CI: Re-enable benchmarking on Mac Mini/RPi4/A55/BPi by mkannwischer · Pull Request #1641 · pq-code-package/mlkem-native

mkannwischer · 2026-03-28T03:14:03Z

Our benchmarking boxes are back online. This renables benchmarking in CI.

One thing to note is this change: We used to have two runners - the RPi4 and RPi5. Both could run benchmarks locally and also execute benchmarks remotely on the A55 and BPi. Now only the RPi4 can connect to the A55 and BPi, so we can no longer run those in parallel.

oqs-bot

Mac Mini (M1, 2020) benchmarks

Details

Benchmark suite	Current: `ae9162e`	Previous: `81a5e72`	Ratio
`ML-KEM-512 keypair`	`12319` cycles	`12319` cycles	`1`
`ML-KEM-512 encaps`	`14999` cycles	`14999` cycles	`1`
`ML-KEM-512 decaps`	`19552` cycles	`19552` cycles	`1`
`ML-KEM-768 keypair`	`21266` cycles	`21083` cycles	`1.01`
`ML-KEM-768 encaps`	`23873` cycles	`23871` cycles	`1.00`
`ML-KEM-768 decaps`	`30417` cycles	`30414` cycles	`1.00`
`ML-KEM-1024 keypair`	`30329` cycles	`30329` cycles	`1`
`ML-KEM-1024 encaps`	`34573` cycles	`34577` cycles	`1.00`
`ML-KEM-1024 decaps`	`44190` cycles	`44184` cycles	`1.00`

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

oqs-bot

ppc64le (POWER10) benchmarks

Details

Benchmark suite	Current: `ae9162e`	Previous: `e2bc4ad`	Ratio
`ML-KEM-512 keypair`	`59868` cycles	`59900` cycles	`1.00`
`ML-KEM-512 encaps`	`72806` cycles	`72669` cycles	`1.00`
`ML-KEM-512 decaps`	`93080` cycles	`92737` cycles	`1.00`
`ML-KEM-768 keypair`	`99886` cycles	`100517` cycles	`0.99`
`ML-KEM-768 encaps`	`116411` cycles	`117320` cycles	`0.99`
`ML-KEM-768 decaps`	`142507` cycles	`143329` cycles	`0.99`
`ML-KEM-1024 keypair`	`151436` cycles	`152695` cycles	`0.99`
`ML-KEM-1024 encaps`	`170338` cycles	`171817` cycles	`0.99`
`ML-KEM-1024 decaps`	`203058` cycles	`203925` 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 'ppc64le (POWER10) benchmarks'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.03.

Benchmark suite	Current: `d1f8a2b`	Previous: `ddbc21d`	Ratio
`ML-KEM-1024 keypair`	`153341` cycles	`148293` cycles	`1.03`
`ML-KEM-1024 decaps`	`204407` cycles	`198126` cycles	`1.03`

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

oqs-bot

AMD EPYC 3rd gen (c6a)

Details

Benchmark suite	Current: `ae9162e`	Previous: `e2bc4ad`	Ratio
`ML-KEM-512 keypair`	`14239` cycles	`14380` cycles	`0.99`
`ML-KEM-512 encaps`	`16033` cycles	`16061` cycles	`1.00`
`ML-KEM-512 decaps`	`21516` cycles	`21569` cycles	`1.00`
`ML-KEM-768 keypair`	`25215` cycles	`24790` cycles	`1.02`
`ML-KEM-768 encaps`	`25743` cycles	`25549` cycles	`1.01`
`ML-KEM-768 decaps`	`33548` cycles	`33362` cycles	`1.01`
`ML-KEM-1024 keypair`	`37506` cycles	`37273` cycles	`1.01`
`ML-KEM-1024 encaps`	`36160` cycles	`37239` cycles	`0.97`
`ML-KEM-1024 decaps`	`47430` cycles	`46827` cycles	`1.01`

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: `d1f8a2b`	Previous: `ddbc21d`	Ratio
`ML-KEM-768 keypair`	`25523` cycles	`24757` cycles	`1.03`

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

oqs-bot

AMD EPYC 4th gen (c7a)

Details

Benchmark suite	Current: `ae9162e`	Previous: `e2bc4ad`	Ratio
`ML-KEM-512 keypair`	`12873` cycles	`12760` cycles	`1.01`
`ML-KEM-512 encaps`	`14269` cycles	`14242` cycles	`1.00`
`ML-KEM-512 decaps`	`19138` cycles	`19119` cycles	`1.00`
`ML-KEM-768 keypair`	`22512` cycles	`22416` cycles	`1.00`
`ML-KEM-768 encaps`	`23051` cycles	`23106` cycles	`1.00`
`ML-KEM-768 decaps`	`30057` cycles	`30082` cycles	`1.00`
`ML-KEM-1024 keypair`	`34269` cycles	`33049` cycles	`1.04`
`ML-KEM-1024 encaps`	`33042` cycles	`33012` cycles	`1.00`
`ML-KEM-1024 decaps`	`42443` cycles	`42417` 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: `ae9162e`	Previous: `e2bc4ad`	Ratio
`ML-KEM-1024 keypair`	`34269` cycles	`33049` cycles	`1.04`

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

oqs-bot

Graviton4

Details

Benchmark suite	Current: `ae9162e`	Previous: `e2bc4ad`	Ratio
`ML-KEM-512 keypair`	`17678` cycles	`17648` cycles	`1.00`
`ML-KEM-512 encaps`	`20596` cycles	`20598` cycles	`1.00`
`ML-KEM-512 decaps`	`27062` cycles	`27062` cycles	`1`
`ML-KEM-768 keypair`	`29923` cycles	`29899` cycles	`1.00`
`ML-KEM-768 encaps`	`32725` cycles	`32770` cycles	`1.00`
`ML-KEM-768 decaps`	`41983` cycles	`41963` cycles	`1.00`
`ML-KEM-1024 keypair`	`43721` cycles	`43739` cycles	`1.00`
`ML-KEM-1024 encaps`	`48773` cycles	`48648` cycles	`1.00`
`ML-KEM-1024 decaps`	`61383` cycles	`61385` 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: `ae9162e`	Previous: `e2bc4ad`	Ratio
`ML-KEM-512 keypair`	`17586` cycles	`17533` cycles	`1.00`
`ML-KEM-512 encaps`	`19855` cycles	`19866` cycles	`1.00`
`ML-KEM-512 decaps`	`26392` cycles	`26401` cycles	`1.00`
`ML-KEM-768 keypair`	`32478` cycles	`32496` cycles	`1.00`
`ML-KEM-768 encaps`	`31063` cycles	`31064` cycles	`1.00`
`ML-KEM-768 decaps`	`41529` cycles	`41498` cycles	`1.00`
`ML-KEM-1024 keypair`	`44381` cycles	`43981` cycles	`1.01`
`ML-KEM-1024 encaps`	`45691` cycles	`46953` cycles	`0.97`
`ML-KEM-1024 decaps`	`58110` cycles	`58232` 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: `ae9162e`	Previous: `e2bc4ad`	Ratio
`ML-KEM-512 keypair`	`40217` cycles	`40259` cycles	`1.00`
`ML-KEM-512 encaps`	`48343` cycles	`48421` cycles	`1.00`
`ML-KEM-512 decaps`	`62487` cycles	`62612` cycles	`1.00`
`ML-KEM-768 keypair`	`64571` cycles	`63737` cycles	`1.01`
`ML-KEM-768 encaps`	`74921` cycles	`74943` cycles	`1.00`
`ML-KEM-768 decaps`	`93427` cycles	`93583` cycles	`1.00`
`ML-KEM-1024 keypair`	`95273` cycles	`95240` cycles	`1.00`
`ML-KEM-1024 encaps`	`109366` cycles	`109391` cycles	`1.00`
`ML-KEM-1024 decaps`	`132099` cycles	`132163` 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: `ae9162e`	Previous: `e2bc4ad`	Ratio
`ML-KEM-512 keypair`	`28283` cycles	`28250` cycles	`1.00`
`ML-KEM-512 encaps`	`34105` cycles	`34099` cycles	`1.00`
`ML-KEM-512 decaps`	`44373` cycles	`44327` cycles	`1.00`
`ML-KEM-768 keypair`	`47685` cycles	`47614` cycles	`1.00`
`ML-KEM-768 encaps`	`53903` cycles	`53937` cycles	`1.00`
`ML-KEM-768 decaps`	`68354` cycles	`68362` cycles	`1.00`
`ML-KEM-1024 keypair`	`70279` cycles	`70211` cycles	`1.00`
`ML-KEM-1024 encaps`	`78694` cycles	`78756` cycles	`1.00`
`ML-KEM-1024 decaps`	`98456` cycles	`98438` 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: `ae9162e`	Previous: `e2bc4ad`	Ratio
`ML-KEM-512 keypair`	`36614` cycles	`36602` cycles	`1.00`
`ML-KEM-512 encaps`	`43136` cycles	`43061` cycles	`1.00`
`ML-KEM-512 decaps`	`55724` cycles	`55697` cycles	`1.00`
`ML-KEM-768 keypair`	`58661` cycles	`58658` cycles	`1.00`
`ML-KEM-768 encaps`	`67583` cycles	`67501` cycles	`1.00`
`ML-KEM-768 decaps`	`84592` cycles	`84426` cycles	`1.00`
`ML-KEM-1024 keypair`	`89042` cycles	`88986` cycles	`1.00`
`ML-KEM-1024 encaps`	`99195` cycles	`99235` cycles	`1.00`
`ML-KEM-1024 decaps`	`120638` cycles	`120547` cycles	`1.00`

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

oqs-bot

Graviton4 (no-opt)

Details

Benchmark suite	Current: `ae9162e`	Previous: `e2bc4ad`	Ratio
`ML-KEM-512 keypair`	`35448` cycles	`35410` cycles	`1.00`
`ML-KEM-512 encaps`	`40092` cycles	`40115` cycles	`1.00`
`ML-KEM-512 decaps`	`51097` cycles	`51137` cycles	`1.00`
`ML-KEM-768 keypair`	`56742` cycles	`56672` cycles	`1.00`
`ML-KEM-768 encaps`	`64537` cycles	`65151` cycles	`0.99`
`ML-KEM-768 decaps`	`79370` cycles	`79301` cycles	`1.00`
`ML-KEM-1024 keypair`	`87844` cycles	`87869` cycles	`1.00`
`ML-KEM-1024 encaps`	`97111` cycles	`96878` cycles	`1.00`
`ML-KEM-1024 decaps`	`115947` cycles	`115828` cycles	`1.00`

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

oqs-bot

Graviton3

Details

Benchmark suite	Current: `ae9162e`	Previous: `e2bc4ad`	Ratio
`ML-KEM-512 keypair`	`18659` cycles	`18637` cycles	`1.00`
`ML-KEM-512 encaps`	`21882` cycles	`21873` cycles	`1.00`
`ML-KEM-512 decaps`	`28878` cycles	`28862` cycles	`1.00`
`ML-KEM-768 keypair`	`31587` cycles	`31539` cycles	`1.00`
`ML-KEM-768 encaps`	`34742` cycles	`34775` cycles	`1.00`
`ML-KEM-768 decaps`	`44810` cycles	`44775` cycles	`1.00`
`ML-KEM-1024 keypair`	`46073` cycles	`46080` cycles	`1.00`
`ML-KEM-1024 encaps`	`51509` cycles	`51496` cycles	`1.00`
`ML-KEM-1024 decaps`	`65017` cycles	`65027` 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: `ae9162e`	Previous: `e2bc4ad`	Ratio
`ML-KEM-512 keypair`	`45688` cycles	`45672` cycles	`1.00`
`ML-KEM-512 encaps`	`54399` cycles	`54295` cycles	`1.00`
`ML-KEM-512 decaps`	`69751` cycles	`69725` cycles	`1.00`
`ML-KEM-768 keypair`	`74146` cycles	`74093` cycles	`1.00`
`ML-KEM-768 encaps`	`86019` cycles	`85991` cycles	`1.00`
`ML-KEM-768 decaps`	`106602` cycles	`106591` cycles	`1.00`
`ML-KEM-1024 keypair`	`112182` cycles	`112084` cycles	`1.00`
`ML-KEM-1024 encaps`	`124671` cycles	`124581` cycles	`1.00`
`ML-KEM-1024 decaps`	`150632` cycles	`150599` cycles	`1.00`

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

oqs-bot

Graviton3 (no-opt)

Details

Benchmark suite	Current: `ae9162e`	Previous: `e2bc4ad`	Ratio
`ML-KEM-512 keypair`	`38937` cycles	`38894` cycles	`1.00`
`ML-KEM-512 encaps`	`44534` cycles	`44595` cycles	`1.00`
`ML-KEM-512 decaps`	`56597` cycles	`56669` cycles	`1.00`
`ML-KEM-768 keypair`	`62332` cycles	`62297` cycles	`1.00`
`ML-KEM-768 encaps`	`71054` cycles	`72327` cycles	`0.98`
`ML-KEM-768 decaps`	`87337` cycles	`87697` cycles	`1.00`
`ML-KEM-1024 keypair`	`96219` cycles	`96159` cycles	`1.00`
`ML-KEM-1024 encaps`	`106365` cycles	`106138` cycles	`1.00`
`ML-KEM-1024 decaps`	`126801` cycles	`126578` cycles	`1.00`

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

oqs-bot

Graviton2

Details

Benchmark suite	Current: `ae9162e`	Previous: `e2bc4ad`	Ratio
`ML-KEM-512 keypair`	`28248` cycles	`28255` cycles	`1.00`
`ML-KEM-512 encaps`	`34179` cycles	`34110` cycles	`1.00`
`ML-KEM-512 decaps`	`44384` cycles	`44390` cycles	`1.00`
`ML-KEM-768 keypair`	`47717` cycles	`47671` cycles	`1.00`
`ML-KEM-768 encaps`	`54081` cycles	`53906` cycles	`1.00`
`ML-KEM-768 decaps`	`68558` cycles	`68361` cycles	`1.00`
`ML-KEM-1024 keypair`	`70346` cycles	`70281` cycles	`1.00`
`ML-KEM-1024 encaps`	`78742` cycles	`78718` cycles	`1.00`
`ML-KEM-1024 decaps`	`98529` cycles	`98476` cycles	`1.00`

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

oqs-bot

Graviton2 (no-opt)

Details

Benchmark suite	Current: `ae9162e`	Previous: `e2bc4ad`	Ratio
`ML-KEM-512 keypair`	`59189` cycles	`59129` cycles	`1.00`
`ML-KEM-512 encaps`	`68622` cycles	`68640` cycles	`1.00`
`ML-KEM-512 decaps`	`87356` cycles	`87347` cycles	`1.00`
`ML-KEM-768 keypair`	`95479` cycles	`95010` cycles	`1.00`
`ML-KEM-768 encaps`	`110402` cycles	`109687` cycles	`1.01`
`ML-KEM-768 decaps`	`134636` cycles	`134169` cycles	`1.00`
`ML-KEM-1024 keypair`	`148046` cycles	`147998` cycles	`1.00`
`ML-KEM-1024 encaps`	`163740` cycles	`163866` cycles	`1.00`
`ML-KEM-1024 decaps`	`195486` cycles	`195539` cycles	`1.00`

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

oqs-bot

Arm Cortex-A55 (Snapdragon 888) benchmarks

Details

Benchmark suite	Current: `ae9162e`	Previous: `81a5e72`	Ratio
`ML-KEM-512 keypair`	`59731` cycles	`59707` cycles	`1.00`
`ML-KEM-512 encaps`	`67471` cycles	`67375` cycles	`1.00`
`ML-KEM-512 decaps`	`86058` cycles	`86040` cycles	`1.00`
`ML-KEM-768 keypair`	`97394` cycles	`97411` cycles	`1.00`
`ML-KEM-768 encaps`	`110836` cycles	`110976` cycles	`1.00`
`ML-KEM-768 decaps`	`137705` cycles	`138041` cycles	`1.00`
`ML-KEM-1024 keypair`	`154757` cycles	`154703` cycles	`1.00`
`ML-KEM-1024 encaps`	`171629` cycles	`171727` cycles	`1.00`
`ML-KEM-1024 decaps`	`207366` cycles	`208007` cycles	`1.00`

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

oqs-bot · 2026-03-28T03:36:37Z

CBMC Results (ML-KEM-512)

⚠️ Attention Required

Proof	Status	Current	Previous	Change
`poly_ntt_native`	⚠️	44s	25s	+76%

Full Results (190 proofs)

Proof	Status	Current	Previous	Change
`TOTAL`	✅	1218s	1068s	+14.0%
`mlk_indcpa_keypair_derand`	✅	164s	131s	+25%
`mlk_indcpa_enc`	✅	146s	133s	+10%
`mlk_rej_uniform_c`	✅	89s	66s	+35%
`mlk_polyvec_basemul_acc_montgomery_cached_c`	✅	67s	61s	+10%
`poly_ntt_native`	⚠️	44s	25s	+76%
`mlk_poly_rej_uniform`	✅	39s	34s	+15%
`mlk_keccak_squeezeblocks_x4`	✅	25s	24s	+4%
`polyvec_basemul_acc_montgomery_cached_native`	✅	22s	21s	+5%
`keccakf1600x4_permute_native_x4`	✅	18s	21s	-14%
`mlk_poly_reduce_native`	✅	16s	15s	+7%
`mlk_polyvec_add`	✅	15s	14s	+7%
`mlk_ntt_layer`	✅	14s	11s	+27%
`mlk_poly_decompress_d10_native`	✅	14s	12s	+17%
`mlk_poly_decompress_d4_native`	✅	13s	12s	+8%
`mlk_indcpa_dec`	✅	12s	9s	+33%
`mlk_keccak_absorb_once_x4`	✅	10s	8s	+25%
`mlk_poly_frommsg`	✅	10s	11s	-9%
`mlk_polymat_permute_bitrev_to_custom`	✅	10s	7s	+43%
`mlk_ntt_butterfly_block`	✅	9s	7s	+29%
`mlk_poly_frombytes_native`	✅	8s	7s	+14%
`mlk_keccak_squeeze_once`	✅	7s	5s	+40%
`mlk_poly_rej_uniform_x4`	✅	7s	6s	+17%
`rej_uniform_native_aarch64`	✅	7s	3s	+133%
`mlk_fqmul`	✅	6s	7s	-14%
`mlk_poly_mulcache_compute_c`	✅	6s	4s	+50%
`poly_decompress_d10_native_x86_64`	✅	6s	4s	+50%
`kem_dec`	✅	5s	5s	+0%
`mlk_ct_cmask_nonzero_u8`	✅	5s	2s	+150%
`mlk_keccak_squeezeblocks`	✅	5s	6s	-17%
`mlk_keccakf1600x4_extract_bytes`	✅	5s	6s	-17%
`mlk_poly_cbd_eta1`	✅	5s	1s	+400%
`mlk_polyvec_tobytes`	✅	5s	2s	+150%
`mlk_shake128x4_squeezeblocks`	✅	5s	3s	+67%
`mlk_shake256x4`	✅	5s	4s	+25%
`poly_compress_d10_native_x86_64`	✅	5s	4s	+25%
`poly_decompress_d4_native_x86_64`	✅	5s	4s	+25%
`poly_frombytes_native_x86_64`	✅	5s	6s	-17%
`polyvec_basemul_acc_montgomery_cached_k2_native_x86_64`	✅	5s	3s	+67%
`intt_native_x86_64`	✅	4s	2s	+100%
`kem_check_pk`	✅	4s	6s	-33%
`kem_enc_derand`	✅	4s	2s	+100%
`mlk_ct_sel_uint8`	✅	4s	2s	+100%
`mlk_invntt_layer`	✅	4s	6s	-33%
`mlk_keccak_absorb_once`	✅	4s	4s	+0%
`mlk_poly_add`	✅	4s	5s	-20%
`mlk_poly_cbd_eta2`	✅	4s	4s	+0%
`mlk_poly_compress_d11_native`	✅	4s	3s	+33%
`mlk_poly_compress_d4`	✅	4s	2s	+100%
`mlk_poly_compress_d5`	✅	4s	2s	+100%
`mlk_poly_decompress_d10_c`	✅	4s	5s	-20%
`mlk_poly_decompress_dv`	✅	4s	2s	+100%
`mlk_poly_getnoise_eta1122_4x`	✅	4s	3s	+33%
`mlk_poly_mulcache_compute_native`	✅	4s	2s	+100%
`mlk_poly_reduce`	✅	4s	2s	+100%
`mlk_poly_tomsg`	✅	4s	3s	+33%
`mlk_polyvec_permute_bitrev_to_custom`	✅	4s	2s	+100%
`mlk_polyvec_reduce`	✅	4s	3s	+33%
`mlk_scalar_compress_d5`	✅	4s	1s	+300%
`mlk_scalar_decompress_d11`	✅	4s	2s	+100%
`mlk_scalar_decompress_d4`	✅	4s	2s	+100%
`nttunpack_native_x86_64`	✅	4s	1s	+300%
`poly_reduce_native_aarch64`	✅	4s	2s	+100%
`poly_tomont_native_x86_64`	✅	4s	1s	+300%
`rej_uniform_native`	✅	4s	3s	+33%
`keccak_f1600_x4_native_aarch64_v8a_v84a_scalar_hybrid`	✅	3s	1s	+200%
`keccak_f1600_x4_native_avx2`	✅	3s	1s	+200%
`keccakf1600_permute_native`	✅	3s	2s	+50%
`keccakf1600x4_xor_bytes_native`	✅	3s	2s	+50%
`mlk_check_pct`	✅	3s	5s	-40%
`mlk_ct_cmask_neg_i16`	✅	3s	3s	+0%
`mlk_ct_cmask_nonzero_u16`	✅	3s	2s	+50%
`mlk_ct_get_optblocker_i32`	✅	3s	2s	+50%
`mlk_ct_memcmp`	✅	3s	1s	+200%
`mlk_ct_sel_int16`	✅	3s	4s	-25%
`mlk_keypair_getnoise_eta1`	✅	3s	3s	+0%
`mlk_poly_compress_d10_c`	✅	3s	2s	+50%
`mlk_poly_compress_d5_c`	✅	3s	1s	+200%
`mlk_poly_compress_dv`	✅	3s	3s	+0%
`mlk_poly_decompress_d10`	✅	3s	1s	+200%
`mlk_poly_decompress_d11_native`	✅	3s	4s	-25%
`mlk_poly_decompress_d4_c`	✅	3s	2s	+50%
`mlk_poly_decompress_d5_c`	✅	3s	2s	+50%
`mlk_poly_frombytes`	✅	3s	3s	+0%
`mlk_poly_frombytes_c`	✅	3s	4s	-25%
`mlk_poly_getnoise_eta1_4x_native`	✅	3s	3s	+0%
`mlk_poly_invntt_tomont_c`	✅	3s	2s	+50%
`mlk_poly_ntt_c`	✅	3s	3s	+0%
`mlk_poly_tobytes_c`	✅	3s	2s	+50%
`mlk_poly_tomont`	✅	3s	3s	+0%
`mlk_poly_tomont_c`	✅	3s	1s	+200%
`mlk_polyvec_decompress_du`	✅	3s	3s	+0%
`mlk_polyvec_permute_bitrev_to_custom_native`	✅	3s	3s	+0%
`mlk_scalar_decompress_d5`	✅	3s	1s	+200%
`mlk_scalar_signed_to_unsigned_q`	✅	3s	3s	+0%
`mlk_sha3_512`	✅	3s	2s	+50%
`mlk_shake128_absorb_once`	✅	3s	1s	+200%
`mlk_shake128x4_absorb_once`	✅	3s	1s	+200%
`mlk_value_barrier_i32`	✅	3s	1s	+200%
`ntt_native_aarch64`	✅	3s	2s	+50%
`ntt_native_x86_64`	✅	3s	2s	+50%
`poly_compress_d11_native_x86_64`	✅	3s	2s	+50%
`poly_compress_d4_native_x86_64`	✅	3s	2s	+50%
`poly_decompress_d11_native_x86_64`	✅	3s	3s	+0%
`poly_getnoise_eta1122_4x_native`	✅	3s	4s	-25%
`poly_tobytes_native_aarch64`	✅	3s	1s	+200%
`poly_tomont_native_aarch64`	✅	3s	1s	+200%
`polyvec_basemul_acc_montgomery_cached_k3_native_x86_64`	✅	3s	1s	+200%
`keccak_f1600_x1_native_aarch64`	✅	2s	3s	-33%
`keccak_f1600_x1_native_aarch64_v84a`	✅	2s	2s	+0%
`keccak_f1600_x4_native_aarch64_v84a`	✅	2s	1s	+100%
`keccakf1600x4_extract_bytes_native`	✅	2s	1s	+100%
`kem_check_sk`	✅	2s	3s	-33%
`kem_enc`	✅	2s	2s	+0%
`kem_keypair`	✅	2s	3s	-33%
`kem_keypair_derand`	✅	2s	6s	-67%
`mlk_barrett_reduce`	✅	2s	3s	-33%
`mlk_ct_cmov_zero`	✅	2s	2s	+0%
`mlk_gen_matrix_serial`	✅	2s	2s	+0%
`mlk_keccakf1600_extract_bytes`	✅	2s	2s	+0%
`mlk_keccakf1600_permute`	✅	2s	4s	-50%
`mlk_keccakf1600_permute_c`	✅	2s	5s	-60%
`mlk_keccakf1600_xor_bytes (big endian)`	✅	2s	1s	+100%
`mlk_keccakf1600x4_extract_bytes_c`	✅	2s	2s	+0%
`mlk_keccakf1600x4_xor_bytes`	✅	2s	1s	+100%
`mlk_matvec_mul`	✅	2s	2s	+0%
`mlk_montgomery_reduce`	✅	2s	1s	+100%
`mlk_poly_compress_d10`	✅	2s	3s	-33%
`mlk_poly_compress_d10_native`	✅	2s	2s	+0%
`mlk_poly_compress_d11`	✅	2s	1s	+100%
`mlk_poly_compress_d11_c`	✅	2s	3s	-33%
`mlk_poly_compress_d4_c`	✅	2s	1s	+100%
`mlk_poly_compress_d4_native`	✅	2s	2s	+0%
`mlk_poly_compress_d5_native`	✅	2s	2s	+0%
`mlk_poly_decompress_d11_c`	✅	2s	4s	-50%
`mlk_poly_getnoise_eta1_4x`	✅	2s	4s	-50%
`mlk_poly_getnoise_eta2`	✅	2s	1s	+100%
`mlk_poly_invntt_tomont`	✅	2s	2s	+0%
`mlk_poly_mulcache_compute`	✅	2s	2s	+0%
`mlk_poly_ntt`	✅	2s	3s	-33%
`mlk_poly_sub`	✅	2s	2s	+0%
`mlk_poly_tobytes`	✅	2s	4s	-50%
`mlk_poly_tomont_native`	✅	2s	2s	+0%
`mlk_polyvec_basemul_acc_montgomery_cached`	✅	2s	2s	+0%
`mlk_polyvec_invntt_tomont`	✅	2s	2s	+0%
`mlk_polyvec_mulcache_compute`	✅	2s	2s	+0%
`mlk_polyvec_ntt`	✅	2s	3s	-33%
`mlk_polyvec_tomont`	✅	2s	1s	+100%
`mlk_scalar_compress_d1`	✅	2s	4s	-50%
`mlk_scalar_compress_d11`	✅	2s	2s	+0%
`mlk_scalar_compress_d4`	✅	2s	1s	+100%
`mlk_scalar_decompress_d10`	✅	2s	2s	+0%
`mlk_sha3_256`	✅	2s	2s	+0%
`mlk_shake256`	✅	2s	2s	+0%
`mlk_value_barrier_u32`	✅	2s	1s	+100%
`mlk_value_barrier_u8`	✅	2s	3s	-33%
`poly_compress_d5_native_x86_64`	✅	2s	2s	+0%
`poly_decompress_d5_native_x86_64`	✅	2s	2s	+0%
`poly_invntt_tomont_native`	✅	2s	1s	+100%
`poly_mulcache_compute_native_aarch64`	✅	2s	3s	-33%
`poly_reduce_native_x86_64`	✅	2s	2s	+0%
`polyvec_basemul_acc_montgomery_cached_k2_native_aarch64`	✅	2s	2s	+0%
`polyvec_basemul_acc_montgomery_cached_k3_native_aarch64`	✅	2s	3s	-33%
`polyvec_basemul_acc_montgomery_cached_k4_native_x86_64`	✅	2s	2s	+0%
`rej_uniform_native_x86_64`	✅	2s	3s	-33%
`sys_check_capability`	✅	2s	4s	-50%
`intt_native_aarch64`	✅	1s	3s	-67%
`keccak_f1600_x4_native_aarch64_v8a_scalar_hybrid`	✅	1s	1s	+0%
`mlk_ct_get_optblocker_u32`	✅	1s	1s	+0%
`mlk_ct_get_optblocker_u8`	✅	1s	2s	-50%
`mlk_gen_matrix`	✅	1s	5s	-80%
`mlk_keccakf1600_extract_bytes (big endian)`	✅	1s	3s	-67%
`mlk_keccakf1600_xor_bytes`	✅	1s	3s	-67%
`mlk_keccakf1600x4_permute`	✅	1s	1s	+0%
`mlk_keccakf1600x4_xor_bytes_c`	✅	1s	2s	-50%
`mlk_poly_compress_du`	✅	1s	3s	-67%
`mlk_poly_decompress_d11`	✅	1s	2s	-50%
`mlk_poly_decompress_d4`	✅	1s	4s	-75%
`mlk_poly_decompress_d5`	✅	1s	4s	-75%
`mlk_poly_decompress_d5_native`	✅	1s	2s	-50%
`mlk_poly_decompress_du`	✅	1s	3s	-67%
`mlk_poly_reduce_c`	✅	1s	2s	-50%
`mlk_poly_tobytes_native`	✅	1s	3s	-67%
`mlk_polyvec_compress_du`	✅	1s	4s	-75%
`mlk_polyvec_frombytes`	✅	1s	2s	-50%
`mlk_rej_uniform`	✅	1s	2s	-50%
`mlk_scalar_compress_d10`	✅	1s	3s	-67%
`mlk_shake128_squeezeblocks`	✅	1s	1s	+0%
`poly_mulcache_compute_native_x86_64`	✅	1s	4s	-75%
`poly_tobytes_native_x86_64`	✅	1s	1s	+0%
`polyvec_basemul_acc_montgomery_cached_k4_native_aarch64`	✅	1s	1s	+0%

oqs-bot · 2026-03-28T03:36:56Z

CBMC Results (ML-KEM-768)

Full Results (190 proofs)

Proof	Status	Current	Previous	Change
`TOTAL`	✅	1330s	1325s	+0.4%
`mlk_indcpa_enc`	✅	243s	258s	-6%
`mlk_indcpa_keypair_derand`	✅	207s	209s	-1%
`mlk_rej_uniform_c`	✅	70s	69s	+1%
`polyvec_basemul_acc_montgomery_cached_native`	✅	57s	64s	-11%
`mlk_polyvec_basemul_acc_montgomery_cached_c`	✅	48s	47s	+2%
`mlk_poly_rej_uniform`	✅	35s	31s	+13%
`mlk_polyvec_add`	✅	27s	27s	+0%
`poly_ntt_native`	✅	27s	23s	+17%
`mlk_keccak_squeezeblocks_x4`	✅	24s	27s	-11%
`keccakf1600x4_permute_native_x4`	✅	20s	20s	+0%
`mlk_poly_reduce_native`	✅	13s	14s	-7%
`mlk_ntt_layer`	✅	12s	12s	+0%
`mlk_poly_frommsg`	✅	12s	10s	+20%
`mlk_indcpa_dec`	✅	11s	14s	-21%
`mlk_poly_decompress_d10_native`	✅	11s	11s	+0%
`mlk_poly_decompress_d4_native`	✅	11s	12s	-8%
`mlk_keccak_absorb_once_x4`	✅	10s	10s	+0%
`mlk_poly_frombytes_native`	✅	9s	7s	+29%
`mlk_keccak_squeeze_once`	✅	8s	6s	+33%
`mlk_poly_rej_uniform_x4`	✅	8s	7s	+14%
`kem_dec`	✅	7s	5s	+40%
`mlk_keccak_squeezeblocks`	✅	7s	6s	+17%
`mlk_fqmul`	✅	6s	8s	-25%
`mlk_ntt_butterfly_block`	✅	6s	7s	-14%
`mlk_poly_cbd_eta1`	✅	6s	2s	+200%
`mlk_polymat_permute_bitrev_to_custom`	✅	6s	6s	+0%
`keccakf1600x4_extract_bytes_native`	✅	5s	3s	+67%
`mlk_invntt_layer`	✅	5s	5s	+0%
`mlk_keccakf1600_permute_c`	✅	5s	6s	-17%
`mlk_poly_add`	✅	5s	6s	-17%
`mlk_scalar_compress_d10`	✅	5s	3s	+67%
`mlk_scalar_decompress_d10`	✅	5s	3s	+67%
`poly_compress_d5_native_x86_64`	✅	5s	6s	-17%
`poly_decompress_d10_native_x86_64`	✅	5s	8s	-38%
`poly_frombytes_native_x86_64`	✅	5s	4s	+25%
`polyvec_basemul_acc_montgomery_cached_k2_native_x86_64`	✅	5s	4s	+25%
`polyvec_basemul_acc_montgomery_cached_k3_native_x86_64`	✅	5s	2s	+150%
`rej_uniform_native_x86_64`	✅	5s	3s	+67%
`kem_check_pk`	✅	4s	3s	+33%
`mlk_ct_memcmp`	✅	4s	3s	+33%
`mlk_ct_sel_uint8`	✅	4s	2s	+100%
`mlk_gen_matrix`	✅	4s	5s	-20%
`mlk_keccak_absorb_once`	✅	4s	4s	+0%
`mlk_poly_compress_d11`	✅	4s	2s	+100%
`mlk_poly_compress_d4`	✅	4s	3s	+33%
`mlk_poly_compress_d5`	✅	4s	2s	+100%
`mlk_poly_compress_d5_native`	✅	4s	2s	+100%
`mlk_poly_getnoise_eta1_4x_native`	✅	4s	2s	+100%
`mlk_poly_invntt_tomont`	✅	4s	2s	+100%
`mlk_poly_mulcache_compute_c`	✅	4s	4s	+0%
`mlk_shake256x4`	✅	4s	3s	+33%
`mlk_value_barrier_u32`	✅	4s	2s	+100%
`mlk_value_barrier_u8`	✅	4s	1s	+300%
`poly_compress_d11_native_x86_64`	✅	4s	3s	+33%
`poly_decompress_d4_native_x86_64`	✅	4s	5s	-20%
`poly_mulcache_compute_native_x86_64`	✅	4s	3s	+33%
`poly_tomont_native_x86_64`	✅	4s	3s	+33%
`polyvec_basemul_acc_montgomery_cached_k3_native_aarch64`	✅	4s	2s	+100%
`polyvec_basemul_acc_montgomery_cached_k4_native_aarch64`	✅	4s	3s	+33%
`polyvec_basemul_acc_montgomery_cached_k4_native_x86_64`	✅	4s	2s	+100%
`keccak_f1600_x1_native_aarch64_v84a`	✅	3s	1s	+200%
`keccak_f1600_x4_native_aarch64_v8a_v84a_scalar_hybrid`	✅	3s	1s	+200%
`keccak_f1600_x4_native_avx2`	✅	3s	4s	-25%
`keccakf1600_permute_native`	✅	3s	2s	+50%
`mlk_check_pct`	✅	3s	2s	+50%
`mlk_ct_cmask_nonzero_u16`	✅	3s	2s	+50%
`mlk_gen_matrix_serial`	✅	3s	3s	+0%
`mlk_keccakf1600_extract_bytes`	✅	3s	2s	+50%
`mlk_keccakf1600_permute`	✅	3s	2s	+50%
`mlk_keccakf1600x4_extract_bytes`	✅	3s	2s	+50%
`mlk_keccakf1600x4_xor_bytes`	✅	3s	2s	+50%
`mlk_keypair_getnoise_eta1`	✅	3s	3s	+0%
`mlk_montgomery_reduce`	✅	3s	1s	+200%
`mlk_poly_compress_d10_c`	✅	3s	5s	-40%
`mlk_poly_compress_d11_c`	✅	3s	3s	+0%
`mlk_poly_compress_d11_native`	✅	3s	2s	+50%
`mlk_poly_compress_d5_c`	✅	3s	2s	+50%
`mlk_poly_compress_du`	✅	3s	5s	-40%
`mlk_poly_decompress_d11_native`	✅	3s	2s	+50%
`mlk_poly_decompress_d4`	✅	3s	2s	+50%
`mlk_poly_decompress_d5_native`	✅	3s	3s	+0%
`mlk_poly_decompress_dv`	✅	3s	3s	+0%
`mlk_poly_frombytes_c`	✅	3s	2s	+50%
`mlk_poly_reduce`	✅	3s	2s	+50%
`mlk_poly_reduce_c`	✅	3s	3s	+0%
`mlk_poly_sub`	✅	3s	3s	+0%
`mlk_poly_tobytes_native`	✅	3s	2s	+50%
`mlk_poly_tomont`	✅	3s	3s	+0%
`mlk_polyvec_compress_du`	✅	3s	3s	+0%
`mlk_polyvec_mulcache_compute`	✅	3s	2s	+50%
`mlk_polyvec_ntt`	✅	3s	1s	+200%
`mlk_polyvec_reduce`	✅	3s	2s	+50%
`mlk_polyvec_tobytes`	✅	3s	2s	+50%
`mlk_polyvec_tomont`	✅	3s	3s	+0%
`mlk_scalar_compress_d1`	✅	3s	2s	+50%
`mlk_scalar_compress_d5`	✅	3s	2s	+50%
`mlk_scalar_signed_to_unsigned_q`	✅	3s	2s	+50%
`mlk_shake128x4_absorb_once`	✅	3s	2s	+50%
`mlk_shake128x4_squeezeblocks`	✅	3s	3s	+0%
`ntt_native_x86_64`	✅	3s	2s	+50%
`nttunpack_native_x86_64`	✅	3s	2s	+50%
`poly_compress_d10_native_x86_64`	✅	3s	3s	+0%
`poly_decompress_d11_native_x86_64`	✅	3s	3s	+0%
`poly_decompress_d5_native_x86_64`	✅	3s	2s	+50%
`poly_reduce_native_x86_64`	✅	3s	2s	+50%
`rej_uniform_native`	✅	3s	2s	+50%
`intt_native_aarch64`	✅	2s	3s	-33%
`intt_native_x86_64`	✅	2s	2s	+0%
`keccak_f1600_x1_native_aarch64`	✅	2s	2s	+0%
`keccak_f1600_x4_native_aarch64_v84a`	✅	2s	1s	+100%
`keccak_f1600_x4_native_aarch64_v8a_scalar_hybrid`	✅	2s	2s	+0%
`keccakf1600x4_xor_bytes_native`	✅	2s	2s	+0%
`kem_check_sk`	✅	2s	3s	-33%
`kem_enc`	✅	2s	2s	+0%
`kem_enc_derand`	✅	2s	2s	+0%
`mlk_barrett_reduce`	✅	2s	2s	+0%
`mlk_ct_cmask_neg_i16`	✅	2s	3s	-33%
`mlk_ct_cmask_nonzero_u8`	✅	2s	2s	+0%
`mlk_ct_cmov_zero`	✅	2s	3s	-33%
`mlk_ct_get_optblocker_u8`	✅	2s	2s	+0%
`mlk_ct_sel_int16`	✅	2s	2s	+0%
`mlk_keccakf1600x4_permute`	✅	2s	2s	+0%
`mlk_matvec_mul`	✅	2s	3s	-33%
`mlk_poly_cbd_eta2`	✅	2s	1s	+100%
`mlk_poly_compress_d10`	✅	2s	5s	-60%
`mlk_poly_compress_d10_native`	✅	2s	2s	+0%
`mlk_poly_compress_d4_c`	✅	2s	1s	+100%
`mlk_poly_compress_d4_native`	✅	2s	3s	-33%
`mlk_poly_compress_dv`	✅	2s	2s	+0%
`mlk_poly_decompress_d10_c`	✅	2s	3s	-33%
`mlk_poly_decompress_d11`	✅	2s	4s	-50%
`mlk_poly_decompress_d11_c`	✅	2s	2s	+0%
`mlk_poly_decompress_d4_c`	✅	2s	1s	+100%
`mlk_poly_decompress_d5`	✅	2s	2s	+0%
`mlk_poly_getnoise_eta1_4x`	✅	2s	1s	+100%
`mlk_poly_getnoise_eta2`	✅	2s	2s	+0%
`mlk_poly_invntt_tomont_c`	✅	2s	2s	+0%
`mlk_poly_mulcache_compute_native`	✅	2s	2s	+0%
`mlk_poly_ntt`	✅	2s	2s	+0%
`mlk_poly_ntt_c`	✅	2s	1s	+100%
`mlk_poly_tobytes`	✅	2s	2s	+0%
`mlk_poly_tobytes_c`	✅	2s	2s	+0%
`mlk_poly_tomont_c`	✅	2s	3s	-33%
`mlk_poly_tomsg`	✅	2s	1s	+100%
`mlk_polyvec_basemul_acc_montgomery_cached`	✅	2s	3s	-33%
`mlk_polyvec_frombytes`	✅	2s	3s	-33%
`mlk_polyvec_invntt_tomont`	✅	2s	4s	-50%
`mlk_polyvec_permute_bitrev_to_custom`	✅	2s	4s	-50%
`mlk_polyvec_permute_bitrev_to_custom_native`	✅	2s	4s	-50%
`mlk_rej_uniform`	✅	2s	1s	+100%
`mlk_scalar_compress_d4`	✅	2s	2s	+0%
`mlk_scalar_decompress_d11`	✅	2s	2s	+0%
`mlk_scalar_decompress_d5`	✅	2s	2s	+0%
`mlk_sha3_256`	✅	2s	3s	-33%
`mlk_shake128_absorb_once`	✅	2s	3s	-33%
`mlk_shake128_squeezeblocks`	✅	2s	1s	+100%
`mlk_shake256`	✅	2s	1s	+100%
`mlk_value_barrier_i32`	✅	2s	2s	+0%
`poly_invntt_tomont_native`	✅	2s	3s	-33%
`poly_mulcache_compute_native_aarch64`	✅	2s	4s	-50%
`poly_reduce_native_aarch64`	✅	2s	1s	+100%
`poly_tomont_native_aarch64`	✅	2s	5s	-60%
`polyvec_basemul_acc_montgomery_cached_k2_native_aarch64`	✅	2s	4s	-50%
`rej_uniform_native_aarch64`	✅	2s	4s	-50%
`sys_check_capability`	✅	2s	2s	+0%
`kem_keypair`	✅	1s	1s	+0%
`kem_keypair_derand`	✅	1s	4s	-75%
`mlk_ct_get_optblocker_i32`	✅	1s	1s	+0%
`mlk_ct_get_optblocker_u32`	✅	1s	2s	-50%
`mlk_keccakf1600_extract_bytes (big endian)`	✅	1s	2s	-50%
`mlk_keccakf1600_xor_bytes`	✅	1s	1s	+0%
`mlk_keccakf1600_xor_bytes (big endian)`	✅	1s	3s	-67%
`mlk_keccakf1600x4_extract_bytes_c`	✅	1s	1s	+0%
`mlk_keccakf1600x4_xor_bytes_c`	✅	1s	2s	-50%
`mlk_poly_decompress_d10`	✅	1s	1s	+0%
`mlk_poly_decompress_d5_c`	✅	1s	2s	-50%
`mlk_poly_decompress_du`	✅	1s	3s	-67%
`mlk_poly_frombytes`	✅	1s	2s	-50%
`mlk_poly_getnoise_eta1122_4x`	✅	1s	2s	-50%
`mlk_poly_mulcache_compute`	✅	1s	1s	+0%
`mlk_poly_tomont_native`	✅	1s	2s	-50%
`mlk_polyvec_decompress_du`	✅	1s	2s	-50%
`mlk_scalar_compress_d11`	✅	1s	1s	+0%
`mlk_scalar_decompress_d4`	✅	1s	1s	+0%
`mlk_sha3_512`	✅	1s	3s	-67%
`ntt_native_aarch64`	✅	1s	4s	-75%
`poly_compress_d4_native_x86_64`	✅	1s	1s	+0%
`poly_getnoise_eta1122_4x_native`	✅	1s	4s	-75%
`poly_tobytes_native_aarch64`	✅	1s	2s	-50%
`poly_tobytes_native_x86_64`	✅	1s	2s	-50%

oqs-bot · 2026-03-28T03:37:47Z

CBMC Results (ML-KEM-1024)

Full Results (190 proofs)

Proof	Status	Current	Previous	Change
`TOTAL`	✅	1431s	1316s	+8.7%
`mlk_indcpa_enc`	✅	272s	256s	+6%
`polyvec_basemul_acc_montgomery_cached_native`	✅	138s	126s	+10%
`mlk_rej_uniform_c`	✅	95s	76s	+25%
`mlk_indcpa_keypair_derand`	✅	91s	81s	+12%
`mlk_polyvec_basemul_acc_montgomery_cached_c`	✅	77s	71s	+8%
`mlk_poly_rej_uniform`	✅	36s	32s	+12%
`poly_ntt_native`	✅	34s	36s	-6%
`mlk_polyvec_add`	✅	26s	25s	+4%
`mlk_keccak_squeezeblocks_x4`	✅	25s	26s	-4%
`keccakf1600x4_permute_native_x4`	✅	18s	18s	+0%
`mlk_poly_reduce_native`	✅	16s	12s	+33%
`mlk_poly_decompress_d11_native`	✅	15s	12s	+25%
`mlk_polyvec_ntt`	✅	15s	11s	+36%
`mlk_poly_decompress_d5_native`	✅	14s	12s	+17%
`mlk_ntt_layer`	✅	13s	13s	+0%
`mlk_indcpa_dec`	✅	12s	16s	-25%
`mlk_poly_frommsg`	✅	12s	11s	+9%
`mlk_poly_compress_d11_c`	✅	11s	9s	+22%
`mlk_keccak_absorb_once_x4`	✅	10s	8s	+25%
`mlk_poly_frombytes_native`	✅	10s	12s	-17%
`mlk_polymat_permute_bitrev_to_custom`	✅	10s	12s	-17%
`mlk_gen_matrix`	✅	9s	7s	+29%
`mlk_poly_rej_uniform_x4`	✅	9s	7s	+29%
`mlk_keccak_squeezeblocks`	✅	8s	5s	+60%
`mlk_ntt_butterfly_block`	✅	8s	7s	+14%
`kem_dec`	✅	7s	5s	+40%
`mlk_keccak_squeeze_once`	✅	7s	7s	+0%
`mlk_fqmul`	✅	6s	4s	+50%
`mlk_poly_add`	✅	6s	5s	+20%
`mlk_poly_mulcache_compute_c`	✅	6s	4s	+50%
`poly_frombytes_native_x86_64`	✅	6s	3s	+100%
`mlk_gen_matrix_serial`	✅	5s	5s	+0%
`mlk_invntt_layer`	✅	5s	5s	+0%
`mlk_keccak_absorb_once`	✅	5s	4s	+25%
`mlk_poly_compress_d4_native`	✅	5s	2s	+150%
`mlk_poly_tobytes_native`	✅	5s	1s	+400%
`poly_decompress_d5_native_x86_64`	✅	5s	3s	+67%
`intt_native_x86_64`	✅	4s	1s	+300%
`kem_check_pk`	✅	4s	4s	+0%
`kem_keypair_derand`	✅	4s	4s	+0%
`mlk_keccakf1600_permute_c`	✅	4s	3s	+33%
`mlk_keccakf1600_xor_bytes`	✅	4s	1s	+300%
`mlk_keccakf1600x4_xor_bytes_c`	✅	4s	1s	+300%
`mlk_poly_cbd_eta2`	✅	4s	1s	+300%
`mlk_poly_compress_d10_c`	✅	4s	1s	+300%
`mlk_poly_decompress_d4_c`	✅	4s	1s	+300%
`mlk_poly_decompress_d4_native`	✅	4s	2s	+100%
`mlk_poly_decompress_d5_c`	✅	4s	4s	+0%
`mlk_poly_getnoise_eta2`	✅	4s	3s	+33%
`mlk_polyvec_compress_du`	✅	4s	2s	+100%
`mlk_polyvec_permute_bitrev_to_custom`	✅	4s	1s	+300%
`mlk_scalar_compress_d1`	✅	4s	1s	+300%
`mlk_sha3_512`	✅	4s	3s	+33%
`mlk_shake256`	✅	4s	1s	+300%
`mlk_shake256x4`	✅	4s	4s	+0%
`poly_decompress_d11_native_x86_64`	✅	4s	5s	-20%
`poly_mulcache_compute_native_aarch64`	✅	4s	4s	+0%
`poly_mulcache_compute_native_x86_64`	✅	4s	1s	+300%
`poly_tobytes_native_aarch64`	✅	4s	3s	+33%
`polyvec_basemul_acc_montgomery_cached_k2_native_aarch64`	✅	4s	1s	+300%
`polyvec_basemul_acc_montgomery_cached_k2_native_x86_64`	✅	4s	2s	+100%
`polyvec_basemul_acc_montgomery_cached_k4_native_x86_64`	✅	4s	2s	+100%
`sys_check_capability`	✅	4s	3s	+33%
`intt_native_aarch64`	✅	3s	3s	+0%
`keccak_f1600_x1_native_aarch64`	✅	3s	1s	+200%
`keccak_f1600_x4_native_aarch64_v8a_v84a_scalar_hybrid`	✅	3s	2s	+50%
`keccakf1600_permute_native`	✅	3s	1s	+200%
`keccakf1600x4_extract_bytes_native`	✅	3s	2s	+50%
`kem_enc`	✅	3s	1s	+200%
`kem_keypair`	✅	3s	3s	+0%
`mlk_check_pct`	✅	3s	4s	-25%
`mlk_ct_cmask_nonzero_u16`	✅	3s	3s	+0%
`mlk_ct_sel_int16`	✅	3s	4s	-25%
`mlk_ct_sel_uint8`	✅	3s	2s	+50%
`mlk_keccakf1600x4_extract_bytes`	✅	3s	3s	+0%
`mlk_keccakf1600x4_permute`	✅	3s	2s	+50%
`mlk_keccakf1600x4_xor_bytes`	✅	3s	2s	+50%
`mlk_matvec_mul`	✅	3s	1s	+200%
`mlk_montgomery_reduce`	✅	3s	1s	+200%
`mlk_poly_compress_d5`	✅	3s	3s	+0%
`mlk_poly_compress_d5_c`	✅	3s	2s	+50%
`mlk_poly_compress_du`	✅	3s	5s	-40%
`mlk_poly_decompress_d11_c`	✅	3s	3s	+0%
`mlk_poly_ntt_c`	✅	3s	1s	+200%
`mlk_poly_tomont_native`	✅	3s	1s	+200%
`mlk_poly_tomsg`	✅	3s	2s	+50%
`mlk_polyvec_basemul_acc_montgomery_cached`	✅	3s	3s	+0%
`mlk_polyvec_invntt_tomont`	✅	3s	1s	+200%
`mlk_polyvec_mulcache_compute`	✅	3s	3s	+0%
`mlk_polyvec_permute_bitrev_to_custom_native`	✅	3s	3s	+0%
`mlk_polyvec_reduce`	✅	3s	3s	+0%
`mlk_polyvec_tobytes`	✅	3s	2s	+50%
`mlk_polyvec_tomont`	✅	3s	5s	-40%
`mlk_scalar_compress_d5`	✅	3s	2s	+50%
`mlk_scalar_decompress_d5`	✅	3s	3s	+0%
`mlk_sha3_256`	✅	3s	2s	+50%
`mlk_shake128x4_absorb_once`	✅	3s	3s	+0%
`mlk_value_barrier_i32`	✅	3s	3s	+0%
`mlk_value_barrier_u32`	✅	3s	2s	+50%
`mlk_value_barrier_u8`	✅	3s	4s	-25%
`ntt_native_aarch64`	✅	3s	4s	-25%
`ntt_native_x86_64`	✅	3s	2s	+50%
`poly_compress_d11_native_x86_64`	✅	3s	3s	+0%
`poly_reduce_native_x86_64`	✅	3s	2s	+50%
`poly_tobytes_native_x86_64`	✅	3s	2s	+50%
`polyvec_basemul_acc_montgomery_cached_k3_native_x86_64`	✅	3s	2s	+50%
`rej_uniform_native`	✅	3s	4s	-25%
`keccak_f1600_x4_native_aarch64_v84a`	✅	2s	2s	+0%
`keccak_f1600_x4_native_aarch64_v8a_scalar_hybrid`	✅	2s	2s	+0%
`kem_check_sk`	✅	2s	2s	+0%
`mlk_barrett_reduce`	✅	2s	2s	+0%
`mlk_ct_cmask_nonzero_u8`	✅	2s	2s	+0%
`mlk_ct_cmov_zero`	✅	2s	2s	+0%
`mlk_ct_get_optblocker_i32`	✅	2s	2s	+0%
`mlk_ct_get_optblocker_u32`	✅	2s	2s	+0%
`mlk_ct_memcmp`	✅	2s	3s	-33%
`mlk_keccakf1600_extract_bytes`	✅	2s	4s	-50%
`mlk_keccakf1600_extract_bytes (big endian)`	✅	2s	2s	+0%
`mlk_keccakf1600_permute`	✅	2s	3s	-33%
`mlk_keccakf1600_xor_bytes (big endian)`	✅	2s	1s	+100%
`mlk_keccakf1600x4_extract_bytes_c`	✅	2s	2s	+0%
`mlk_poly_cbd_eta1`	✅	2s	4s	-50%
`mlk_poly_compress_d10`	✅	2s	1s	+100%
`mlk_poly_compress_d10_native`	✅	2s	2s	+0%
`mlk_poly_compress_d4_c`	✅	2s	2s	+0%
`mlk_poly_compress_d5_native`	✅	2s	2s	+0%
`mlk_poly_compress_dv`	✅	2s	2s	+0%
`mlk_poly_decompress_d10_c`	✅	2s	3s	-33%
`mlk_poly_decompress_d10_native`	✅	2s	1s	+100%
`mlk_poly_decompress_d11`	✅	2s	4s	-50%
`mlk_poly_decompress_d4`	✅	2s	3s	-33%
`mlk_poly_decompress_du`	✅	2s	2s	+0%
`mlk_poly_decompress_dv`	✅	2s	4s	-50%
`mlk_poly_frombytes`	✅	2s	2s	+0%
`mlk_poly_frombytes_c`	✅	2s	4s	-50%
`mlk_poly_getnoise_eta1122_4x`	✅	2s	1s	+100%
`mlk_poly_getnoise_eta1_4x`	✅	2s	3s	-33%
`mlk_poly_getnoise_eta1_4x_native`	✅	2s	2s	+0%
`mlk_poly_invntt_tomont_c`	✅	2s	2s	+0%
`mlk_poly_mulcache_compute`	✅	2s	3s	-33%
`mlk_poly_mulcache_compute_native`	✅	2s	4s	-50%
`mlk_poly_ntt`	✅	2s	1s	+100%
`mlk_poly_reduce`	✅	2s	1s	+100%
`mlk_poly_reduce_c`	✅	2s	1s	+100%
`mlk_poly_sub`	✅	2s	1s	+100%
`mlk_poly_tobytes`	✅	2s	2s	+0%
`mlk_poly_tobytes_c`	✅	2s	3s	-33%
`mlk_poly_tomont`	✅	2s	2s	+0%
`mlk_poly_tomont_c`	✅	2s	1s	+100%
`mlk_polyvec_frombytes`	✅	2s	3s	-33%
`mlk_rej_uniform`	✅	2s	2s	+0%
`mlk_scalar_compress_d11`	✅	2s	3s	-33%
`mlk_scalar_compress_d4`	✅	2s	3s	-33%
`mlk_scalar_decompress_d11`	✅	2s	2s	+0%
`mlk_scalar_signed_to_unsigned_q`	✅	2s	3s	-33%
`mlk_shake128_absorb_once`	✅	2s	1s	+100%
`mlk_shake128_squeezeblocks`	✅	2s	2s	+0%
`nttunpack_native_x86_64`	✅	2s	2s	+0%
`poly_compress_d4_native_x86_64`	✅	2s	1s	+100%
`poly_compress_d5_native_x86_64`	✅	2s	3s	-33%
`poly_decompress_d10_native_x86_64`	✅	2s	4s	-50%
`poly_decompress_d4_native_x86_64`	✅	2s	2s	+0%
`poly_invntt_tomont_native`	✅	2s	2s	+0%
`poly_tomont_native_aarch64`	✅	2s	2s	+0%
`poly_tomont_native_x86_64`	✅	2s	4s	-50%
`polyvec_basemul_acc_montgomery_cached_k4_native_aarch64`	✅	2s	3s	-33%
`rej_uniform_native_aarch64`	✅	2s	4s	-50%
`keccak_f1600_x1_native_aarch64_v84a`	✅	1s	2s	-50%
`keccak_f1600_x4_native_avx2`	✅	1s	3s	-67%
`keccakf1600x4_xor_bytes_native`	✅	1s	3s	-67%
`kem_enc_derand`	✅	1s	2s	-50%
`mlk_ct_cmask_neg_i16`	✅	1s	4s	-75%
`mlk_ct_get_optblocker_u8`	✅	1s	2s	-50%
`mlk_keypair_getnoise_eta1`	✅	1s	3s	-67%
`mlk_poly_compress_d11`	✅	1s	3s	-67%
`mlk_poly_compress_d11_native`	✅	1s	4s	-75%
`mlk_poly_compress_d4`	✅	1s	1s	+0%
`mlk_poly_decompress_d10`	✅	1s	3s	-67%
`mlk_poly_decompress_d5`	✅	1s	2s	-50%
`mlk_poly_invntt_tomont`	✅	1s	3s	-67%
`mlk_polyvec_decompress_du`	✅	1s	2s	-50%
`mlk_scalar_compress_d10`	✅	1s	3s	-67%
`mlk_scalar_decompress_d10`	✅	1s	1s	+0%
`mlk_scalar_decompress_d4`	✅	1s	2s	-50%
`mlk_shake128x4_squeezeblocks`	✅	1s	3s	-67%
`poly_compress_d10_native_x86_64`	✅	1s	1s	+0%
`poly_getnoise_eta1122_4x_native`	✅	1s	2s	-50%
`poly_reduce_native_aarch64`	✅	1s	1s	+0%
`polyvec_basemul_acc_montgomery_cached_k3_native_aarch64`	✅	1s	4s	-75%
`rej_uniform_native_x86_64`	✅	1s	4s	-75%

oqs-bot

SpacemiT K1 8 (Banana Pi F3) benchmarks

Details

Benchmark suite	Current: `ae9162e`	Previous: `81a5e72`	Ratio
`ML-KEM-512 keypair`	`155527` cycles	`155396` cycles	`1.00`
`ML-KEM-512 encaps`	`163416` cycles	`163248` cycles	`1.00`
`ML-KEM-512 decaps`	`206678` cycles	`206498` cycles	`1.00`
`ML-KEM-768 keypair`	`249881` cycles	`249749` cycles	`1.00`
`ML-KEM-768 encaps`	`270385` cycles	`270217` cycles	`1.00`
`ML-KEM-768 decaps`	`332200` cycles	`332485` cycles	`1.00`
`ML-KEM-1024 keypair`	`395810` cycles	`395420` cycles	`1.00`
`ML-KEM-1024 encaps`	`423624` cycles	`423209` cycles	`1.00`
`ML-KEM-1024 decaps`	`507309` cycles	`506804` cycles	`1.00`

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

oqs-bot

Arm Cortex-A72 (Raspberry Pi 4) benchmarks

Details

Benchmark suite	Current: `ae9162e`	Previous: `81a5e72`	Ratio
`ML-KEM-512 keypair`	`50673` cycles	`51363` cycles	`0.99`
`ML-KEM-512 encaps`	`58542` cycles	`59494` cycles	`0.98`
`ML-KEM-512 decaps`	`75315` cycles	`75347` cycles	`1.00`
`ML-KEM-768 keypair`	`86171` cycles	`86863` cycles	`0.99`
`ML-KEM-768 encaps`	`94320` cycles	`94915` cycles	`0.99`
`ML-KEM-768 decaps`	`117702` cycles	`119808` cycles	`0.98`
`ML-KEM-1024 keypair`	`129513` cycles	`131224` cycles	`0.99`
`ML-KEM-1024 encaps`	`142127` cycles	`144261` cycles	`0.99`
`ML-KEM-1024 decaps`	`174310` cycles	`176677` cycles	`0.99`

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

oqs-bot

Intel Xeon 4th gen (c7i)

Details

Benchmark suite	Current: `ae9162e`	Previous: `e2bc4ad`	Ratio
`ML-KEM-512 keypair`	`12056` cycles	`12037` cycles	`1.00`
`ML-KEM-512 encaps`	`13593` cycles	`13756` cycles	`0.99`
`ML-KEM-512 decaps`	`17746` cycles	`17776` cycles	`1.00`
`ML-KEM-768 keypair`	`21251` cycles	`21044` cycles	`1.01`
`ML-KEM-768 encaps`	`22080` cycles	`22040` cycles	`1.00`
`ML-KEM-768 decaps`	`28477` cycles	`28416` cycles	`1.00`
`ML-KEM-1024 keypair`	`29985` cycles	`29879` cycles	`1.00`
`ML-KEM-1024 encaps`	`31703` cycles	`31716` cycles	`1.00`
`ML-KEM-1024 decaps`	`39427` cycles	`39467` 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: `ae9162e`	Previous: `e2bc4ad`	Ratio
`ML-KEM-512 keypair`	`28235` cycles	`28191` cycles	`1.00`
`ML-KEM-512 encaps`	`36632` cycles	`36609` cycles	`1.00`
`ML-KEM-512 decaps`	`45151` cycles	`45197` cycles	`1.00`
`ML-KEM-768 keypair`	`46233` cycles	`46321` cycles	`1.00`
`ML-KEM-768 encaps`	`55779` cycles	`55810` cycles	`1.00`
`ML-KEM-768 decaps`	`69830` cycles	`69876` cycles	`1.00`
`ML-KEM-1024 keypair`	`70568` cycles	`70306` cycles	`1.00`
`ML-KEM-1024 encaps`	`82482` cycles	`82479` cycles	`1.00`
`ML-KEM-1024 decaps`	`99377` cycles	`98894` cycles	`1.00`

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

Our benchmarking boxes are back online. This renables benchmarking in CI. Signed-off-by: Matthias J. Kannwischer <matthias@zerorisc.com>

Port of pq-code-package/mlkem-native#1641. Signed-off-by: Matthias J. Kannwischer <matthias@zerorisc.com>

mkannwischer added the benchmark this PR should be benchmarked in CI label Mar 28, 2026

oqs-bot reviewed Mar 28, 2026

View reviewed changes

mkannwischer added the needs-mldsa-native-port label Mar 28, 2026

mkannwischer marked this pull request as ready for review March 28, 2026 06:23

mkannwischer requested a review from a team as a code owner March 28, 2026 06:23

oqs-bot reviewed Mar 28, 2026

View reviewed changes

mkannwischer assigned hanno-becker Apr 2, 2026

mkannwischer force-pushed the reenable-bench branch from d1f8a2b to fb93b21 Compare April 5, 2026 03:57

mkannwischer added benchmark this PR should be benchmarked in CI and removed benchmark this PR should be benchmarked in CI labels Apr 5, 2026

oqs-bot reviewed Apr 5, 2026

View reviewed changes

hanno-becker approved these changes Apr 12, 2026

View reviewed changes

CI: Re-enable benchmarking on Mac Mini/RPi4/A55/BPi

ae9162e

Our benchmarking boxes are back online. This renables benchmarking in CI. Signed-off-by: Matthias J. Kannwischer <matthias@zerorisc.com>

hanno-becker force-pushed the reenable-bench branch from fb93b21 to ae9162e Compare April 12, 2026 03:29

mkannwischer added benchmark this PR should be benchmarked in CI and removed benchmark this PR should be benchmarked in CI labels Apr 12, 2026

mkannwischer merged commit 174f15b into main Apr 12, 2026
419 checks passed

mkannwischer deleted the reenable-bench branch April 12, 2026 06:07

mkannwischer added a commit to pq-code-package/mldsa-native that referenced this pull request Apr 12, 2026

CI: Re-enable benchmarking on Mac Mini/RPi4/A55/BPi

640fa71

Port of pq-code-package/mlkem-native#1641. Signed-off-by: Matthias J. Kannwischer <matthias@zerorisc.com>

mkannwischer removed the needs-mldsa-native-port label Apr 12, 2026

mkannwischer mentioned this pull request Apr 12, 2026

CI: Sync with mlkem-native (RISC-V runners, fork guards, examples, benchmarking) pq-code-package/mldsa-native#1034

Merged

hanno-becker pushed a commit to pq-code-package/mldsa-native that referenced this pull request Apr 14, 2026

CI: Re-enable benchmarking on Mac Mini/RPi4/A55/BPi

1f0777f

Port of pq-code-package/mlkem-native#1641. Signed-off-by: Matthias J. Kannwischer <matthias@zerorisc.com>

mkannwischer added a commit to pq-code-package/mldsa-native that referenced this pull request Apr 14, 2026

CI: Re-enable benchmarking on Mac Mini/RPi4/A55/BPi

5228476

Port of pq-code-package/mlkem-native#1641. Signed-off-by: Matthias J. Kannwischer <matthias@zerorisc.com>

hanno-becker pushed a commit to pq-code-package/mldsa-native that referenced this pull request Apr 14, 2026

CI: Re-enable benchmarking on Mac Mini/RPi4/A55/BPi

4d393aa

Port of pq-code-package/mlkem-native#1641. Signed-off-by: Matthias J. Kannwischer <matthias@zerorisc.com>

Conversation

mkannwischer commented Mar 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

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

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

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 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

⚠️ 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

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

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

Arm Cortex-A76 (Raspberry Pi 5) 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

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

Graviton4 (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

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

mkannwischer commented Mar 28, 2026 •

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 Mar 28, 2026 •

edited

Loading

oqs-bot commented Mar 28, 2026 •

edited

Loading

oqs-bot commented Mar 28, 2026 •

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