

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: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 sign`	`1182571` cycles	`1017744` cycles	`1.16`
`ML-DSA-44 verify`	`352257` cycles	`327753` cycles	`1.07`
`ML-DSA-65 sign`	`1906890` cycles	`1663334` cycles	`1.15`
`ML-DSA-65 verify`	`548650` cycles	`513217` cycles	`1.07`
`ML-DSA-87 sign`	`2452000` cycles	`2186302` cycles	`1.12`
`ML-DSA-87 verify`	`906276` cycles	`856528` cycles	`1.06`

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

oqs-bot

Intel Xeon 4th gen (c7i)

Details

Benchmark suite	Current: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 keypair`	`37542` cycles	`36230` cycles	`1.04`
`ML-DSA-44 sign`	`166790` cycles	`138487` cycles	`1.20`
`ML-DSA-44 verify`	`50124` cycles	`46045` cycles	`1.09`
`ML-DSA-65 keypair`	`65625` cycles	`64566` cycles	`1.02`
`ML-DSA-65 sign`	`274615` cycles	`230597` cycles	`1.19`
`ML-DSA-65 verify`	`78471` cycles	`71433` cycles	`1.10`
`ML-DSA-87 keypair`	`101090` cycles	`98741` cycles	`1.02`
`ML-DSA-87 sign`	`320494` cycles	`274945` cycles	`1.17`
`ML-DSA-87 verify`	`116555` cycles	`107165` cycles	`1.09`

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

oqs-bot

⚠️ Performance Alert ⚠️

Possible performance regression was detected for benchmark 'Intel Xeon 4th gen (c7i)'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.03.

Benchmark suite	Current: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 keypair`	`37542` cycles	`36230` cycles	`1.04`
`ML-DSA-44 sign`	`166790` cycles	`138487` cycles	`1.20`
`ML-DSA-44 verify`	`50124` cycles	`46045` cycles	`1.09`
`ML-DSA-65 sign`	`274615` cycles	`230597` cycles	`1.19`
`ML-DSA-65 verify`	`78471` cycles	`71433` cycles	`1.10`
`ML-DSA-87 sign`	`320494` cycles	`274945` cycles	`1.17`
`ML-DSA-87 verify`	`116555` cycles	`107165` cycles	`1.09`

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

oqs-bot

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

Details

Benchmark suite	Current: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 keypair`	`96155` cycles	`95582` cycles	`1.01`
`ML-DSA-44 sign`	`347296` cycles	`322946` cycles	`1.08`
`ML-DSA-44 verify`	`104855` cycles	`101532` cycles	`1.03`
`ML-DSA-65 keypair`	`164112` cycles	`162604` cycles	`1.01`
`ML-DSA-65 sign`	`568798` cycles	`522372` cycles	`1.09`
`ML-DSA-65 verify`	`168809` cycles	`162706` cycles	`1.04`
`ML-DSA-87 keypair`	`276032` cycles	`273426` cycles	`1.01`
`ML-DSA-87 sign`	`724967` cycles	`673052` cycles	`1.08`
`ML-DSA-87 verify`	`281498` cycles	`271672` cycles	`1.04`

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

oqs-bot

⚠️ Performance Alert ⚠️

Possible performance regression was detected for benchmark 'Intel Xeon 4th gen (c7i) (no-opt)'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.03.

Benchmark suite	Current: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 sign`	`347296` cycles	`322946` cycles	`1.08`
`ML-DSA-44 verify`	`104855` cycles	`101532` cycles	`1.03`
`ML-DSA-65 sign`	`568798` cycles	`522372` cycles	`1.09`
`ML-DSA-65 verify`	`168809` cycles	`162706` cycles	`1.04`
`ML-DSA-87 sign`	`724967` cycles	`673052` cycles	`1.08`
`ML-DSA-87 verify`	`281498` cycles	`271672` cycles	`1.04`

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

oqs-bot

AMD EPYC 3rd gen (c6a)

Details

Benchmark suite	Current: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 keypair`	`76926` cycles	`72434` cycles	`1.06`
`ML-DSA-44 sign`	`278561` cycles	`224246` cycles	`1.24`
`ML-DSA-44 verify`	`92289` cycles	`83522` cycles	`1.10`
`ML-DSA-65 keypair`	`126427` cycles	`124793` cycles	`1.01`
`ML-DSA-65 sign`	`429590` cycles	`362808` cycles	`1.18`
`ML-DSA-65 verify`	`142678` cycles	`132717` cycles	`1.08`
`ML-DSA-87 keypair`	`210652` cycles	`207762` cycles	`1.01`
`ML-DSA-87 sign`	`541159` cycles	`464500` cycles	`1.17`
`ML-DSA-87 verify`	`229863` cycles	`216400` cycles	`1.06`

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: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 keypair`	`76926` cycles	`72434` cycles	`1.06`
`ML-DSA-44 sign`	`278561` cycles	`224246` cycles	`1.24`
`ML-DSA-44 verify`	`92289` cycles	`83522` cycles	`1.10`
`ML-DSA-65 sign`	`429590` cycles	`362808` cycles	`1.18`
`ML-DSA-65 verify`	`142678` cycles	`132717` cycles	`1.08`
`ML-DSA-87 sign`	`541159` cycles	`464500` cycles	`1.17`
`ML-DSA-87 verify`	`229863` cycles	`216400` cycles	`1.06`

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

oqs-bot

Intel Xeon 3rd gen (c6i)

Details

Benchmark suite	Current: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 keypair`	`61162` cycles	`59212` cycles	`1.03`
`ML-DSA-44 sign`	`259898` cycles	`211541` cycles	`1.23`
`ML-DSA-44 verify`	`81508` cycles	`73884` cycles	`1.10`
`ML-DSA-65 keypair`	`106825` cycles	`104326` cycles	`1.02`
`ML-DSA-65 sign`	`426164` cycles	`347410` cycles	`1.23`
`ML-DSA-65 verify`	`126994` cycles	`115218` cycles	`1.10`
`ML-DSA-87 keypair`	`165860` cycles	`162599` cycles	`1.02`
`ML-DSA-87 sign`	`510581` cycles	`421422` cycles	`1.21`
`ML-DSA-87 verify`	`191919` cycles	`177621` cycles	`1.08`

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

oqs-bot

⚠️ Performance Alert ⚠️

Possible performance regression was detected for benchmark 'Intel Xeon 3rd gen (c6i)'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.03.

Benchmark suite	Current: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 keypair`	`61162` cycles	`59212` cycles	`1.03`
`ML-DSA-44 sign`	`259898` cycles	`211541` cycles	`1.23`
`ML-DSA-44 verify`	`81508` cycles	`73884` cycles	`1.10`
`ML-DSA-65 sign`	`426164` cycles	`347410` cycles	`1.23`
`ML-DSA-65 verify`	`126994` cycles	`115218` cycles	`1.10`
`ML-DSA-87 sign`	`510581` cycles	`421422` cycles	`1.21`
`ML-DSA-87 verify`	`191919` cycles	`177621` cycles	`1.08`

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

oqs-bot

Graviton4

Details

Benchmark suite	Current: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 keypair`	`72951` cycles	`71188` cycles	`1.02`
`ML-DSA-44 sign`	`275790` cycles	`236961` cycles	`1.16`
`ML-DSA-44 verify`	`86426` cycles	`81190` cycles	`1.06`
`ML-DSA-65 keypair`	`128574` cycles	`125629` cycles	`1.02`
`ML-DSA-65 sign`	`448212` cycles	`384709` cycles	`1.17`
`ML-DSA-65 verify`	`138566` cycles	`131226` cycles	`1.06`
`ML-DSA-87 keypair`	`208342` cycles	`203823` cycles	`1.02`
`ML-DSA-87 sign`	`551582` cycles	`486383` cycles	`1.13`
`ML-DSA-87 verify`	`221933` cycles	`212160` cycles	`1.05`

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

oqs-bot

⚠️ Performance Alert ⚠️

Possible performance regression was detected for benchmark 'Graviton4'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.03.

Benchmark suite	Current: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 sign`	`275790` cycles	`236961` cycles	`1.16`
`ML-DSA-44 verify`	`86426` cycles	`81190` cycles	`1.06`
`ML-DSA-65 sign`	`448212` cycles	`384709` cycles	`1.17`
`ML-DSA-65 verify`	`138566` cycles	`131226` cycles	`1.06`
`ML-DSA-87 sign`	`551582` cycles	`486383` cycles	`1.13`
`ML-DSA-87 verify`	`221933` cycles	`212160` 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: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 keypair`	`464821` cycles	`458473` cycles	`1.01`
`ML-DSA-44 sign`	`2196905` cycles	`2031159` cycles	`1.08`
`ML-DSA-44 verify`	`557100` cycles	`532359` cycles	`1.05`
`ML-DSA-65 keypair`	`779114` cycles	`769535` cycles	`1.01`
`ML-DSA-65 sign`	`3582406` cycles	`3353844` cycles	`1.07`
`ML-DSA-65 verify`	`862497` cycles	`827247` cycles	`1.04`
`ML-DSA-87 keypair`	`1256966` cycles	`1241079` cycles	`1.01`
`ML-DSA-87 sign`	`4412738` cycles	`4186890` cycles	`1.05`
`ML-DSA-87 verify`	`1377727` cycles	`1328345` cycles	`1.04`

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

github-actions

⚠️ Performance Alert ⚠️

Possible performance regression was detected for benchmark 'Arm Cortex-A55 (Snapdragon 888) benchmarks (no-opt)'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.03.

Benchmark suite	Current: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 sign`	`2196905` cycles	`2031159` cycles	`1.08`
`ML-DSA-44 verify`	`557100` cycles	`532359` cycles	`1.05`
`ML-DSA-65 sign`	`3582406` cycles	`3353844` cycles	`1.07`
`ML-DSA-65 verify`	`862497` cycles	`827247` cycles	`1.04`
`ML-DSA-87 sign`	`4412738` cycles	`4186890` cycles	`1.05`
`ML-DSA-87 verify`	`1377727` cycles	`1328345` cycles	`1.04`

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

oqs-bot

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

Details

Benchmark suite	Current: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 keypair`	`136633` cycles	`135071` cycles	`1.01`
`ML-DSA-44 sign`	`551099` cycles	`510595` cycles	`1.08`
`ML-DSA-44 verify`	`155069` cycles	`149244` cycles	`1.04`
`ML-DSA-65 keypair`	`228231` cycles	`226517` cycles	`1.01`
`ML-DSA-65 sign`	`886555` cycles	`820635` cycles	`1.08`
`ML-DSA-65 verify`	`244279` cycles	`234097` cycles	`1.04`
`ML-DSA-87 keypair`	`376203` cycles	`372954` cycles	`1.01`
`ML-DSA-87 sign`	`1115054` cycles	`1035711` cycles	`1.08`
`ML-DSA-87 verify`	`398395` cycles	`384434` cycles	`1.04`

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) (no-opt)'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.03.

Benchmark suite	Current: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 sign`	`551099` cycles	`510595` cycles	`1.08`
`ML-DSA-44 verify`	`155069` cycles	`149244` cycles	`1.04`
`ML-DSA-65 sign`	`886555` cycles	`820635` cycles	`1.08`
`ML-DSA-65 verify`	`244279` cycles	`234097` cycles	`1.04`
`ML-DSA-87 sign`	`1115054` cycles	`1035711` cycles	`1.08`
`ML-DSA-87 verify`	`398395` cycles	`384434` cycles	`1.04`

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

oqs-bot

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

Details

Benchmark suite	Current: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 keypair`	`159175` cycles	`157139` cycles	`1.01`
`ML-DSA-44 sign`	`569022` cycles	`521064` cycles	`1.09`
`ML-DSA-44 verify`	`174916` cycles	`167947` cycles	`1.04`
`ML-DSA-65 keypair`	`270585` cycles	`271243` cycles	`1.00`
`ML-DSA-65 sign`	`924086` cycles	`852337` cycles	`1.08`
`ML-DSA-65 verify`	`281964` cycles	`272681` cycles	`1.03`
`ML-DSA-87 keypair`	`452239` cycles	`448838` cycles	`1.01`
`ML-DSA-87 sign`	`1179542` cycles	`1089952` cycles	`1.08`
`ML-DSA-87 verify`	`468749` cycles	`451807` cycles	`1.04`

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

oqs-bot

⚠️ Performance Alert ⚠️

Possible performance regression was detected for benchmark 'Intel Xeon 3rd gen (c6i) (no-opt)'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.03.

Benchmark suite	Current: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 sign`	`569022` cycles	`521064` cycles	`1.09`
`ML-DSA-44 verify`	`174916` cycles	`167947` cycles	`1.04`
`ML-DSA-65 sign`	`924086` cycles	`852337` cycles	`1.08`
`ML-DSA-65 verify`	`281964` cycles	`272681` cycles	`1.03`
`ML-DSA-87 sign`	`1179542` cycles	`1089952` cycles	`1.08`
`ML-DSA-87 verify`	`468749` cycles	`451807` cycles	`1.04`

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

oqs-bot

AMD EPYC 4th gen (c7a)

Details

Benchmark suite	Current: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 keypair`	`44147` cycles	`43880` cycles	`1.01`
`ML-DSA-44 sign`	`192701` cycles	`153569` cycles	`1.25`
`ML-DSA-44 verify`	`60176` cycles	`55229` cycles	`1.09`
`ML-DSA-65 keypair`	`77972` cycles	`74426` cycles	`1.05`
`ML-DSA-65 sign`	`319890` cycles	`246006` cycles	`1.30`
`ML-DSA-65 verify`	`97664` cycles	`84283` cycles	`1.16`
`ML-DSA-87 keypair`	`114975` cycles	`114067` cycles	`1.01`
`ML-DSA-87 sign`	`359514` cycles	`295714` cycles	`1.22`
`ML-DSA-87 verify`	`140767` cycles	`128169` cycles	`1.10`

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: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 sign`	`192701` cycles	`153569` cycles	`1.25`
`ML-DSA-44 verify`	`60176` cycles	`55229` cycles	`1.09`
`ML-DSA-65 keypair`	`77972` cycles	`74426` cycles	`1.05`
`ML-DSA-65 sign`	`319890` cycles	`246006` cycles	`1.30`
`ML-DSA-65 verify`	`97664` cycles	`84283` cycles	`1.16`
`ML-DSA-87 sign`	`359514` cycles	`295714` cycles	`1.22`
`ML-DSA-87 verify`	`140767` cycles	`128169` cycles	`1.10`

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

oqs-bot

Graviton4 (no-opt)

Details

Benchmark suite	Current: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 keypair`	`134668` cycles	`132708` cycles	`1.01`
`ML-DSA-44 sign`	`499979` cycles	`471356` cycles	`1.06`
`ML-DSA-44 verify`	`149132` cycles	`143518` cycles	`1.04`
`ML-DSA-65 keypair`	`228685` cycles	`226009` cycles	`1.01`
`ML-DSA-65 sign`	`805992` cycles	`739063` cycles	`1.09`
`ML-DSA-65 verify`	`236898` cycles	`228696` cycles	`1.04`
`ML-DSA-87 keypair`	`377151` cycles	`374200` cycles	`1.01`
`ML-DSA-87 sign`	`1017722` cycles	`945390` cycles	`1.08`
`ML-DSA-87 verify`	`390838` cycles	`378937` cycles	`1.03`

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

oqs-bot

⚠️ Performance Alert ⚠️

Possible performance regression was detected for benchmark 'Graviton4 (no-opt)'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.03.

Benchmark suite	Current: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 sign`	`499979` cycles	`471356` cycles	`1.06`
`ML-DSA-44 verify`	`149132` cycles	`143518` cycles	`1.04`
`ML-DSA-65 sign`	`805992` cycles	`739063` cycles	`1.09`
`ML-DSA-65 verify`	`236898` cycles	`228696` cycles	`1.04`
`ML-DSA-87 sign`	`1017722` cycles	`945390` cycles	`1.08`
`ML-DSA-87 verify`	`390838` cycles	`378937` cycles	`1.03`

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

oqs-bot

Graviton2

Details

Benchmark suite	Current: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 keypair`	`120917` cycles	`117704` cycles	`1.03`
`ML-DSA-44 sign`	`480627` cycles	`419317` cycles	`1.15`
`ML-DSA-44 verify`	`145847` cycles	`137009` cycles	`1.06`
`ML-DSA-65 keypair`	`207782` cycles	`202971` cycles	`1.02`
`ML-DSA-65 sign`	`786650` cycles	`691577` cycles	`1.14`
`ML-DSA-65 verify`	`231653` cycles	`220514` cycles	`1.05`
`ML-DSA-87 keypair`	`336591` cycles	`331100` cycles	`1.02`
`ML-DSA-87 sign`	`967345` cycles	`869010` cycles	`1.11`
`ML-DSA-87 verify`	`369108` cycles	`354605` cycles	`1.04`

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

oqs-bot

⚠️ Performance Alert ⚠️

Possible performance regression was detected for benchmark 'Graviton2'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.03.

Benchmark suite	Current: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 sign`	`480627` cycles	`419317` cycles	`1.15`
`ML-DSA-44 verify`	`145847` cycles	`137009` cycles	`1.06`
`ML-DSA-65 sign`	`786650` cycles	`691577` cycles	`1.14`
`ML-DSA-65 verify`	`231653` cycles	`220514` cycles	`1.05`
`ML-DSA-87 sign`	`967345` cycles	`869010` cycles	`1.11`
`ML-DSA-87 verify`	`369108` cycles	`354605` cycles	`1.04`

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

oqs-bot

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

Details

Benchmark suite	Current: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 keypair`	`122467` cycles	`120631` cycles	`1.02`
`ML-DSA-44 sign`	`460478` cycles	`424505` cycles	`1.08`
`ML-DSA-44 verify`	`136758` cycles	`132270` cycles	`1.03`
`ML-DSA-65 keypair`	`206273` cycles	`203949` cycles	`1.01`
`ML-DSA-65 sign`	`746191` cycles	`677550` cycles	`1.10`
`ML-DSA-65 verify`	`216398` cycles	`207064` cycles	`1.05`
`ML-DSA-87 keypair`	`340114` cycles	`337594` cycles	`1.01`
`ML-DSA-87 sign`	`944638` cycles	`871359` cycles	`1.08`
`ML-DSA-87 verify`	`355602` cycles	`343109` cycles	`1.04`

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: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 sign`	`799828` cycles	`746549` cycles	`1.07`
`ML-DSA-44 verify`	`266832` cycles	`245039` cycles	`1.09`
`ML-DSA-65 sign`	`1293746` cycles	`1184030` cycles	`1.09`
`ML-DSA-65 verify`	`423456` cycles	`403607` cycles	`1.05`
`ML-DSA-87 sign`	`1666424` cycles	`1538472` cycles	`1.08`

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

oqs-bot

Graviton3 (no-opt)

Details

Benchmark suite	Current: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 keypair`	`139809` cycles	`137464` cycles	`1.02`
`ML-DSA-44 sign`	`506922` cycles	`454171` cycles	`1.12`
`ML-DSA-44 verify`	`154242` cycles	`147815` cycles	`1.04`
`ML-DSA-65 keypair`	`245075` cycles	`242209` cycles	`1.01`
`ML-DSA-65 sign`	`810119` cycles	`740730` cycles	`1.09`
`ML-DSA-65 verify`	`248459` cycles	`238979` cycles	`1.04`
`ML-DSA-87 keypair`	`397946` cycles	`392666` cycles	`1.01`
`ML-DSA-87 sign`	`1030480` cycles	`955507` cycles	`1.08`
`ML-DSA-87 verify`	`410884` cycles	`398960` cycles	`1.03`

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

oqs-bot

⚠️ Performance Alert ⚠️

Possible performance regression was detected for benchmark 'Graviton3 (no-opt)'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.03.

Benchmark suite	Current: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 sign`	`506922` cycles	`454171` cycles	`1.12`
`ML-DSA-44 verify`	`154242` cycles	`147815` cycles	`1.04`
`ML-DSA-65 sign`	`810119` cycles	`740730` cycles	`1.09`
`ML-DSA-65 verify`	`248459` cycles	`238979` cycles	`1.04`
`ML-DSA-87 sign`	`1030480` cycles	`955507` cycles	`1.08`

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

oqs-bot

Graviton2 (no-opt)

Details

Benchmark suite	Current: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 keypair`	`215634` cycles	`212496` cycles	`1.01`
`ML-DSA-44 sign`	`790020` cycles	`742145` cycles	`1.06`
`ML-DSA-44 verify`	`239451` cycles	`230539` cycles	`1.04`
`ML-DSA-65 keypair`	`383574` cycles	`380509` cycles	`1.01`
`ML-DSA-65 sign`	`1292839` cycles	`1199264` cycles	`1.08`
`ML-DSA-65 verify`	`384924` cycles	`373938` cycles	`1.03`
`ML-DSA-87 keypair`	`612155` cycles	`606509` cycles	`1.01`
`ML-DSA-87 sign`	`1652474` cycles	`1547667` cycles	`1.07`
`ML-DSA-87 verify`	`637552` cycles	`621502` cycles	`1.03`

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: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 sign`	`790020` cycles	`742145` cycles	`1.06`
`ML-DSA-44 verify`	`239451` cycles	`230539` cycles	`1.04`
`ML-DSA-65 sign`	`1292839` cycles	`1199264` cycles	`1.08`
`ML-DSA-87 sign`	`1652474` cycles	`1547667` cycles	`1.07`

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: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 keypair`	`946626` cycles	`944595` cycles	`1.00`
`ML-DSA-44 sign`	`4411928` cycles	`4357225` cycles	`1.01`
`ML-DSA-44 verify`	`1091753` cycles	`1075353` cycles	`1.02`
`ML-DSA-65 keypair`	`1582387` cycles	`1570727` cycles	`1.01`
`ML-DSA-65 sign`	`7317735` cycles	`7186436` cycles	`1.02`
`ML-DSA-65 verify`	`1728569` cycles	`1701260` cycles	`1.02`
`ML-DSA-87 keypair`	`2548834` cycles	`2540962` cycles	`1.00`
`ML-DSA-87 sign`	`8977512` cycles	`8815371` cycles	`1.02`
`ML-DSA-87 verify`	`2755678` cycles	`2720024` cycles	`1.01`

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: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 keypair`	`320126` cycles	`307872` cycles	`1.04`
`ML-DSA-44 sign`	`1233790` cycles	`1140862` cycles	`1.08`
`ML-DSA-44 verify`	`349626` cycles	`337492` cycles	`1.04`
`ML-DSA-65 keypair`	`590709` cycles	`578597` cycles	`1.02`
`ML-DSA-65 sign`	`2024614` cycles	`1867708` cycles	`1.08`
`ML-DSA-65 verify`	`565192` cycles	`551173` cycles	`1.03`
`ML-DSA-87 keypair`	`899030` cycles	`872509` cycles	`1.03`
`ML-DSA-87 sign`	`2542452` cycles	`2349579` cycles	`1.08`
`ML-DSA-87 verify`	`926170` cycles	`879687` cycles	`1.05`

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: `5110e63`	Previous: `334c7a2`	Ratio
`ML-DSA-44 keypair`	`320126` cycles	`307872` cycles	`1.04`
`ML-DSA-44 sign`	`1233790` cycles	`1140862` cycles	`1.08`
`ML-DSA-44 verify`	`349626` cycles	`337492` cycles	`1.04`
`ML-DSA-65 sign`	`2024614` cycles	`1867708` cycles	`1.08`
`ML-DSA-87 keypair`	`899030` cycles	`872509` cycles	`1.03`
`ML-DSA-87 sign`	`2542452` cycles	`2349579` cycles	`1.08`
`ML-DSA-87 verify`	`926170` cycles	`879687` cycles	`1.05`

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

This commit adds a config variation test to CI to test the alternative constant time functions (when no inline asm is available). To do so, first a MLD_CONFIG_FILE configuration option is introduced allowing to define a custom configurtion file. This option is then used in a workflow to include a config that has MLD_CONFIG_NO_ASM set. Signed-off-by: Matthias J. Kannwischer <matthias@kannwischer.eu>

…config.h Signed-off-by: Matthias J. Kannwischer <matthias@kannwischer.eu>

mldsa/ct.h

hanno-becker

Some spurious TODOs left, but otherwise LGTM

Based on pq-code-package/mldsa-native#371. Signed-off-by: dkostic <dkostic@amazon.com>

) ML-DSA constant-time hardening for caddq, poly_chknorm, decompose Based on pq-code-package/mldsa-native#371.

mkannwischer added the benchmark label Jul 17, 2025

github-actions bot reviewed Jul 17, 2025

View reviewed changes

oqs-bot reviewed Jul 17, 2025

View reviewed changes

github-actions bot reviewed Jul 17, 2025

View reviewed changes

oqs-bot reviewed Jul 17, 2025

View reviewed changes

github-actions bot reviewed Jul 28, 2025

View reviewed changes

oqs-bot reviewed Jul 28, 2025

View reviewed changes

github-actions bot reviewed Jul 28, 2025

View reviewed changes

mkannwischer added 2 commits July 28, 2025 16:37

CT: Improve documentation and add MLD_CONFIG_NO_ASM_VALUE_BARRIER to …

2da7a4f

…config.h Signed-off-by: Matthias J. Kannwischer <matthias@kannwischer.eu>

mkannwischer force-pushed the constant-time branch from 5110e63 to 2da7a4f Compare July 28, 2025 08:37

This was referenced Jul 28, 2025

Add valgrind-based constant-time tests #391

Merged

CT: Make poly_chknorm constant flow #392

Merged

willieyz mentioned this pull request Jul 29, 2025

CI: add the PCT enabled, PCT enabled + broken, Custom zeroization (explicit_bzero) #393

Merged

mkannwischer mentioned this pull request Jul 30, 2025

Clean up mld_attempt_signature_generation #395

Closed

hanno-becker reviewed Jul 30, 2025

View reviewed changes

mldsa/ct.h Show resolved Hide resolved

hanno-becker approved these changes Jul 30, 2025

View reviewed changes

hanno-becker merged commit 6369a7e into main Jul 30, 2025
170 checks passed

hanno-becker deleted the constant-time branch July 30, 2025 06:14

dkostic added a commit to dkostic/aws-lc that referenced this pull request Aug 5, 2025

ML-DSA constant-time hardening for caddq, poly_chknorm, decompose

9f71d54

Based on pq-code-package/mldsa-native#371. Signed-off-by: dkostic <dkostic@amazon.com>

dkostic added a commit to dkostic/aws-lc that referenced this pull request Aug 5, 2025

ML-DSA constant-time hardening for caddq, poly_chknorm, decompose

579a7a5

Based on pq-code-package/mldsa-native#371. Signed-off-by: dkostic <dkostic@amazon.com>

dkostic mentioned this pull request Aug 5, 2025

ML-DSA constant-time hardening for caddq, poly_chknorm, decompose aws/aws-lc#2602

Merged

dkostic added a commit to aws/aws-lc that referenced this pull request Aug 12, 2025

ML-DSA constant-time hardening for caddq, poly_chknorm, decompose (#2602

2a184bd

) ML-DSA constant-time hardening for caddq, poly_chknorm, decompose Based on pq-code-package/mldsa-native#371.

mkannwischer mentioned this pull request Nov 13, 2025

mldsa-native integration open-quantum-safe/liboqs#2284

Merged

Conversation

mkannwischer commented Jul 16, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

github-actions bot left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Mac Mini (M1, 2020) benchmarks (opt)

Uh oh!

github-actions bot left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

⚠️ Performance Alert ⚠️

Uh oh!

github-actions bot left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

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

Uh oh!

github-actions bot left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

⚠️ Performance Alert ⚠️

Uh oh!

github-actions bot left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

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

Uh oh!

github-actions bot left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

⚠️ Performance Alert ⚠️

Uh oh!

oqs-bot left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Intel Xeon 4th gen (c7i)

Uh oh!

oqs-bot left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

⚠️ Performance Alert ⚠️

Uh oh!

oqs-bot left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

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

Uh oh!

oqs-bot left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

⚠️ Performance Alert ⚠️

Uh oh!

oqs-bot left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

AMD EPYC 3rd gen (c6a)

Uh oh!

oqs-bot left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

⚠️ Performance Alert ⚠️

Uh oh!

oqs-bot left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Intel Xeon 3rd gen (c6i)

Uh oh!

oqs-bot left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

⚠️ Performance Alert ⚠️

Uh oh!

oqs-bot left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Graviton4

Uh oh!

oqs-bot left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

mkannwischer commented Jul 16, 2025 •

edited

Loading

github-actions bot left a comment •

edited

Loading

github-actions bot left a comment •

edited

Loading

github-actions bot left a comment •

edited

Loading

github-actions bot left a comment •

edited

Loading

github-actions bot left a comment •

edited

Loading

github-actions bot left a comment •

edited

Loading

oqs-bot left a comment •

edited

Loading

oqs-bot left a comment •

edited

Loading

oqs-bot left a comment •

edited

Loading

oqs-bot left a comment •

edited

Loading

oqs-bot left a comment •

edited

Loading

oqs-bot left a comment •

edited

Loading

oqs-bot left a comment •

edited

Loading

oqs-bot left a comment •

edited

Loading

oqs-bot left a comment •

edited

Loading

oqs-bot left a comment •

edited

Loading

github-actions bot left a comment •

edited

Loading

github-actions bot left a comment •

edited

Loading

oqs-bot left a comment •

edited

Loading

oqs-bot left a comment •

edited

Loading

oqs-bot left a comment •

edited

Loading

oqs-bot left a comment •

edited

Loading

oqs-bot left a comment •

edited

Loading

oqs-bot left a comment •

edited

Loading

oqs-bot left a comment •

edited

Loading

oqs-bot left a comment •

edited

Loading

oqs-bot left a comment •

edited

Loading

oqs-bot left a comment •

edited

Loading