Add ppc64le backend (supports p8 and above architectures) [full CI] by mkannwischer · Pull Request #1677 · pq-code-package/mlkem-native

mkannwischer · 2026-05-05T08:24:30Z

Running the full CI on #1648

oqs-bot

Mac Mini (M1, 2020) benchmarks

Details

Benchmark suite	Current: `12036a9`	Previous: `db75353`	Ratio
`ML-KEM-512 keypair`	`12319` cycles	`12319` cycles	`1`
`ML-KEM-512 encaps`	`14997` cycles	`14997` cycles	`1`
`ML-KEM-512 decaps`	`19549` cycles	`19551` cycles	`1.00`
`ML-KEM-768 keypair`	`21263` cycles	`21264` cycles	`1.00`
`ML-KEM-768 encaps`	`23873` cycles	`23874` cycles	`1.00`
`ML-KEM-768 decaps`	`30417` cycles	`30425` cycles	`1.00`
`ML-KEM-1024 keypair`	`30328` cycles	`30327` cycles	`1.00`
`ML-KEM-1024 encaps`	`34573` cycles	`34573` cycles	`1`
`ML-KEM-1024 decaps`	`44189` cycles	`44190` cycles	`1.00`

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

oqs-bot

ppc64le (POWER10) benchmarks

Details

Benchmark suite	Current: `12036a9`	Previous: `db75353`	Ratio
`ML-KEM-512 keypair`	`38059` cycles	`59591` cycles	`0.64`
`ML-KEM-512 encaps`	`43668` cycles	`72335` cycles	`0.60`
`ML-KEM-512 decaps`	`53803` cycles	`92227` cycles	`0.58`
`ML-KEM-768 keypair`	`67131` cycles	`97410` cycles	`0.69`
`ML-KEM-768 encaps`	`76341` cycles	`114292` cycles	`0.67`
`ML-KEM-768 decaps`	`90588` cycles	`139751` cycles	`0.65`
`ML-KEM-1024 keypair`	`108574` cycles	`151020` cycles	`0.72`
`ML-KEM-1024 encaps`	`119069` cycles	`169141` cycles	`0.70`
`ML-KEM-1024 decaps`	`137495` cycles	`200776` cycles	`0.68`

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

oqs-bot

Intel Xeon 4th gen (c7i)

Details

Benchmark suite	Current: `12036a9`	Previous: `db75353`	Ratio
`ML-KEM-512 keypair`	`12048` cycles	`12038` cycles	`1.00`
`ML-KEM-512 encaps`	`13632` cycles	`13787` cycles	`0.99`
`ML-KEM-512 decaps`	`17778` cycles	`17801` cycles	`1.00`
`ML-KEM-768 keypair`	`21266` cycles	`21014` cycles	`1.01`
`ML-KEM-768 encaps`	`22146` cycles	`22184` cycles	`1.00`
`ML-KEM-768 decaps`	`28443` cycles	`28329` cycles	`1.00`
`ML-KEM-1024 keypair`	`29577` cycles	`29959` cycles	`0.99`
`ML-KEM-1024 encaps`	`31745` cycles	`31722` cycles	`1.00`
`ML-KEM-1024 decaps`	`39476` cycles	`39346` 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 'Intel Xeon 4th gen (c7i)'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.03.

Benchmark suite	Current: `5f3d329`	Previous: `2ee902c`	Ratio
`ML-KEM-512 keypair`	`12044` cycles	`9751` cycles	`1.24`
`ML-KEM-512 encaps`	`13624` cycles	`11423` cycles	`1.19`
`ML-KEM-512 decaps`	`17783` cycles	`15570` cycles	`1.14`
`ML-KEM-768 keypair`	`21292` cycles	`16302` cycles	`1.31`
`ML-KEM-768 encaps`	`22015` cycles	`17954` cycles	`1.23`
`ML-KEM-768 decaps`	`28023` cycles	`23461` cycles	`1.19`
`ML-KEM-1024 keypair`	`29562` cycles	`22439` cycles	`1.32`
`ML-KEM-1024 encaps`	`31715` cycles	`24509` cycles	`1.29`
`ML-KEM-1024 decaps`	`39395` cycles	`32178` cycles	`1.22`

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

oqs-bot

AMD EPYC 3rd gen (c6a)

Details

Benchmark suite	Current: `12036a9`	Previous: `db75353`	Ratio
`ML-KEM-512 keypair`	`14207` cycles	`14376` cycles	`0.99`
`ML-KEM-512 encaps`	`15984` cycles	`16060` cycles	`1.00`
`ML-KEM-512 decaps`	`21538` cycles	`21627` cycles	`1.00`
`ML-KEM-768 keypair`	`25114` cycles	`24794` cycles	`1.01`
`ML-KEM-768 encaps`	`25658` cycles	`25550` cycles	`1.00`
`ML-KEM-768 decaps`	`33523` cycles	`33409` cycles	`1.00`
`ML-KEM-1024 keypair`	`34848` cycles	`37228` cycles	`0.94`
`ML-KEM-1024 encaps`	`36114` cycles	`37346` cycles	`0.97`
`ML-KEM-1024 decaps`	`47236` cycles	`46787` cycles	`1.01`

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

oqs-bot

AMD EPYC 4th gen (c7a)

Details

Benchmark suite	Current: `12036a9`	Previous: `db75353`	Ratio
`ML-KEM-512 keypair`	`12796` cycles	`12777` cycles	`1.00`
`ML-KEM-512 encaps`	`14285` cycles	`14269` cycles	`1.00`
`ML-KEM-512 decaps`	`19148` cycles	`19117` cycles	`1.00`
`ML-KEM-768 keypair`	`22525` cycles	`22412` cycles	`1.01`
`ML-KEM-768 encaps`	`23071` cycles	`23051` cycles	`1.00`
`ML-KEM-768 decaps`	`30094` cycles	`30064` cycles	`1.00`
`ML-KEM-1024 keypair`	`34224` cycles	`32997` cycles	`1.04`
`ML-KEM-1024 encaps`	`33002` cycles	`33104` cycles	`1.00`
`ML-KEM-1024 decaps`	`42405` cycles	`42483` cycles	`1.00`

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

oqs-bot

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

Details

Benchmark suite	Current: `12036a9`	Previous: `db75353`	Ratio
`ML-KEM-512 keypair`	`28337` cycles	`28237` cycles	`1.00`
`ML-KEM-512 encaps`	`36782` cycles	`36623` cycles	`1.00`
`ML-KEM-512 decaps`	`45398` cycles	`45123` cycles	`1.01`
`ML-KEM-768 keypair`	`46215` cycles	`46315` cycles	`1.00`
`ML-KEM-768 encaps`	`55787` cycles	`55593` cycles	`1.00`
`ML-KEM-768 decaps`	`69875` cycles	`69917` cycles	`1.00`
`ML-KEM-1024 keypair`	`70417` cycles	`70363` cycles	`1.00`
`ML-KEM-1024 encaps`	`82459` cycles	`82510` cycles	`1.00`
`ML-KEM-1024 decaps`	`99343` cycles	`99218` 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 4th gen (c7a)'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.03.

Benchmark suite	Current: `12036a9`	Previous: `db75353`	Ratio
`ML-KEM-1024 keypair`	`34224` cycles	`32997` cycles	`1.04`

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

oqs-bot

Intel Xeon 3rd gen (c6i)

Details

Benchmark suite	Current: `12036a9`	Previous: `db75353`	Ratio
`ML-KEM-512 keypair`	`17603` cycles	`17547` cycles	`1.00`
`ML-KEM-512 encaps`	`19900` cycles	`19938` cycles	`1.00`
`ML-KEM-512 decaps`	`26420` cycles	`26450` cycles	`1.00`
`ML-KEM-768 keypair`	`31203` cycles	`31168` cycles	`1.00`
`ML-KEM-768 encaps`	`31989` cycles	`32415` cycles	`0.99`
`ML-KEM-768 decaps`	`41468` cycles	`41536` cycles	`1.00`
`ML-KEM-1024 keypair`	`43770` cycles	`43998` cycles	`0.99`
`ML-KEM-1024 encaps`	`45855` cycles	`46270` cycles	`0.99`
`ML-KEM-1024 decaps`	`58042` cycles	`58266` 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 'Intel Xeon 3rd gen (c6i)'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.03.

Benchmark suite	Current: `5f3d329`	Previous: `2ee902c`	Ratio
`ML-KEM-512 keypair`	`17588` cycles	`16301` cycles	`1.08`
`ML-KEM-512 encaps`	`19896` cycles	`18736` cycles	`1.06`
`ML-KEM-512 decaps`	`26412` cycles	`25234` cycles	`1.05`
`ML-KEM-768 keypair`	`31190` cycles	`28649` cycles	`1.09`
`ML-KEM-768 encaps`	`31768` cycles	`30001` cycles	`1.06`
`ML-KEM-1024 keypair`	`43790` cycles	`37884` cycles	`1.16`
`ML-KEM-1024 encaps`	`45790` cycles	`40704` cycles	`1.12`
`ML-KEM-1024 decaps`	`58108` cycles	`54265` cycles	`1.07`

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

oqs-bot

Graviton4

Details

Benchmark suite	Current: `12036a9`	Previous: `db75353`	Ratio
`ML-KEM-512 keypair`	`17686` cycles	`17646` cycles	`1.00`
`ML-KEM-512 encaps`	`20642` cycles	`20608` cycles	`1.00`
`ML-KEM-512 decaps`	`27077` cycles	`27084` cycles	`1.00`
`ML-KEM-768 keypair`	`29981` cycles	`29899` cycles	`1.00`
`ML-KEM-768 encaps`	`32757` cycles	`32774` cycles	`1.00`
`ML-KEM-768 decaps`	`42007` cycles	`41962` cycles	`1.00`
`ML-KEM-1024 keypair`	`43716` cycles	`43745` cycles	`1.00`
`ML-KEM-1024 encaps`	`48773` cycles	`48719` cycles	`1.00`
`ML-KEM-1024 decaps`	`61379` cycles	`61386` 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: `12036a9`	Previous: `db75353`	Ratio
`ML-KEM-512 keypair`	`40197` cycles	`40266` cycles	`1.00`
`ML-KEM-512 encaps`	`48368` cycles	`48417` cycles	`1.00`
`ML-KEM-512 decaps`	`62501` cycles	`62596` cycles	`1.00`
`ML-KEM-768 keypair`	`63804` cycles	`63800` cycles	`1.00`
`ML-KEM-768 encaps`	`74937` cycles	`74978` cycles	`1.00`
`ML-KEM-768 decaps`	`93413` cycles	`93631` cycles	`1.00`
`ML-KEM-1024 keypair`	`95310` cycles	`95102` cycles	`1.00`
`ML-KEM-1024 encaps`	`109384` cycles	`109294` cycles	`1.00`
`ML-KEM-1024 decaps`	`132137` cycles	`132065` 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: `12036a9`	Previous: `db75353`	Ratio
`ML-KEM-512 keypair`	`36602` cycles	`36810` cycles	`0.99`
`ML-KEM-512 encaps`	`43115` cycles	`43058` cycles	`1.00`
`ML-KEM-512 decaps`	`55703` cycles	`55671` cycles	`1.00`
`ML-KEM-768 keypair`	`58678` cycles	`58693` cycles	`1.00`
`ML-KEM-768 encaps`	`67624` cycles	`67471` cycles	`1.00`
`ML-KEM-768 decaps`	`84521` cycles	`84392` cycles	`1.00`
`ML-KEM-1024 keypair`	`89114` cycles	`89088` cycles	`1.00`
`ML-KEM-1024 encaps`	`99256` cycles	`99346` cycles	`1.00`
`ML-KEM-1024 decaps`	`120774` cycles	`120756` 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: `12036a9`	Previous: `db75353`	Ratio
`ML-KEM-512 keypair`	`45749` cycles	`45722` cycles	`1.00`
`ML-KEM-512 encaps`	`54475` cycles	`54376` cycles	`1.00`
`ML-KEM-512 decaps`	`69855` cycles	`69830` cycles	`1.00`
`ML-KEM-768 keypair`	`74173` cycles	`74187` cycles	`1.00`
`ML-KEM-768 encaps`	`86050` cycles	`86041` cycles	`1.00`
`ML-KEM-768 decaps`	`106672` cycles	`106532` cycles	`1.00`
`ML-KEM-1024 keypair`	`112123` cycles	`112130` cycles	`1.00`
`ML-KEM-1024 encaps`	`124717` cycles	`124654` cycles	`1.00`
`ML-KEM-1024 decaps`	`150632` cycles	`150714` cycles	`1.00`

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

oqs-bot

Graviton3

Details

Benchmark suite	Current: `12036a9`	Previous: `db75353`	Ratio
`ML-KEM-512 keypair`	`18675` cycles	`18639` cycles	`1.00`
`ML-KEM-512 encaps`	`21889` cycles	`21878` cycles	`1.00`
`ML-KEM-512 decaps`	`28890` cycles	`28864` cycles	`1.00`
`ML-KEM-768 keypair`	`31630` cycles	`31545` cycles	`1.00`
`ML-KEM-768 encaps`	`34788` cycles	`34776` cycles	`1.00`
`ML-KEM-768 decaps`	`44839` cycles	`44778` cycles	`1.00`
`ML-KEM-1024 keypair`	`46069` cycles	`46079` cycles	`1.00`
`ML-KEM-1024 encaps`	`51492` cycles	`51492` cycles	`1`
`ML-KEM-1024 decaps`	`65005` cycles	`65023` cycles	`1.00`

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

oqs-bot

Graviton4 (no-opt)

Details

Benchmark suite	Current: `12036a9`	Previous: `db75353`	Ratio
`ML-KEM-512 keypair`	`35504` cycles	`35410` cycles	`1.00`
`ML-KEM-512 encaps`	`40175` cycles	`40114` cycles	`1.00`
`ML-KEM-512 decaps`	`51132` cycles	`51139` cycles	`1.00`
`ML-KEM-768 keypair`	`56800` cycles	`56670` cycles	`1.00`
`ML-KEM-768 encaps`	`64827` cycles	`65147` cycles	`1.00`
`ML-KEM-768 decaps`	`78931` cycles	`79294` cycles	`1.00`
`ML-KEM-1024 keypair`	`87846` cycles	`87857` cycles	`1.00`
`ML-KEM-1024 encaps`	`97109` cycles	`96871` cycles	`1.00`
`ML-KEM-1024 decaps`	`115956` cycles	`115822` cycles	`1.00`

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

oqs-bot

Arm Cortex-A76 (Raspberry Pi 5) benchmarks

Details

Benchmark suite	Current: `12036a9`	Previous: `db75353`	Ratio
`ML-KEM-512 keypair`	`28265` cycles	`28220` cycles	`1.00`
`ML-KEM-512 encaps`	`34157` cycles	`34107` cycles	`1.00`
`ML-KEM-512 decaps`	`44377` cycles	`44335` cycles	`1.00`
`ML-KEM-768 keypair`	`47618` cycles	`47614` cycles	`1.00`
`ML-KEM-768 encaps`	`53934` cycles	`53937` cycles	`1.00`
`ML-KEM-768 decaps`	`68340` cycles	`68365` cycles	`1.00`
`ML-KEM-1024 keypair`	`70248` cycles	`70248` cycles	`1`
`ML-KEM-1024 encaps`	`78733` cycles	`78728` cycles	`1.00`
`ML-KEM-1024 decaps`	`98416` cycles	`98444` cycles	`1.00`

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

oqs-bot

Graviton3 (no-opt)

Details

Benchmark suite	Current: `12036a9`	Previous: `db75353`	Ratio
`ML-KEM-512 keypair`	`39016` cycles	`38886` cycles	`1.00`
`ML-KEM-512 encaps`	`44562` cycles	`44589` cycles	`1.00`
`ML-KEM-512 decaps`	`56630` cycles	`56665` cycles	`1.00`
`ML-KEM-768 keypair`	`62456` cycles	`62296` cycles	`1.00`
`ML-KEM-768 encaps`	`71385` cycles	`72308` cycles	`0.99`
`ML-KEM-768 decaps`	`86856` cycles	`87700` cycles	`0.99`
`ML-KEM-1024 keypair`	`96224` cycles	`96159` cycles	`1.00`
`ML-KEM-1024 encaps`	`106363` cycles	`106136` cycles	`1.00`
`ML-KEM-1024 decaps`	`126811` cycles	`126585` cycles	`1.00`

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

oqs-bot

Graviton2

Details

Benchmark suite	Current: `12036a9`	Previous: `db75353`	Ratio
`ML-KEM-512 keypair`	`28238` cycles	`28274` cycles	`1.00`
`ML-KEM-512 encaps`	`34161` cycles	`34125` cycles	`1.00`
`ML-KEM-512 decaps`	`44340` cycles	`44382` cycles	`1.00`
`ML-KEM-768 keypair`	`47638` cycles	`47672` cycles	`1.00`
`ML-KEM-768 encaps`	`53920` cycles	`53906` cycles	`1.00`
`ML-KEM-768 decaps`	`68398` cycles	`68361` cycles	`1.00`
`ML-KEM-1024 keypair`	`70366` cycles	`70253` cycles	`1.00`
`ML-KEM-1024 encaps`	`78752` cycles	`78754` cycles	`1.00`
`ML-KEM-1024 decaps`	`98551` cycles	`98440` cycles	`1.00`

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

oqs-bot

SpacemiT K1 8 (Banana Pi F3) benchmarks

Details

Benchmark suite	Current: `12036a9`	Previous: `db75353`	Ratio
`ML-KEM-512 keypair`	`155506` cycles	`155497` cycles	`1.00`
`ML-KEM-512 encaps`	`163419` cycles	`163389` cycles	`1.00`
`ML-KEM-512 decaps`	`206655` cycles	`206624` cycles	`1.00`
`ML-KEM-768 keypair`	`249866` cycles	`249882` cycles	`1.00`
`ML-KEM-768 encaps`	`270396` cycles	`270411` cycles	`1.00`
`ML-KEM-768 decaps`	`332775` cycles	`332827` cycles	`1.00`
`ML-KEM-1024 keypair`	`395754` cycles	`395617` cycles	`1.00`
`ML-KEM-1024 encaps`	`423609` cycles	`422610` cycles	`1.00`
`ML-KEM-1024 decaps`	`507214` cycles	`506225` cycles	`1.00`

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

oqs-bot

Graviton2 (no-opt)

Details

Benchmark suite	Current: `12036a9`	Previous: `db75353`	Ratio
`ML-KEM-512 keypair`	`59230` cycles	`59139` cycles	`1.00`
`ML-KEM-512 encaps`	`68655` cycles	`68634` cycles	`1.00`
`ML-KEM-512 decaps`	`87379` cycles	`87351` cycles	`1.00`
`ML-KEM-768 keypair`	`95187` cycles	`95327` cycles	`1.00`
`ML-KEM-768 encaps`	`109224` cycles	`109878` cycles	`0.99`
`ML-KEM-768 decaps`	`134022` cycles	`134352` cycles	`1.00`
`ML-KEM-1024 keypair`	`146800` cycles	`148090` cycles	`0.99`
`ML-KEM-1024 encaps`	`162753` cycles	`163969` cycles	`0.99`
`ML-KEM-1024 decaps`	`194331` cycles	`195624` cycles	`0.99`

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

oqs-bot

Arm Cortex-A55 (Snapdragon 888) benchmarks

Details

Benchmark suite	Current: `12036a9`	Previous: `db75353`	Ratio
`ML-KEM-512 keypair`	`59765` cycles	`59775` cycles	`1.00`
`ML-KEM-512 encaps`	`67418` cycles	`67520` cycles	`1.00`
`ML-KEM-512 decaps`	`86122` cycles	`86168` cycles	`1.00`
`ML-KEM-768 keypair`	`97488` cycles	`97434` cycles	`1.00`
`ML-KEM-768 encaps`	`110849` cycles	`110991` cycles	`1.00`
`ML-KEM-768 decaps`	`138197` cycles	`138336` cycles	`1.00`
`ML-KEM-1024 keypair`	`155074` cycles	`154826` cycles	`1.00`
`ML-KEM-1024 encaps`	`172210` cycles	`172753` cycles	`1.00`
`ML-KEM-1024 decaps`	`209438` cycles	`208701` cycles	`1.00`

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

oqs-bot · 2026-05-05T08:46:34Z

CBMC Results (ML-KEM-1024)

Full Results (191 proofs)

Proof	Status	Current	Previous	Change
`TOTAL`	✅	1246s	1200s	+3.8%
`mlk_indcpa_enc`	✅	141s	135s	+4%
`mlk_rej_uniform_c`	✅	133s	125s	+6%
`mlk_indcpa_keypair_derand`	✅	128s	116s	+10%
`mlk_polyvec_basemul_acc_montgomery_cached_c`	✅	76s	69s	+10%
`polyvec_basemul_acc_montgomery_cached_native`	✅	34s	39s	-13%
`mlk_ntt_layer`	✅	33s	32s	+3%
`mlk_poly_rej_uniform`	✅	32s	30s	+7%
`poly_ntt_native`	✅	28s	27s	+4%
`mlk_keccak_squeezeblocks_x4`	✅	24s	26s	-8%
`mlk_poly_reduce_native`	✅	21s	19s	+11%
`keccakf1600x4_permute_native_x4`	✅	17s	16s	+6%
`mlk_fqmul`	✅	16s	17s	-6%
`mlk_poly_decompress_d11_native`	✅	15s	16s	-6%
`mlk_poly_decompress_d5_native`	✅	12s	13s	-8%
`mlk_keccak_squeezeblocks`	✅	11s	9s	+22%
`mlk_polyvec_add`	✅	11s	11s	+0%
`mlk_poly_frommsg`	✅	9s	8s	+12%
`mlk_indcpa_dec`	✅	8s	8s	+0%
`mlk_ntt_butterfly_block`	✅	8s	6s	+33%
`mlk_keccak_squeeze_once`	✅	7s	8s	-12%
`mlk_poly_frombytes_native`	✅	7s	7s	+0%
`mlk_poly_rej_uniform_x4`	✅	7s	8s	-12%
`kem_check_sk`	✅	6s	2s	+200%
`kem_dec`	✅	6s	5s	+20%
`mlk_poly_ntt`	✅	6s	6s	+0%
`mlk_polymat_permute_bitrev_to_custom`	✅	6s	6s	+0%
`mlk_polyvec_ntt`	✅	6s	3s	+100%
`mlk_gen_matrix`	✅	5s	4s	+25%
`mlk_invntt_layer`	✅	5s	4s	+25%
`mlk_keccak_absorb_once`	✅	5s	3s	+67%
`mlk_keccak_absorb_once_x4`	✅	5s	7s	-29%
`mlk_keccakf1600_permute_c`	✅	5s	6s	-17%
`mlk_polyvec_tomont`	✅	5s	2s	+150%
`mlk_shake256`	✅	5s	1s	+400%
`mlk_shake256x4`	✅	5s	4s	+25%
`poly_decompress_d5_native_x86_64`	✅	5s	5s	+0%
`poly_tomont_native_x86_64`	✅	5s	5s	+0%
`polyvec_basemul_acc_montgomery_cached_k3_native_x86_64`	✅	5s	3s	+67%
`rej_uniform_native_x86_64`	✅	5s	5s	+0%
`intt_native_aarch64`	✅	4s	3s	+33%
`keccak_f1600_x4_native_aarch64_v84a`	✅	4s	2s	+100%
`kem_enc`	✅	4s	3s	+33%
`mlk_gen_matrix_serial`	✅	4s	6s	-33%
`mlk_keypair_getnoise_eta1`	✅	4s	3s	+33%
`mlk_poly_cbd_eta2`	✅	4s	4s	+0%
`mlk_poly_compress_d11_c`	✅	4s	4s	+0%
`mlk_poly_compress_d11_native`	✅	4s	2s	+100%
`mlk_poly_compress_d5_native`	✅	4s	1s	+300%
`mlk_poly_decompress_d10`	✅	4s	2s	+100%
`mlk_poly_decompress_d5_c`	✅	4s	1s	+300%
`mlk_poly_getnoise_eta1_4x_native`	✅	4s	4s	+0%
`mlk_poly_mulcache_compute`	✅	4s	2s	+100%
`mlk_poly_mulcache_compute_c`	✅	4s	2s	+100%
`mlk_poly_sub`	✅	4s	2s	+100%
`mlk_poly_tobytes_native`	✅	4s	2s	+100%
`mlk_poly_tomsg`	✅	4s	3s	+33%
`mlk_polyvec_reduce`	✅	4s	3s	+33%
`mlk_rej_uniform`	✅	4s	2s	+100%
`mlk_scalar_decompress_d11`	✅	4s	1s	+300%
`nttunpack_native_x86_64`	✅	4s	3s	+33%
`poly_compress_d5_native_x86_64`	✅	4s	2s	+100%
`poly_decompress_d10_native_x86_64`	✅	4s	3s	+33%
`poly_decompress_d11_native_x86_64`	✅	4s	4s	+0%
`poly_frombytes_native_x86_64`	✅	4s	5s	-20%
`poly_reduce_native_aarch64`	✅	4s	2s	+100%
`polyvec_basemul_acc_montgomery_cached_k2_native_aarch64`	✅	4s	2s	+100%
`keccak_f1600_x4_native_aarch64_v8a_v84a_scalar_hybrid`	✅	3s	3s	+0%
`kem_check_pk`	✅	3s	3s	+0%
`kem_keypair_derand`	✅	3s	4s	-25%
`mlk_barrett_reduce`	✅	3s	3s	+0%
`mlk_ct_cmask_neg_i16`	✅	3s	5s	-40%
`mlk_ct_cmov_zero`	✅	3s	3s	+0%
`mlk_ct_get_optblocker_u32`	✅	3s	2s	+50%
`mlk_ct_get_optblocker_u8`	✅	3s	2s	+50%
`mlk_enc_getnoise_eta1_eta2`	✅	3s	2s	+50%
`mlk_keccakf1600_xor_bytes`	✅	3s	4s	-25%
`mlk_keccakf1600_xor_bytes (big endian)`	✅	3s	1s	+200%
`mlk_keccakf1600x4_extract_bytes_c`	✅	3s	4s	-25%
`mlk_keccakf1600x4_xor_bytes`	✅	3s	2s	+50%
`mlk_matvec_mul`	✅	3s	3s	+0%
`mlk_poly_compress_d11`	✅	3s	3s	+0%
`mlk_poly_compress_d4_native`	✅	3s	3s	+0%
`mlk_poly_compress_d5_c`	✅	3s	2s	+50%
`mlk_poly_decompress_d4_c`	✅	3s	1s	+200%
`mlk_poly_decompress_d5`	✅	3s	4s	-25%
`mlk_poly_decompress_du`	✅	3s	2s	+50%
`mlk_poly_getnoise_eta1_4x`	✅	3s	3s	+0%
`mlk_poly_invntt_tomont`	✅	3s	3s	+0%
`mlk_poly_mulcache_compute_native`	✅	3s	3s	+0%
`mlk_poly_reduce`	✅	3s	1s	+200%
`mlk_poly_reduce_c`	✅	3s	3s	+0%
`mlk_poly_tobytes`	✅	3s	2s	+50%
`mlk_poly_tomont`	✅	3s	2s	+50%
`mlk_polyvec_basemul_acc_montgomery_cached`	✅	3s	4s	-25%
`mlk_polyvec_compress_du`	✅	3s	1s	+200%
`mlk_polyvec_invntt_tomont`	✅	3s	2s	+50%
`mlk_polyvec_permute_bitrev_to_custom`	✅	3s	3s	+0%
`mlk_scalar_compress_d10`	✅	3s	2s	+50%
`mlk_scalar_compress_d4`	✅	3s	1s	+200%
`mlk_scalar_decompress_d4`	✅	3s	1s	+200%
`mlk_sha3_256`	✅	3s	1s	+200%
`mlk_shake128_absorb_once`	✅	3s	1s	+200%
`mlk_shake128x4_absorb_once`	✅	3s	2s	+50%
`ntt_native_x86_64`	✅	3s	3s	+0%
`poly_compress_d10_native_x86_64`	✅	3s	2s	+50%
`poly_invntt_tomont_native`	✅	3s	6s	-50%
`poly_mulcache_compute_native_x86_64`	✅	3s	2s	+50%
`poly_tobytes_native_x86_64`	✅	3s	2s	+50%
`rej_uniform_native`	✅	3s	4s	-25%
`intt_native_x86_64`	✅	2s	3s	-33%
`keccak_f1600_x1_native_aarch64`	✅	2s	2s	+0%
`keccak_f1600_x1_native_aarch64_v84a`	✅	2s	2s	+0%
`keccak_f1600_x4_native_aarch64_v8a_scalar_hybrid`	✅	2s	2s	+0%
`keccakf1600_permute_native`	✅	2s	3s	-33%
`keccakf1600x4_extract_bytes_native`	✅	2s	3s	-33%
`keccakf1600x4_xor_bytes_native`	✅	2s	1s	+100%
`kem_keypair`	✅	2s	2s	+0%
`mlk_check_pct`	✅	2s	3s	-33%
`mlk_ct_cmask_nonzero_u16`	✅	2s	2s	+0%
`mlk_ct_get_optblocker_i32`	✅	2s	2s	+0%
`mlk_ct_memcmp`	✅	2s	4s	-50%
`mlk_ct_sel_int16`	✅	2s	1s	+100%
`mlk_ct_sel_uint8`	✅	2s	1s	+100%
`mlk_keccakf1600_extract_bytes`	✅	2s	3s	-33%
`mlk_keccakf1600_permute`	✅	2s	3s	-33%
`mlk_keccakf1600x4_extract_bytes`	✅	2s	1s	+100%
`mlk_keccakf1600x4_xor_bytes_c`	✅	2s	3s	-33%
`mlk_montgomery_reduce`	✅	2s	3s	-33%
`mlk_poly_add`	✅	2s	1s	+100%
`mlk_poly_compress_d10`	✅	2s	4s	-50%
`mlk_poly_compress_d4`	✅	2s	1s	+100%
`mlk_poly_compress_d4_c`	✅	2s	4s	-50%
`mlk_poly_compress_d5`	✅	2s	3s	-33%
`mlk_poly_compress_du`	✅	2s	1s	+100%
`mlk_poly_compress_dv`	✅	2s	2s	+0%
`mlk_poly_decompress_d10_native`	✅	2s	2s	+0%
`mlk_poly_decompress_d11`	✅	2s	2s	+0%
`mlk_poly_decompress_d11_c`	✅	2s	1s	+100%
`mlk_poly_decompress_d4`	✅	2s	4s	-50%
`mlk_poly_decompress_d4_native`	✅	2s	4s	-50%
`mlk_poly_decompress_dv`	✅	2s	2s	+0%
`mlk_poly_frombytes`	✅	2s	1s	+100%
`mlk_poly_getnoise_eta1122_4x`	✅	2s	2s	+0%
`mlk_poly_invntt_tomont_c`	✅	2s	3s	-33%
`mlk_poly_ntt_c`	✅	2s	3s	-33%
`mlk_poly_tobytes_c`	✅	2s	2s	+0%
`mlk_poly_tomont_c`	✅	2s	3s	-33%
`mlk_poly_tomont_native`	✅	2s	3s	-33%
`mlk_polyvec_decompress_du`	✅	2s	5s	-60%
`mlk_polyvec_frombytes`	✅	2s	3s	-33%
`mlk_polyvec_mulcache_compute`	✅	2s	4s	-50%
`mlk_polyvec_permute_bitrev_to_custom_native`	✅	2s	3s	-33%
`mlk_polyvec_tobytes`	✅	2s	2s	+0%
`mlk_scalar_compress_d5`	✅	2s	1s	+100%
`mlk_scalar_signed_to_unsigned_q`	✅	2s	1s	+100%
`mlk_shake128_squeezeblocks`	✅	2s	1s	+100%
`mlk_value_barrier_i32`	✅	2s	2s	+0%
`mlk_value_barrier_u32`	✅	2s	4s	-50%
`ntt_native_aarch64`	✅	2s	2s	+0%
`poly_compress_d11_native_x86_64`	✅	2s	3s	-33%
`poly_compress_d4_native_x86_64`	✅	2s	2s	+0%
`poly_getnoise_eta1122_4x_native`	✅	2s	1s	+100%
`poly_mulcache_compute_native_aarch64`	✅	2s	3s	-33%
`poly_reduce_native_x86_64`	✅	2s	4s	-50%
`poly_tomont_native_aarch64`	✅	2s	2s	+0%
`polyvec_basemul_acc_montgomery_cached_k2_native_x86_64`	✅	2s	2s	+0%
`polyvec_basemul_acc_montgomery_cached_k4_native_aarch64`	✅	2s	1s	+100%
`polyvec_basemul_acc_montgomery_cached_k4_native_x86_64`	✅	2s	1s	+100%
`keccak_f1600_x4_native_avx2`	✅	1s	1s	+0%
`kem_enc_derand`	✅	1s	5s	-80%
`mlk_ct_cmask_nonzero_u8`	✅	1s	3s	-67%
`mlk_keccakf1600_extract_bytes (big endian)`	✅	1s	2s	-50%
`mlk_keccakf1600x4_permute`	✅	1s	3s	-67%
`mlk_poly_cbd_eta1`	✅	1s	4s	-75%
`mlk_poly_compress_d10_c`	✅	1s	3s	-67%
`mlk_poly_compress_d10_native`	✅	1s	2s	-50%
`mlk_poly_decompress_d10_c`	✅	1s	2s	-50%
`mlk_poly_frombytes_c`	✅	1s	2s	-50%
`mlk_poly_getnoise_eta2`	✅	1s	3s	-67%
`mlk_scalar_compress_d1`	✅	1s	2s	-50%
`mlk_scalar_compress_d11`	✅	1s	3s	-67%
`mlk_scalar_decompress_d10`	✅	1s	2s	-50%
`mlk_scalar_decompress_d5`	✅	1s	2s	-50%
`mlk_sha3_512`	✅	1s	2s	-50%
`mlk_shake128x4_squeezeblocks`	✅	1s	2s	-50%
`mlk_value_barrier_u8`	✅	1s	3s	-67%
`poly_decompress_d4_native_x86_64`	✅	1s	3s	-67%
`poly_tobytes_native_aarch64`	✅	1s	2s	-50%
`polyvec_basemul_acc_montgomery_cached_k3_native_aarch64`	✅	1s	3s	-67%
`rej_uniform_native_aarch64`	✅	1s	5s	-80%
`sys_check_capability`	✅	1s	3s	-67%

oqs-bot · 2026-05-05T08:47:08Z

CBMC Results (ML-KEM-512)

Full Results (191 proofs)

Proof	Status	Current	Previous	Change
`TOTAL`	✅	1443s	1608s	-10.3%
`mlk_indcpa_keypair_derand`	✅	263s	310s	-15%
`mlk_indcpa_enc`	✅	187s	200s	-6%
`mlk_rej_uniform_c`	✅	159s	182s	-13%
`mlk_polyvec_basemul_acc_montgomery_cached_c`	✅	63s	75s	-16%
`mlk_ntt_layer`	✅	43s	52s	-17%
`mlk_poly_rej_uniform`	✅	32s	36s	-11%
`poly_ntt_native`	✅	30s	32s	-6%
`mlk_keccak_squeezeblocks_x4`	✅	28s	35s	-20%
`mlk_poly_reduce_native`	✅	24s	24s	+0%
`mlk_fqmul`	✅	20s	21s	-5%
`keccakf1600x4_permute_native_x4`	✅	18s	17s	+6%
`mlk_indcpa_dec`	✅	17s	18s	-6%
`mlk_poly_decompress_d10_native`	✅	17s	16s	+6%
`mlk_poly_decompress_d4_native`	✅	16s	18s	-11%
`mlk_polyvec_add`	✅	12s	14s	-14%
`mlk_poly_frombytes_native`	✅	10s	12s	-17%
`mlk_keccak_squeeze_once`	✅	9s	10s	-10%
`mlk_poly_frommsg`	✅	9s	13s	-31%
`mlk_keccak_squeezeblocks`	✅	8s	9s	-11%
`mlk_ntt_butterfly_block`	✅	8s	11s	-27%
`mlk_poly_getnoise_eta2`	✅	8s	3s	+167%
`mlk_poly_ntt`	✅	8s	11s	-27%
`mlk_poly_rej_uniform_x4`	✅	8s	10s	-20%
`mlk_keccak_absorb_once_x4`	✅	7s	8s	-12%
`polyvec_basemul_acc_montgomery_cached_native`	✅	7s	7s	+0%
`rej_uniform_native_x86_64`	✅	7s	7s	+0%
`mlk_invntt_layer`	✅	6s	6s	+0%
`mlk_poly_cbd_eta2`	✅	6s	7s	-14%
`mlk_poly_tobytes`	✅	6s	3s	+100%
`poly_decompress_d10_native_x86_64`	✅	6s	5s	+20%
`mlk_keccakf1600_permute_c`	✅	5s	5s	+0%
`mlk_scalar_compress_d11`	✅	5s	2s	+150%
`mlk_shake256x4`	✅	5s	5s	+0%
`poly_frombytes_native_x86_64`	✅	5s	4s	+25%
`sys_check_capability`	✅	5s	2s	+150%
`intt_native_x86_64`	✅	4s	1s	+300%
`keccak_f1600_x4_native_aarch64_v84a`	✅	4s	1s	+300%
`kem_check_pk`	✅	4s	4s	+0%
`kem_dec`	✅	4s	4s	+0%
`kem_keypair_derand`	✅	4s	3s	+33%
`mlk_barrett_reduce`	✅	4s	1s	+300%
`mlk_ct_cmask_nonzero_u8`	✅	4s	1s	+300%
`mlk_ct_get_optblocker_i32`	✅	4s	2s	+100%
`mlk_keccak_absorb_once`	✅	4s	4s	+0%
`mlk_keccakf1600_xor_bytes`	✅	4s	2s	+100%
`mlk_poly_compress_d11_c`	✅	4s	2s	+100%
`mlk_poly_compress_d4_c`	✅	4s	4s	+0%
`mlk_poly_decompress_d4_c`	✅	4s	2s	+100%
`mlk_poly_getnoise_eta1122_4x`	✅	4s	4s	+0%
`mlk_poly_getnoise_eta1_4x`	✅	4s	4s	+0%
`mlk_poly_ntt_c`	✅	4s	4s	+0%
`mlk_poly_tomsg`	✅	4s	1s	+300%
`mlk_polymat_permute_bitrev_to_custom`	✅	4s	4s	+0%
`mlk_scalar_decompress_d10`	✅	4s	2s	+100%
`mlk_scalar_signed_to_unsigned_q`	✅	4s	2s	+100%
`mlk_value_barrier_i32`	✅	4s	4s	+0%
`poly_compress_d5_native_x86_64`	✅	4s	3s	+33%
`poly_decompress_d11_native_x86_64`	✅	4s	4s	+0%
`poly_decompress_d4_native_x86_64`	✅	4s	7s	-43%
`poly_mulcache_compute_native_aarch64`	✅	4s	4s	+0%
`poly_mulcache_compute_native_x86_64`	✅	4s	2s	+100%
`poly_reduce_native_x86_64`	✅	4s	2s	+100%
`keccak_f1600_x4_native_aarch64_v8a_v84a_scalar_hybrid`	✅	3s	2s	+50%
`mlk_check_pct`	✅	3s	3s	+0%
`mlk_ct_get_optblocker_u32`	✅	3s	5s	-40%
`mlk_ct_get_optblocker_u8`	✅	3s	1s	+200%
`mlk_enc_getnoise_eta1_eta2`	✅	3s	4s	-25%
`mlk_gen_matrix_serial`	✅	3s	1s	+200%
`mlk_keccakf1600_permute`	✅	3s	2s	+50%
`mlk_keccakf1600x4_extract_bytes_c`	✅	3s	3s	+0%
`mlk_keccakf1600x4_permute`	✅	3s	4s	-25%
`mlk_keccakf1600x4_xor_bytes`	✅	3s	1s	+200%
`mlk_matvec_mul`	✅	3s	4s	-25%
`mlk_poly_add`	✅	3s	1s	+200%
`mlk_poly_cbd_eta1`	✅	3s	4s	-25%
`mlk_poly_compress_d10`	✅	3s	2s	+50%
`mlk_poly_compress_d4`	✅	3s	3s	+0%
`mlk_poly_compress_dv`	✅	3s	2s	+50%
`mlk_poly_decompress_d10`	✅	3s	2s	+50%
`mlk_poly_decompress_d11_native`	✅	3s	4s	-25%
`mlk_poly_decompress_dv`	✅	3s	3s	+0%
`mlk_poly_mulcache_compute`	✅	3s	3s	+0%
`mlk_poly_mulcache_compute_native`	✅	3s	2s	+50%
`mlk_poly_reduce`	✅	3s	4s	-25%
`mlk_poly_sub`	✅	3s	1s	+200%
`mlk_poly_tobytes_native`	✅	3s	3s	+0%
`mlk_polyvec_basemul_acc_montgomery_cached`	✅	3s	2s	+50%
`mlk_polyvec_mulcache_compute`	✅	3s	2s	+50%
`mlk_polyvec_permute_bitrev_to_custom_native`	✅	3s	3s	+0%
`mlk_rej_uniform`	✅	3s	3s	+0%
`mlk_scalar_compress_d10`	✅	3s	2s	+50%
`mlk_scalar_decompress_d4`	✅	3s	3s	+0%
`mlk_sha3_512`	✅	3s	2s	+50%
`ntt_native_aarch64`	✅	3s	2s	+50%
`nttunpack_native_x86_64`	✅	3s	4s	-25%
`poly_compress_d10_native_x86_64`	✅	3s	1s	+200%
`poly_invntt_tomont_native`	✅	3s	3s	+0%
`polyvec_basemul_acc_montgomery_cached_k2_native_aarch64`	✅	3s	2s	+50%
`polyvec_basemul_acc_montgomery_cached_k2_native_x86_64`	✅	3s	4s	-25%
`polyvec_basemul_acc_montgomery_cached_k4_native_aarch64`	✅	3s	2s	+50%
`rej_uniform_native_aarch64`	✅	3s	5s	-40%
`keccak_f1600_x1_native_aarch64`	✅	2s	2s	+0%
`keccak_f1600_x4_native_aarch64_v8a_scalar_hybrid`	✅	2s	1s	+100%
`keccak_f1600_x4_native_avx2`	✅	2s	2s	+0%
`keccakf1600_permute_native`	✅	2s	2s	+0%
`kem_check_sk`	✅	2s	3s	-33%
`kem_enc_derand`	✅	2s	2s	+0%
`mlk_ct_cmask_neg_i16`	✅	2s	2s	+0%
`mlk_ct_cmask_nonzero_u16`	✅	2s	2s	+0%
`mlk_ct_cmov_zero`	✅	2s	4s	-50%
`mlk_ct_memcmp`	✅	2s	4s	-50%
`mlk_ct_sel_int16`	✅	2s	4s	-50%
`mlk_ct_sel_uint8`	✅	2s	4s	-50%
`mlk_gen_matrix`	✅	2s	3s	-33%
`mlk_keccakf1600_extract_bytes (big endian)`	✅	2s	1s	+100%
`mlk_keccakf1600_xor_bytes (big endian)`	✅	2s	1s	+100%
`mlk_keccakf1600x4_extract_bytes`	✅	2s	2s	+0%
`mlk_keypair_getnoise_eta1`	✅	2s	3s	-33%
`mlk_montgomery_reduce`	✅	2s	6s	-67%
`mlk_poly_compress_d10_c`	✅	2s	3s	-33%
`mlk_poly_compress_d11`	✅	2s	3s	-33%
`mlk_poly_compress_d4_native`	✅	2s	3s	-33%
`mlk_poly_compress_d5`	✅	2s	2s	+0%
`mlk_poly_compress_d5_c`	✅	2s	1s	+100%
`mlk_poly_compress_du`	✅	2s	2s	+0%
`mlk_poly_decompress_d11_c`	✅	2s	3s	-33%
`mlk_poly_decompress_d4`	✅	2s	2s	+0%
`mlk_poly_decompress_d5_c`	✅	2s	1s	+100%
`mlk_poly_decompress_d5_native`	✅	2s	2s	+0%
`mlk_poly_decompress_du`	✅	2s	2s	+0%
`mlk_poly_frombytes`	✅	2s	3s	-33%
`mlk_poly_frombytes_c`	✅	2s	1s	+100%
`mlk_poly_invntt_tomont`	✅	2s	1s	+100%
`mlk_poly_mulcache_compute_c`	✅	2s	5s	-60%
`mlk_poly_tomont`	✅	2s	2s	+0%
`mlk_poly_tomont_c`	✅	2s	2s	+0%
`mlk_poly_tomont_native`	✅	2s	3s	-33%
`mlk_polyvec_ntt`	✅	2s	4s	-50%
`mlk_polyvec_permute_bitrev_to_custom`	✅	2s	4s	-50%
`mlk_polyvec_reduce`	✅	2s	2s	+0%
`mlk_polyvec_tobytes`	✅	2s	3s	-33%
`mlk_scalar_compress_d1`	✅	2s	1s	+100%
`mlk_scalar_compress_d4`	✅	2s	3s	-33%
`mlk_scalar_compress_d5`	✅	2s	3s	-33%
`mlk_scalar_decompress_d11`	✅	2s	3s	-33%
`mlk_scalar_decompress_d5`	✅	2s	3s	-33%
`mlk_sha3_256`	✅	2s	3s	-33%
`mlk_shake128_absorb_once`	✅	2s	3s	-33%
`mlk_shake128x4_absorb_once`	✅	2s	4s	-50%
`mlk_shake256`	✅	2s	1s	+100%
`mlk_value_barrier_u32`	✅	2s	2s	+0%
`mlk_value_barrier_u8`	✅	2s	2s	+0%
`poly_compress_d11_native_x86_64`	✅	2s	3s	-33%
`poly_compress_d4_native_x86_64`	✅	2s	2s	+0%
`poly_getnoise_eta1122_4x_native`	✅	2s	2s	+0%
`poly_reduce_native_aarch64`	✅	2s	3s	-33%
`poly_tomont_native_aarch64`	✅	2s	6s	-67%
`poly_tomont_native_x86_64`	✅	2s	3s	-33%
`polyvec_basemul_acc_montgomery_cached_k3_native_aarch64`	✅	2s	1s	+100%
`polyvec_basemul_acc_montgomery_cached_k3_native_x86_64`	✅	2s	2s	+0%
`polyvec_basemul_acc_montgomery_cached_k4_native_x86_64`	✅	2s	1s	+100%
`rej_uniform_native`	✅	2s	5s	-60%
`intt_native_aarch64`	✅	1s	2s	-50%
`keccak_f1600_x1_native_aarch64_v84a`	✅	1s	1s	+0%
`keccakf1600x4_extract_bytes_native`	✅	1s	2s	-50%
`keccakf1600x4_xor_bytes_native`	✅	1s	2s	-50%
`kem_enc`	✅	1s	4s	-75%
`kem_keypair`	✅	1s	3s	-67%
`mlk_keccakf1600_extract_bytes`	✅	1s	3s	-67%
`mlk_keccakf1600x4_xor_bytes_c`	✅	1s	2s	-50%
`mlk_poly_compress_d10_native`	✅	1s	4s	-75%
`mlk_poly_compress_d11_native`	✅	1s	3s	-67%
`mlk_poly_compress_d5_native`	✅	1s	2s	-50%
`mlk_poly_decompress_d10_c`	✅	1s	1s	+0%
`mlk_poly_decompress_d11`	✅	1s	3s	-67%
`mlk_poly_decompress_d5`	✅	1s	3s	-67%
`mlk_poly_getnoise_eta1_4x_native`	✅	1s	3s	-67%
`mlk_poly_invntt_tomont_c`	✅	1s	3s	-67%
`mlk_poly_reduce_c`	✅	1s	1s	+0%
`mlk_poly_tobytes_c`	✅	1s	1s	+0%
`mlk_polyvec_compress_du`	✅	1s	5s	-80%
`mlk_polyvec_decompress_du`	✅	1s	3s	-67%
`mlk_polyvec_frombytes`	✅	1s	4s	-75%
`mlk_polyvec_invntt_tomont`	✅	1s	2s	-50%
`mlk_polyvec_tomont`	✅	1s	5s	-80%
`mlk_shake128_squeezeblocks`	✅	1s	3s	-67%
`mlk_shake128x4_squeezeblocks`	✅	1s	3s	-67%
`ntt_native_x86_64`	✅	1s	1s	+0%
`poly_decompress_d5_native_x86_64`	✅	1s	4s	-75%
`poly_tobytes_native_aarch64`	✅	1s	1s	+0%
`poly_tobytes_native_x86_64`	✅	1s	4s	-75%

oqs-bot · 2026-05-05T08:48:17Z

CBMC Results (ML-KEM-768)

Full Results (191 proofs)

Proof	Status	Current	Previous	Change
`TOTAL`	✅	1336s	1285s	+4.0%
`mlk_indcpa_keypair_derand`	✅	203s	201s	+1%
`mlk_indcpa_enc`	✅	182s	178s	+2%
`mlk_rej_uniform_c`	✅	145s	126s	+15%
`mlk_polyvec_basemul_acc_montgomery_cached_c`	✅	47s	46s	+2%
`mlk_ntt_layer`	✅	38s	34s	+12%
`mlk_poly_rej_uniform`	✅	32s	30s	+7%
`poly_ntt_native`	✅	31s	26s	+19%
`mlk_keccak_squeezeblocks_x4`	✅	26s	26s	+0%
`mlk_poly_reduce_native`	✅	24s	20s	+20%
`keccakf1600x4_permute_native_x4`	✅	17s	16s	+6%
`mlk_fqmul`	✅	17s	16s	+6%
`polyvec_basemul_acc_montgomery_cached_native`	✅	17s	20s	-15%
`mlk_indcpa_dec`	✅	16s	13s	+23%
`mlk_poly_decompress_d10_native`	✅	15s	14s	+7%
`mlk_poly_decompress_d4_native`	✅	15s	18s	-17%
`mlk_poly_ntt`	✅	11s	11s	+0%
`mlk_poly_frommsg`	✅	10s	9s	+11%
`mlk_keccak_squeezeblocks`	✅	9s	8s	+12%
`mlk_keccak_squeeze_once`	✅	8s	11s	-27%
`mlk_poly_frombytes_native`	✅	8s	9s	-11%
`mlk_polyvec_add`	✅	8s	10s	-20%
`mlk_ntt_butterfly_block`	✅	7s	7s	+0%
`kem_dec`	✅	6s	6s	+0%
`mlk_gen_matrix_serial`	✅	6s	3s	+100%
`mlk_keccak_absorb_once_x4`	✅	6s	7s	-14%
`mlk_poly_rej_uniform_x4`	✅	6s	6s	+0%
`poly_decompress_d4_native_x86_64`	✅	6s	7s	-14%
`kem_enc`	✅	5s	2s	+150%
`mlk_enc_getnoise_eta1_eta2`	✅	5s	1s	+400%
`mlk_keccakf1600_permute_c`	✅	5s	5s	+0%
`mlk_poly_compress_d4_native`	✅	5s	1s	+400%
`mlk_poly_tobytes`	✅	5s	2s	+150%
`mlk_shake128x4_squeezeblocks`	✅	5s	1s	+400%
`poly_invntt_tomont_native`	✅	5s	4s	+25%
`poly_tomont_native_x86_64`	✅	5s	2s	+150%
`rej_uniform_native_x86_64`	✅	5s	6s	-17%
`keccak_f1600_x4_native_aarch64_v8a_scalar_hybrid`	✅	4s	2s	+100%
`keccakf1600x4_extract_bytes_native`	✅	4s	2s	+100%
`kem_enc_derand`	✅	4s	3s	+33%
`mlk_ct_cmov_zero`	✅	4s	2s	+100%
`mlk_gen_matrix`	✅	4s	4s	+0%
`mlk_invntt_layer`	✅	4s	7s	-43%
`mlk_matvec_mul`	✅	4s	4s	+0%
`mlk_poly_cbd_eta1`	✅	4s	4s	+0%
`mlk_poly_compress_d10_c`	✅	4s	3s	+33%
`mlk_poly_compress_d11_c`	✅	4s	1s	+300%
`mlk_poly_decompress_d10_c`	✅	4s	2s	+100%
`mlk_poly_decompress_d11`	✅	4s	2s	+100%
`mlk_poly_reduce`	✅	4s	5s	-20%
`mlk_poly_tobytes_c`	✅	4s	2s	+100%
`mlk_polymat_permute_bitrev_to_custom`	✅	4s	4s	+0%
`mlk_polyvec_invntt_tomont`	✅	4s	3s	+33%
`mlk_polyvec_mulcache_compute`	✅	4s	4s	+0%
`mlk_shake128x4_absorb_once`	✅	4s	3s	+33%
`mlk_shake256x4`	✅	4s	4s	+0%
`poly_compress_d10_native_x86_64`	✅	4s	3s	+33%
`poly_frombytes_native_x86_64`	✅	4s	4s	+0%
`poly_reduce_native_aarch64`	✅	4s	2s	+100%
`polyvec_basemul_acc_montgomery_cached_k3_native_x86_64`	✅	4s	1s	+300%
`polyvec_basemul_acc_montgomery_cached_k4_native_x86_64`	✅	4s	2s	+100%
`rej_uniform_native_aarch64`	✅	4s	3s	+33%
`sys_check_capability`	✅	4s	2s	+100%
`keccak_f1600_x4_native_avx2`	✅	3s	2s	+50%
`kem_check_pk`	✅	3s	4s	-25%
`kem_check_sk`	✅	3s	1s	+200%
`kem_keypair_derand`	✅	3s	1s	+200%
`mlk_ct_cmask_nonzero_u16`	✅	3s	2s	+50%
`mlk_keccak_absorb_once`	✅	3s	4s	-25%
`mlk_keccakf1600_permute`	✅	3s	2s	+50%
`mlk_keccakf1600_xor_bytes`	✅	3s	4s	-25%
`mlk_keccakf1600x4_extract_bytes`	✅	3s	1s	+200%
`mlk_keccakf1600x4_extract_bytes_c`	✅	3s	3s	+0%
`mlk_keypair_getnoise_eta1`	✅	3s	4s	-25%
`mlk_montgomery_reduce`	✅	3s	1s	+200%
`mlk_poly_add`	✅	3s	2s	+50%
`mlk_poly_cbd_eta2`	✅	3s	3s	+0%
`mlk_poly_compress_d11`	✅	3s	1s	+200%
`mlk_poly_compress_d4`	✅	3s	3s	+0%
`mlk_poly_compress_d5`	✅	3s	2s	+50%
`mlk_poly_compress_dv`	✅	3s	2s	+50%
`mlk_poly_decompress_d10`	✅	3s	1s	+200%
`mlk_poly_decompress_d11_native`	✅	3s	1s	+200%
`mlk_poly_frombytes_c`	✅	3s	3s	+0%
`mlk_poly_getnoise_eta2`	✅	3s	3s	+0%
`mlk_poly_ntt_c`	✅	3s	4s	-25%
`mlk_poly_reduce_c`	✅	3s	2s	+50%
`mlk_poly_sub`	✅	3s	2s	+50%
`mlk_poly_tobytes_native`	✅	3s	2s	+50%
`mlk_poly_tomont_native`	✅	3s	5s	-40%
`mlk_polyvec_basemul_acc_montgomery_cached`	✅	3s	2s	+50%
`mlk_polyvec_compress_du`	✅	3s	1s	+200%
`mlk_polyvec_ntt`	✅	3s	1s	+200%
`mlk_polyvec_permute_bitrev_to_custom`	✅	3s	2s	+50%
`mlk_polyvec_tomont`	✅	3s	1s	+200%
`mlk_scalar_compress_d11`	✅	3s	3s	+0%
`mlk_scalar_compress_d5`	✅	3s	3s	+0%
`mlk_scalar_decompress_d10`	✅	3s	3s	+0%
`mlk_scalar_signed_to_unsigned_q`	✅	3s	4s	-25%
`mlk_sha3_256`	✅	3s	2s	+50%
`mlk_value_barrier_u32`	✅	3s	2s	+50%
`ntt_native_aarch64`	✅	3s	1s	+200%
`poly_decompress_d10_native_x86_64`	✅	3s	7s	-57%
`poly_decompress_d5_native_x86_64`	✅	3s	1s	+200%
`poly_getnoise_eta1122_4x_native`	✅	3s	2s	+50%
`polyvec_basemul_acc_montgomery_cached_k2_native_aarch64`	✅	3s	3s	+0%
`intt_native_aarch64`	✅	2s	4s	-50%
`intt_native_x86_64`	✅	2s	3s	-33%
`keccak_f1600_x1_native_aarch64_v84a`	✅	2s	1s	+100%
`keccak_f1600_x4_native_aarch64_v84a`	✅	2s	1s	+100%
`keccak_f1600_x4_native_aarch64_v8a_v84a_scalar_hybrid`	✅	2s	1s	+100%
`keccakf1600_permute_native`	✅	2s	2s	+0%
`keccakf1600x4_xor_bytes_native`	✅	2s	3s	-33%
`kem_keypair`	✅	2s	1s	+100%
`mlk_check_pct`	✅	2s	3s	-33%
`mlk_ct_cmask_nonzero_u8`	✅	2s	1s	+100%
`mlk_ct_get_optblocker_u8`	✅	2s	1s	+100%
`mlk_ct_sel_int16`	✅	2s	4s	-50%
`mlk_ct_sel_uint8`	✅	2s	2s	+0%
`mlk_keccakf1600_extract_bytes`	✅	2s	1s	+100%
`mlk_keccakf1600x4_permute`	✅	2s	2s	+0%
`mlk_keccakf1600x4_xor_bytes`	✅	2s	3s	-33%
`mlk_poly_compress_d11_native`	✅	2s	3s	-33%
`mlk_poly_compress_d4_c`	✅	2s	2s	+0%
`mlk_poly_compress_d5_c`	✅	2s	2s	+0%
`mlk_poly_compress_d5_native`	✅	2s	2s	+0%
`mlk_poly_compress_du`	✅	2s	2s	+0%
`mlk_poly_decompress_d11_c`	✅	2s	2s	+0%
`mlk_poly_decompress_d4_c`	✅	2s	4s	-50%
`mlk_poly_decompress_d5`	✅	2s	2s	+0%
`mlk_poly_decompress_d5_c`	✅	2s	2s	+0%
`mlk_poly_decompress_d5_native`	✅	2s	2s	+0%
`mlk_poly_decompress_du`	✅	2s	1s	+100%
`mlk_poly_frombytes`	✅	2s	1s	+100%
`mlk_poly_getnoise_eta1_4x_native`	✅	2s	4s	-50%
`mlk_poly_invntt_tomont`	✅	2s	1s	+100%
`mlk_poly_invntt_tomont_c`	✅	2s	2s	+0%
`mlk_poly_mulcache_compute`	✅	2s	4s	-50%
`mlk_poly_mulcache_compute_c`	✅	2s	3s	-33%
`mlk_poly_mulcache_compute_native`	✅	2s	2s	+0%
`mlk_poly_tomont`	✅	2s	3s	-33%
`mlk_poly_tomont_c`	✅	2s	2s	+0%
`mlk_polyvec_decompress_du`	✅	2s	3s	-33%
`mlk_polyvec_frombytes`	✅	2s	1s	+100%
`mlk_polyvec_tobytes`	✅	2s	2s	+0%
`mlk_rej_uniform`	✅	2s	1s	+100%
`mlk_scalar_compress_d10`	✅	2s	2s	+0%
`mlk_scalar_decompress_d4`	✅	2s	2s	+0%
`mlk_scalar_decompress_d5`	✅	2s	4s	-50%
`mlk_sha3_512`	✅	2s	4s	-50%
`mlk_shake128_absorb_once`	✅	2s	3s	-33%
`mlk_shake128_squeezeblocks`	✅	2s	2s	+0%
`mlk_shake256`	✅	2s	2s	+0%
`mlk_value_barrier_i32`	✅	2s	1s	+100%
`mlk_value_barrier_u8`	✅	2s	2s	+0%
`nttunpack_native_x86_64`	✅	2s	4s	-50%
`poly_compress_d4_native_x86_64`	✅	2s	2s	+0%
`poly_compress_d5_native_x86_64`	✅	2s	3s	-33%
`poly_reduce_native_x86_64`	✅	2s	2s	+0%
`poly_tobytes_native_x86_64`	✅	2s	2s	+0%
`polyvec_basemul_acc_montgomery_cached_k4_native_aarch64`	✅	2s	1s	+100%
`rej_uniform_native`	✅	2s	4s	-50%
`keccak_f1600_x1_native_aarch64`	✅	1s	2s	-50%
`mlk_barrett_reduce`	✅	1s	3s	-67%
`mlk_ct_cmask_neg_i16`	✅	1s	3s	-67%
`mlk_ct_get_optblocker_i32`	✅	1s	3s	-67%
`mlk_ct_get_optblocker_u32`	✅	1s	2s	-50%
`mlk_ct_memcmp`	✅	1s	5s	-80%
`mlk_keccakf1600_extract_bytes (big endian)`	✅	1s	2s	-50%
`mlk_keccakf1600_xor_bytes (big endian)`	✅	1s	2s	-50%
`mlk_keccakf1600x4_xor_bytes_c`	✅	1s	4s	-75%
`mlk_poly_compress_d10`	✅	1s	1s	+0%
`mlk_poly_compress_d10_native`	✅	1s	3s	-67%
`mlk_poly_decompress_d4`	✅	1s	4s	-75%
`mlk_poly_decompress_dv`	✅	1s	1s	+0%
`mlk_poly_getnoise_eta1122_4x`	✅	1s	1s	+0%
`mlk_poly_getnoise_eta1_4x`	✅	1s	2s	-50%
`mlk_poly_tomsg`	✅	1s	4s	-75%
`mlk_polyvec_permute_bitrev_to_custom_native`	✅	1s	3s	-67%
`mlk_polyvec_reduce`	✅	1s	2s	-50%
`mlk_scalar_compress_d1`	✅	1s	3s	-67%
`mlk_scalar_compress_d4`	✅	1s	4s	-75%
`mlk_scalar_decompress_d11`	✅	1s	2s	-50%
`ntt_native_x86_64`	✅	1s	1s	+0%
`poly_compress_d11_native_x86_64`	✅	1s	2s	-50%
`poly_decompress_d11_native_x86_64`	✅	1s	2s	-50%
`poly_mulcache_compute_native_aarch64`	✅	1s	3s	-67%
`poly_mulcache_compute_native_x86_64`	✅	1s	3s	-67%
`poly_tobytes_native_aarch64`	✅	1s	2s	-50%
`poly_tomont_native_aarch64`	✅	1s	2s	-50%
`polyvec_basemul_acc_montgomery_cached_k2_native_x86_64`	✅	1s	2s	-50%
`polyvec_basemul_acc_montgomery_cached_k3_native_aarch64`	✅	1s	3s	-67%

oqs-bot

Arm Cortex-A72 (Raspberry Pi 4) benchmarks

Details

Benchmark suite	Current: `12036a9`	Previous: `db75353`	Ratio
`ML-KEM-512 keypair`	`50711` cycles	`50835` cycles	`1.00`
`ML-KEM-512 encaps`	`59128` cycles	`58693` cycles	`1.01`
`ML-KEM-512 decaps`	`75104` cycles	`74946` cycles	`1.00`
`ML-KEM-768 keypair`	`86483` cycles	`86333` cycles	`1.00`
`ML-KEM-768 encaps`	`94967` cycles	`95399` cycles	`1.00`
`ML-KEM-768 decaps`	`117949` cycles	`119161` cycles	`0.99`
`ML-KEM-1024 keypair`	`129328` cycles	`130734` cycles	`0.99`
`ML-KEM-1024 encaps`	`141943` cycles	`143257` cycles	`0.99`
`ML-KEM-1024 decaps`	`174407` cycles	`173329` cycles	`1.01`

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

Add a little-endian POWER (ppc64le) native backend targeting POWER9 / ISA 2.07 and above, implementing the forward NTT, inverse NTT, Montgomery reduction, and poly_tomont in vector assembly. Extend scripts/autogen and scripts/simpasm to support the ppc64le backend (including auto-generated consts), add the development sources under dev/ppc64le, and wire up CI and the cross toolchain. Co-authored-by: Basil Hess <bhe@zurich.ibm.com> Signed-off-by: Danny Tsen <dtsen@us.ibm.com>

Comments and renames; no semantic change. - VSX register-naming primer; INTT VSR constant-stash explained. - SAVE_REGS frame layout, byte-offset convention, INTT layer-4 zeta rewinds, barrett_fqmul_4x NTT/INTT asymmetry documented. - barrett_reduce_4x entry-block: "Restore" -> "Materialize" (it pulls VR copies from the VSR stash). - Rename layer macros to CT_Butterfly_*_len{2,4} / GS_Butterfly_* so names describe shape, not hard-coded layer numbers. Lift AddSub_4x into the wrappers so all 7 layers share the same explicit load->mul->addsub->store sequence. - Rename Load/Write helpers for symmetry: LoadPermL{24,44}, PermStoreL{24,44}, Load_4_{High,Low,Both}, Write_4_{Low,High}, Write_8X. - Extract SaveLow/RecoverLow/RestoreGlobals helpers in INTT. - Introduce VS_*_STASH defines for the long-term VSR slots. Auto-derived copies under mlkem/src/native/ppc64le synced in a follow-up. Signed-off-by: Hanno Becker <beckphan@amazon.co.uk>

mkannwischer added the benchmark this PR should be benchmarked in CI label May 5, 2026

oqs-bot reviewed May 5, 2026

View reviewed changes

mkannwischer mentioned this pull request May 5, 2026

Add ppc64le backend (supports p8 and above architectures) #1648

Open

mkannwischer force-pushed the ppc64le branch from ebbaf3e to b698193 Compare May 5, 2026 10:21

mkannwischer force-pushed the ppc64le branch from b698193 to 12036a9 Compare May 21, 2026 14:44

mkannwischer added benchmark this PR should be benchmarked in CI and removed benchmark this PR should be benchmarked in CI labels May 21, 2026

mkannwischer force-pushed the ppc64le branch from 12036a9 to ef3392d Compare June 4, 2026 03:40

Conversation

mkannwischer commented May 5, 2026

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

Mac Mini (M1, 2020) benchmarks

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

ppc64le (POWER10) benchmarks

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

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

AMD EPYC 4th gen (c7a)

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

Intel Xeon 3rd gen (c6i)

Uh oh!

oqs-bot left a comment

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.

Uh oh!

Choose a reason for hiding this comment

AMD EPYC 3rd gen (c6a) (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

AMD EPYC 4th gen (c7a) (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

Intel Xeon 3rd gen (c6i) (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

Graviton3

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 (no-opt)

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

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

oqs-bot commented May 5, 2026 •

edited

Loading

oqs-bot commented May 5, 2026 •

edited

Loading

oqs-bot commented May 5, 2026 •

edited

Loading

oqs-bot left a comment •

edited

Loading