zephyr: fix NUCLEO benchmark timing and TCM layout by bremoran · Pull Request #1796 · pq-code-package/mlkem-native

bremoran · 2026-07-09T06:56:17Z

Use Zephyr's 64-bit cycle counter so long benchmark windows do not wrap, and make benchmark percentile sorting robust for large uint64_t values.

Link NUCLEO-N657X0-Q Zephyr images as XIP from ITCM with runtime data and stacks in DTCM. Force the STM32N6 VTOR setup to the ITCM vector table and scrub the expanded DTCM window in the early reset hook.

Keep project CFLAGS out of Zephyr's cached global CMake flags, and print the Zephyr build log when a suppressed Ninja build fails.

Tested:

nix develop .#zephyr --command make run_func EXTRA_MAKEFILE=test/zephyr/platform.mk ZEPHYR_TARGET=mps3-an547 CYCLES=NO -j1
nix develop .#zephyr --command make run_bench EXTRA_MAKEFILE=test/zephyr/platform.mk ZEPHYR_TARGET=nucleo-n657x0-q CYCLES=NO -j1

Use Zephyr's 64-bit cycle counter so long benchmark windows do not wrap, and make benchmark percentile sorting robust for large uint64_t values. Link NUCLEO-N657X0-Q Zephyr images as XIP from ITCM with runtime data and stacks in DTCM. Force the STM32N6 VTOR setup to the ITCM vector table and scrub the expanded DTCM window in the early reset hook. Keep project CFLAGS out of Zephyr's cached global CMake flags, and print the Zephyr build log when a suppressed Ninja build fails. Tested: - nix develop .#zephyr --command make run_func EXTRA_MAKEFILE=test/zephyr/platform.mk ZEPHYR_TARGET=mps3-an547 CYCLES=NO -j1 - nix develop .#zephyr --command make run_bench EXTRA_MAKEFILE=test/zephyr/platform.mk ZEPHYR_TARGET=nucleo-n657x0-q CYCLES=NO -j1 Signed-off-by: Brendan Moran <brendan.moran@arm.com>

oqs-bot

Mac Mini (M1, 2020) benchmarks

Details

Benchmark suite	Current: `0e34d09`	Previous: `cfd1912`	Ratio
`ML-KEM-512 keypair`	`12318` cycles	`12317` cycles	`1.00`
`ML-KEM-512 encaps`	`14763` cycles	`14761` cycles	`1.00`
`ML-KEM-512 decaps`	`19318` cycles	`19315` cycles	`1.00`
`ML-KEM-768 keypair`	`21263` cycles	`21263` cycles	`1`
`ML-KEM-768 encaps`	`23510` cycles	`23508` cycles	`1.00`
`ML-KEM-768 decaps`	`30054` cycles	`30054` cycles	`1`
`ML-KEM-1024 keypair`	`30316` cycles	`30315` cycles	`1.00`
`ML-KEM-1024 encaps`	`34082` cycles	`34080` cycles	`1.00`
`ML-KEM-1024 decaps`	`43696` cycles	`43695` cycles	`1.00`

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

oqs-bot

ppc64le (POWER10) benchmarks

Details

Benchmark suite	Current: `0e34d09`	Previous: `cfd1912`	Ratio
`ML-KEM-512 keypair`	`38585` cycles	`37977` cycles	`1.02`
`ML-KEM-512 encaps`	`43426` cycles	`43448` cycles	`1.00`
`ML-KEM-512 decaps`	`54121` cycles	`53423` cycles	`1.01`
`ML-KEM-768 keypair`	`67931` cycles	`67467` cycles	`1.01`
`ML-KEM-768 encaps`	`76428` cycles	`76412` cycles	`1.00`
`ML-KEM-768 decaps`	`91027` cycles	`90618` cycles	`1.00`
`ML-KEM-1024 keypair`	`113397` cycles	`108562` cycles	`1.04`
`ML-KEM-1024 encaps`	`124126` cycles	`118622` cycles	`1.05`
`ML-KEM-1024 decaps`	`142680` cycles	`137281` cycles	`1.04`

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

oqs-bot

⚠️ Performance Alert ⚠️

Possible performance regression was detected for benchmark 'ppc64le (POWER10) benchmarks'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.03.

Benchmark suite	Current: `0e34d09`	Previous: `cfd1912`	Ratio
`ML-KEM-1024 keypair`	`113397` cycles	`108562` cycles	`1.04`
`ML-KEM-1024 encaps`	`124126` cycles	`118622` cycles	`1.05`
`ML-KEM-1024 decaps`	`142680` cycles	`137281` cycles	`1.04`

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

oqs-bot

Arm Cortex-A76 (Raspberry Pi 5) benchmarks

Details

Benchmark suite	Current: `0e34d09`	Previous: `cfd1912`	Ratio
`ML-KEM-512 keypair`	`28219` cycles	`28242` cycles	`1.00`
`ML-KEM-512 encaps`	`34064` cycles	`34125` cycles	`1.00`
`ML-KEM-512 decaps`	`44497` cycles	`44473` cycles	`1.00`
`ML-KEM-768 keypair`	`47586` cycles	`47611` cycles	`1.00`
`ML-KEM-768 encaps`	`53885` cycles	`53862` cycles	`1.00`
`ML-KEM-768 decaps`	`68533` cycles	`68521` cycles	`1.00`
`ML-KEM-1024 keypair`	`70182` cycles	`70247` cycles	`1.00`
`ML-KEM-1024 encaps`	`78707` cycles	`78704` cycles	`1.00`
`ML-KEM-1024 decaps`	`98255` cycles	`98273` cycles	`1.00`

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

oqs-bot

Intel Xeon 4th gen (c7i)

Details

Benchmark suite	Current: `0e34d09`	Previous: `cfd1912`	Ratio
`ML-KEM-512 keypair`	`11928` cycles	`11784` cycles	`1.01`
`ML-KEM-512 encaps`	`13210` cycles	`13079` cycles	`1.01`
`ML-KEM-512 decaps`	`17055` cycles	`17091` cycles	`1.00`
`ML-KEM-768 keypair`	`19451` cycles	`19364` cycles	`1.00`
`ML-KEM-768 encaps`	`20773` cycles	`20596` cycles	`1.01`
`ML-KEM-768 decaps`	`26372` cycles	`26281` cycles	`1.00`
`ML-KEM-1024 keypair`	`28083` cycles	`28109` cycles	`1.00`
`ML-KEM-1024 encaps`	`30214` cycles	`30378` cycles	`0.99`
`ML-KEM-1024 decaps`	`37460` cycles	`37448` 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: `0e34d09`	Previous: `cfd1912`	Ratio
`ML-KEM-512 keypair`	`17952` cycles	`17868` cycles	`1.00`
`ML-KEM-512 encaps`	`20159` cycles	`20057` cycles	`1.01`
`ML-KEM-512 decaps`	`26750` cycles	`26606` cycles	`1.01`
`ML-KEM-768 keypair`	`30092` cycles	`29950` cycles	`1.00`
`ML-KEM-768 encaps`	`34094` cycles	`31996` cycles	`1.07`
`ML-KEM-768 decaps`	`41525` cycles	`41491` cycles	`1.00`
`ML-KEM-1024 keypair`	`42133` cycles	`42078` cycles	`1.00`
`ML-KEM-1024 encaps`	`44892` cycles	`45040` cycles	`1.00`
`ML-KEM-1024 decaps`	`58032` cycles	`57971` 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: `0e34d09`	Previous: `cfd1912`	Ratio
`ML-KEM-768 encaps`	`34094` cycles	`31996` cycles	`1.07`

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

oqs-bot

AMD EPYC 4th gen (c7a)

Details

Benchmark suite	Current: `0e34d09`	Previous: `cfd1912`	Ratio
`ML-KEM-512 keypair`	`12869` cycles	`12801` cycles	`1.01`
`ML-KEM-512 encaps`	`14250` cycles	`14237` cycles	`1.00`
`ML-KEM-512 decaps`	`18993` cycles	`18975` cycles	`1.00`
`ML-KEM-768 keypair`	`21539` cycles	`21566` cycles	`1.00`
`ML-KEM-768 encaps`	`22722` cycles	`22732` cycles	`1.00`
`ML-KEM-768 decaps`	`29783` cycles	`29717` cycles	`1.00`
`ML-KEM-1024 keypair`	`30651` cycles	`30514` cycles	`1.00`
`ML-KEM-1024 encaps`	`32698` cycles	`32678` cycles	`1.00`
`ML-KEM-1024 decaps`	`41861` cycles	`41973` cycles	`1.00`

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

oqs-bot

Graviton5

Details

Benchmark suite	Current: `0e34d09`	Previous: `cfd1912`	Ratio
`ML-KEM-512 keypair`	`13431` cycles	`13452` cycles	`1.00`
`ML-KEM-512 encaps`	`15773` cycles	`15781` cycles	`1.00`
`ML-KEM-512 decaps`	`21298` cycles	`21312` cycles	`1.00`
`ML-KEM-768 keypair`	`22870` cycles	`22907` cycles	`1.00`
`ML-KEM-768 encaps`	`25294` cycles	`25225` cycles	`1.00`
`ML-KEM-768 decaps`	`33170` cycles	`33173` cycles	`1.00`
`ML-KEM-1024 keypair`	`33297` cycles	`33232` cycles	`1.00`
`ML-KEM-1024 encaps`	`36925` cycles	`37044` cycles	`1.00`
`ML-KEM-1024 decaps`	`47246` cycles	`47104` cycles	`1.00`

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

oqs-bot

AMD EPYC 3rd gen (c6a)

Details

Benchmark suite	Current: `0e34d09`	Previous: `cfd1912`	Ratio
`ML-KEM-512 keypair`	`14443` cycles	`14443` cycles	`1`
`ML-KEM-512 encaps`	`15959` cycles	`15927` cycles	`1.00`
`ML-KEM-512 decaps`	`21374` cycles	`21327` cycles	`1.00`
`ML-KEM-768 keypair`	`23679` cycles	`23806` cycles	`0.99`
`ML-KEM-768 encaps`	`25126` cycles	`25250` cycles	`1.00`
`ML-KEM-768 decaps`	`32832` cycles	`33030` cycles	`0.99`
`ML-KEM-1024 keypair`	`33474` cycles	`33344` cycles	`1.00`
`ML-KEM-1024 encaps`	`35929` cycles	`35877` cycles	`1.00`
`ML-KEM-1024 decaps`	`46192` cycles	`46192` cycles	`1`

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

oqs-bot

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

Details

Benchmark suite	Current: `0e34d09`	Previous: `cfd1912`	Ratio
`ML-KEM-512 keypair`	`27760` cycles	`27523` cycles	`1.01`
`ML-KEM-512 encaps`	`34812` cycles	`35607` cycles	`0.98`
`ML-KEM-512 decaps`	`44420` cycles	`44308` cycles	`1.00`
`ML-KEM-768 keypair`	`44243` cycles	`44159` cycles	`1.00`
`ML-KEM-768 encaps`	`54897` cycles	`54927` cycles	`1.00`
`ML-KEM-768 decaps`	`68243` cycles	`68284` cycles	`1.00`
`ML-KEM-1024 keypair`	`67469` cycles	`67850` cycles	`0.99`
`ML-KEM-1024 encaps`	`78941` cycles	`78888` cycles	`1.00`
`ML-KEM-1024 decaps`	`95978` cycles	`96098` cycles	`1.00`

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

oqs-bot

Graviton3

Details

Benchmark suite	Current: `0e34d09`	Previous: `cfd1912`	Ratio
`ML-KEM-512 keypair`	`18651` cycles	`18665` cycles	`1.00`
`ML-KEM-512 encaps`	`21829` cycles	`21857` cycles	`1.00`
`ML-KEM-512 decaps`	`28830` cycles	`28837` cycles	`1.00`
`ML-KEM-768 keypair`	`31535` cycles	`31576` cycles	`1.00`
`ML-KEM-768 encaps`	`34743` cycles	`34667` cycles	`1.00`
`ML-KEM-768 decaps`	`44801` cycles	`44803` cycles	`1.00`
`ML-KEM-1024 keypair`	`46207` cycles	`46143` cycles	`1.00`
`ML-KEM-1024 encaps`	`51380` cycles	`51481` cycles	`1.00`
`ML-KEM-1024 decaps`	`64988` cycles	`64971` cycles	`1.00`

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

oqs-bot

Graviton4

Details

Benchmark suite	Current: `0e34d09`	Previous: `cfd1912`	Ratio
`ML-KEM-512 keypair`	`17669` cycles	`17686` cycles	`1.00`
`ML-KEM-512 encaps`	`20547` cycles	`20594` cycles	`1.00`
`ML-KEM-512 decaps`	`27012` cycles	`27005` cycles	`1.00`
`ML-KEM-768 keypair`	`29871` cycles	`29896` cycles	`1.00`
`ML-KEM-768 encaps`	`32699` cycles	`32630` cycles	`1.00`
`ML-KEM-768 decaps`	`41917` cycles	`41919` cycles	`1.00`
`ML-KEM-1024 keypair`	`43746` cycles	`43717` cycles	`1.00`
`ML-KEM-1024 encaps`	`48596` cycles	`48686` cycles	`1.00`
`ML-KEM-1024 decaps`	`61421` cycles	`61381` 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: `0e34d09`	Previous: `cfd1912`	Ratio
`ML-KEM-512 keypair`	`46921` cycles	`46764` cycles	`1.00`
`ML-KEM-512 encaps`	`55746` cycles	`55568` cycles	`1.00`
`ML-KEM-512 decaps`	`71351` cycles	`71228` cycles	`1.00`
`ML-KEM-768 keypair`	`74387` cycles	`74430` cycles	`1.00`
`ML-KEM-768 encaps`	`85985` cycles	`85934` cycles	`1.00`
`ML-KEM-768 decaps`	`107021` cycles	`107113` cycles	`1.00`
`ML-KEM-1024 keypair`	`111438` cycles	`111586` cycles	`1.00`
`ML-KEM-1024 encaps`	`126243` cycles	`126184` cycles	`1.00`
`ML-KEM-1024 decaps`	`152191` cycles	`152270` cycles	`1.00`

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

oqs-bot

Graviton5 (no-opt)

Details

Benchmark suite	Current: `0e34d09`	Previous: `cfd1912`	Ratio
`ML-KEM-512 keypair`	`27365` cycles	`27472` cycles	`1.00`
`ML-KEM-512 encaps`	`31712` cycles	`31683` cycles	`1.00`
`ML-KEM-512 decaps`	`40427` cycles	`40493` cycles	`1.00`
`ML-KEM-768 keypair`	`44011` cycles	`44191` cycles	`1.00`
`ML-KEM-768 encaps`	`50710` cycles	`50710` cycles	`1`
`ML-KEM-768 decaps`	`62119` cycles	`62158` cycles	`1.00`
`ML-KEM-1024 keypair`	`68334` cycles	`67977` cycles	`1.01`
`ML-KEM-1024 encaps`	`75967` cycles	`76071` cycles	`1.00`
`ML-KEM-1024 decaps`	`90694` cycles	`90823` 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: `0e34d09`	Previous: `cfd1912`	Ratio
`ML-KEM-512 keypair`	`37199` cycles	`36791` cycles	`1.01`
`ML-KEM-512 encaps`	`42849` cycles	`42839` cycles	`1.00`
`ML-KEM-512 decaps`	`55552` cycles	`55590` cycles	`1.00`
`ML-KEM-768 keypair`	`58355` cycles	`58027` cycles	`1.01`
`ML-KEM-768 encaps`	`67050` cycles	`66923` cycles	`1.00`
`ML-KEM-768 decaps`	`83925` cycles	`83737` cycles	`1.00`
`ML-KEM-1024 keypair`	`88483` cycles	`88594` cycles	`1.00`
`ML-KEM-1024 encaps`	`99111` cycles	`99087` cycles	`1.00`
`ML-KEM-1024 decaps`	`120395` 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: `0e34d09`	Previous: `cfd1912`	Ratio
`ML-KEM-512 keypair`	`35230` cycles	`35280` cycles	`1.00`
`ML-KEM-512 encaps`	`40033` cycles	`40025` cycles	`1.00`
`ML-KEM-512 decaps`	`50570` cycles	`50626` cycles	`1.00`
`ML-KEM-768 keypair`	`56691` cycles	`56804` cycles	`1.00`
`ML-KEM-768 encaps`	`64087` cycles	`64045` cycles	`1.00`
`ML-KEM-768 decaps`	`78425` cycles	`78496` cycles	`1.00`
`ML-KEM-1024 keypair`	`87404` cycles	`87379` cycles	`1.00`
`ML-KEM-1024 encaps`	`96484` cycles	`96676` cycles	`1.00`
`ML-KEM-1024 decaps`	`114841` cycles	`114706` 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: `0e34d09`	Previous: `cfd1912`	Ratio
`ML-KEM-512 keypair`	`39913` cycles	`40230` cycles	`0.99`
`ML-KEM-512 encaps`	`48250` cycles	`48379` cycles	`1.00`
`ML-KEM-512 decaps`	`61661` cycles	`62191` cycles	`0.99`
`ML-KEM-768 keypair`	`62550` cycles	`62555` cycles	`1.00`
`ML-KEM-768 encaps`	`74792` cycles	`74853` cycles	`1.00`
`ML-KEM-768 decaps`	`92264` cycles	`92436` cycles	`1.00`
`ML-KEM-1024 keypair`	`95131` cycles	`95010` cycles	`1.00`
`ML-KEM-1024 encaps`	`110061` cycles	`109881` cycles	`1.00`
`ML-KEM-1024 decaps`	`132664` cycles	`132297` cycles	`1.00`

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

oqs-bot

Graviton2

Details

Benchmark suite	Current: `0e34d09`	Previous: `cfd1912`	Ratio
`ML-KEM-512 keypair`	`28211` cycles	`28237` cycles	`1.00`
`ML-KEM-512 encaps`	`34076` cycles	`34121` cycles	`1.00`
`ML-KEM-512 decaps`	`44500` cycles	`44477` cycles	`1.00`
`ML-KEM-768 keypair`	`47592` cycles	`47614` cycles	`1.00`
`ML-KEM-768 encaps`	`53891` cycles	`53872` cycles	`1.00`
`ML-KEM-768 decaps`	`68559` cycles	`68558` cycles	`1.00`
`ML-KEM-1024 keypair`	`70166` cycles	`70259` cycles	`1.00`
`ML-KEM-1024 encaps`	`78645` cycles	`78638` cycles	`1.00`
`ML-KEM-1024 decaps`	`98338` cycles	`98289` cycles	`1.00`

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

oqs-bot

Graviton3 (no-opt)

Details

Benchmark suite	Current: `0e34d09`	Previous: `cfd1912`	Ratio
`ML-KEM-512 keypair`	`38827` cycles	`38899` cycles	`1.00`
`ML-KEM-512 encaps`	`44466` cycles	`44383` cycles	`1.00`
`ML-KEM-512 decaps`	`56099` cycles	`56203` cycles	`1.00`
`ML-KEM-768 keypair`	`62352` cycles	`62495` cycles	`1.00`
`ML-KEM-768 encaps`	`70577` cycles	`70438` cycles	`1.00`
`ML-KEM-768 decaps`	`86233` cycles	`86368` cycles	`1.00`
`ML-KEM-1024 keypair`	`95826` cycles	`95867` cycles	`1.00`
`ML-KEM-1024 encaps`	`105850` cycles	`106090` cycles	`1.00`
`ML-KEM-1024 decaps`	`126006` cycles	`125739` cycles	`1.00`

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

oqs-bot

Graviton2 (no-opt)

Details

Benchmark suite	Current: `0e34d09`	Previous: `cfd1912`	Ratio
`ML-KEM-512 keypair`	`58736` cycles	`58755` cycles	`1.00`
`ML-KEM-512 encaps`	`68552` cycles	`68467` cycles	`1.00`
`ML-KEM-512 decaps`	`87403` cycles	`87387` cycles	`1.00`
`ML-KEM-768 keypair`	`94857` cycles	`94524` cycles	`1.00`
`ML-KEM-768 encaps`	`109220` cycles	`108935` cycles	`1.00`
`ML-KEM-768 decaps`	`134224` cycles	`134030` cycles	`1.00`
`ML-KEM-1024 keypair`	`150872` cycles	`150855` cycles	`1.00`
`ML-KEM-1024 encaps`	`165511` cycles	`165563` cycles	`1.00`
`ML-KEM-1024 decaps`	`198679` cycles	`198753` 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: `0e34d09`	Previous: `cfd1912`	Ratio
`ML-KEM-512 keypair`	`143631` cycles	`143013` cycles	`1.00`
`ML-KEM-512 encaps`	`151961` cycles	`151351` cycles	`1.00`
`ML-KEM-512 decaps`	`190370` cycles	`189644` cycles	`1.00`
`ML-KEM-768 keypair`	`233900` cycles	`233091` cycles	`1.00`
`ML-KEM-768 encaps`	`251060` cycles	`250300` cycles	`1.00`
`ML-KEM-768 decaps`	`305899` cycles	`304927` cycles	`1.00`
`ML-KEM-1024 keypair`	`367025` cycles	`365509` cycles	`1.00`
`ML-KEM-1024 encaps`	`389439` cycles	`387853` cycles	`1.00`
`ML-KEM-1024 decaps`	`460090` cycles	`458266` cycles	`1.00`

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

oqs-bot · 2026-07-09T07:32:43Z

CBMC Results (ML-KEM-768)

Full Results (191 proofs)

Proof	Status	Current	Previous	Change
`TOTAL`	✅	1742s	1768s	-1.5%
`mlk_indcpa_keypair_derand`	✅	338s	342s	-1%
`mlk_indcpa_enc`	✅	259s	252s	+3%
`mlk_poly_rej_uniform`	✅	141s	148s	-5%
`mlk_rej_uniform_c`	✅	127s	128s	-1%
`rej_uniform_native_x86_64`	✅	58s	61s	-5%
`mlk_polyvec_basemul_acc_montgomery_cached_c`	✅	52s	56s	-7%
`poly_ntt_native`	✅	40s	45s	-11%
`rej_uniform_native_aarch64`	✅	40s	38s	+5%
`mlk_poly_reduce_native`	✅	34s	36s	-6%
`mlk_ntt_layer`	✅	32s	41s	-22%
`mlk_keccak_squeezeblocks_x4`	✅	26s	24s	+8%
`keccakf1600x4_permute_native_x4`	✅	18s	17s	+6%
`mlk_fqmul`	✅	17s	17s	+0%
`polyvec_basemul_acc_montgomery_cached_native`	✅	17s	16s	+6%
`mlk_poly_decompress_d10_native`	✅	16s	13s	+23%
`mlk_poly_decompress_d4_native`	✅	15s	16s	-6%
`mlk_poly_frombytes_native`	✅	10s	8s	+25%
`mlk_polyvec_add`	✅	10s	12s	-17%
`mlk_indcpa_dec`	✅	9s	11s	-18%
`mlk_keccak_squeezeblocks`	✅	9s	9s	+0%
`mlk_poly_frommsg`	✅	9s	10s	-10%
`mlk_poly_sub`	✅	8s	8s	+0%
`mlk_keccak_squeeze_once`	✅	7s	7s	+0%
`mlk_poly_ntt`	✅	7s	9s	-22%
`kem_dec`	✅	6s	2s	+200%
`mlk_invntt_layer`	✅	6s	5s	+20%
`mlk_keccak_absorb_once_x4`	✅	6s	6s	+0%
`mlk_ntt_butterfly_block`	✅	6s	8s	-25%
`mlk_poly_compress_d10_c`	✅	6s	2s	+200%
`mlk_poly_rej_uniform_x4`	✅	6s	6s	+0%
`mlk_gen_matrix`	✅	5s	5s	+0%
`mlk_gen_matrix_serial`	✅	5s	2s	+150%
`mlk_keccakf1600_permute_c`	✅	5s	5s	+0%
`mlk_poly_add`	✅	5s	5s	+0%
`mlk_poly_getnoise_eta2`	✅	5s	2s	+150%
`mlk_shake256x4`	✅	5s	5s	+0%
`poly_decompress_d4_native_x86_64`	✅	5s	5s	+0%
`poly_decompress_d5_native_x86_64`	✅	5s	2s	+150%
`poly_frombytes_native_x86_64`	✅	5s	4s	+25%
`keccak_f1600_x1_native_aarch64`	✅	4s	1s	+300%
`keccak_f1600_x4_native_aarch64_v84a`	✅	4s	2s	+100%
`kem_enc`	✅	4s	1s	+300%
`mlk_ct_cmask_nonzero_u8`	✅	4s	2s	+100%
`mlk_ct_sel_uint8`	✅	4s	3s	+33%
`mlk_enc_getnoise_eta1_eta2`	✅	4s	2s	+100%
`mlk_keccak_absorb_once`	✅	4s	3s	+33%
`mlk_keccakf1600_xor_bytes`	✅	4s	3s	+33%
`mlk_poly_compress_d4`	✅	4s	5s	-20%
`mlk_poly_compress_d5_native`	✅	4s	3s	+33%
`mlk_poly_decompress_d4_c`	✅	4s	5s	-20%
`mlk_poly_frombytes_c`	✅	4s	5s	-20%
`mlk_poly_getnoise_eta1_4x`	✅	4s	3s	+33%
`mlk_poly_invntt_tomont_c`	✅	4s	2s	+100%
`mlk_poly_reduce_c`	✅	4s	2s	+100%
`mlk_scalar_compress_d11`	✅	4s	2s	+100%
`nttunpack_native_x86_64`	✅	4s	3s	+33%
`poly_decompress_d10_native_x86_64`	✅	4s	6s	-33%
`polyvec_basemul_acc_montgomery_cached_k3_native_x86_64`	✅	4s	5s	-20%
`kem_check_pk`	✅	3s	3s	+0%
`kem_enc_derand`	✅	3s	3s	+0%
`mlk_barrett_reduce`	✅	3s	1s	+200%
`mlk_ct_get_optblocker_i32`	✅	3s	3s	+0%
`mlk_ct_get_optblocker_u32`	✅	3s	1s	+200%
`mlk_keccakf1600x4_extract_bytes`	✅	3s	4s	-25%
`mlk_keccakf1600x4_permute`	✅	3s	2s	+50%
`mlk_keccakf1600x4_xor_bytes`	✅	3s	3s	+0%
`mlk_keccakf1600x4_xor_bytes_c`	✅	3s	2s	+50%
`mlk_montgomery_reduce`	✅	3s	1s	+200%
`mlk_poly_cbd_eta1`	✅	3s	1s	+200%
`mlk_poly_cbd_eta2`	✅	3s	3s	+0%
`mlk_poly_compress_d10_native`	✅	3s	1s	+200%
`mlk_poly_compress_d11`	✅	3s	2s	+50%
`mlk_poly_compress_d4_c`	✅	3s	4s	-25%
`mlk_poly_compress_d5_c`	✅	3s	2s	+50%
`mlk_poly_decompress_d10_c`	✅	3s	3s	+0%
`mlk_poly_decompress_d4`	✅	3s	3s	+0%
`mlk_poly_decompress_d5_c`	✅	3s	2s	+50%
`mlk_poly_decompress_d5_native`	✅	3s	2s	+50%
`mlk_poly_getnoise_eta1_4x_native`	✅	3s	5s	-40%
`mlk_poly_invntt_tomont`	✅	3s	6s	-50%
`mlk_poly_mulcache_compute`	✅	3s	3s	+0%
`mlk_poly_mulcache_compute_c`	✅	3s	4s	-25%
`mlk_poly_ntt_c`	✅	3s	3s	+0%
`mlk_poly_tobytes`	✅	3s	3s	+0%
`mlk_polyvec_basemul_acc_montgomery_cached`	✅	3s	3s	+0%
`mlk_polyvec_frombytes`	✅	3s	1s	+200%
`mlk_polyvec_invntt_tomont`	✅	3s	3s	+0%
`mlk_polyvec_permute_bitrev_to_custom`	✅	3s	5s	-40%
`mlk_polyvec_permute_bitrev_to_custom_native`	✅	3s	3s	+0%
`mlk_polyvec_reduce`	✅	3s	2s	+50%
`mlk_polyvec_tobytes`	✅	3s	4s	-25%
`mlk_scalar_compress_d1`	✅	3s	1s	+200%
`mlk_scalar_decompress_d10`	✅	3s	4s	-25%
`mlk_scalar_decompress_d11`	✅	3s	2s	+50%
`mlk_scalar_decompress_d5`	✅	3s	4s	-25%
`mlk_sha3_256`	✅	3s	1s	+200%
`mlk_shake128_absorb_once`	✅	3s	1s	+200%
`mlk_value_barrier_i32`	✅	3s	3s	+0%
`mlk_value_barrier_u8`	✅	3s	3s	+0%
`poly_compress_d4_native_x86_64`	✅	3s	2s	+50%
`poly_compress_d5_native_x86_64`	✅	3s	1s	+200%
`poly_decompress_d11_native_x86_64`	✅	3s	3s	+0%
`poly_mulcache_compute_native_aarch64`	✅	3s	2s	+50%
`poly_mulcache_compute_native_x86_64`	✅	3s	2s	+50%
`polyvec_basemul_acc_montgomery_cached_k2_native_aarch64`	✅	3s	2s	+50%
`polyvec_basemul_acc_montgomery_cached_k2_native_x86_64`	✅	3s	3s	+0%
`polyvec_basemul_acc_montgomery_cached_k4_native_aarch64`	✅	3s	1s	+200%
`intt_native_aarch64`	✅	2s	3s	-33%
`intt_native_x86_64`	✅	2s	4s	-50%
`keccak_f1600_x1_native_aarch64_v84a`	✅	2s	3s	-33%
`keccakf1600_permute_native`	✅	2s	3s	-33%
`keccakf1600x4_xor_bytes_native`	✅	2s	4s	-50%
`kem_keypair`	✅	2s	1s	+100%
`kem_keypair_derand`	✅	2s	3s	-33%
`mlk_ct_cmask_neg_i16`	✅	2s	1s	+100%
`mlk_ct_cmask_nonzero_u16`	✅	2s	2s	+0%
`mlk_ct_cmov_zero`	✅	2s	3s	-33%
`mlk_keccakf1600_extract_bytes (big endian)`	✅	2s	3s	-33%
`mlk_keccakf1600_permute`	✅	2s	2s	+0%
`mlk_keccakf1600x4_extract_bytes_c`	✅	2s	2s	+0%
`mlk_keypair_getnoise_eta1`	✅	2s	3s	-33%
`mlk_matvec_mul`	✅	2s	3s	-33%
`mlk_poly_compress_d10`	✅	2s	3s	-33%
`mlk_poly_compress_d11_native`	✅	2s	1s	+100%
`mlk_poly_compress_d4_native`	✅	2s	3s	-33%
`mlk_poly_compress_d5`	✅	2s	3s	-33%
`mlk_poly_decompress_d10`	✅	2s	2s	+0%
`mlk_poly_decompress_d11`	✅	2s	1s	+100%
`mlk_poly_decompress_d11_c`	✅	2s	2s	+0%
`mlk_poly_decompress_d5`	✅	2s	4s	-50%
`mlk_poly_decompress_du`	✅	2s	2s	+0%
`mlk_poly_frombytes`	✅	2s	2s	+0%
`mlk_poly_getnoise_eta1122_4x`	✅	2s	1s	+100%
`mlk_poly_mulcache_compute_native`	✅	2s	1s	+100%
`mlk_poly_reduce`	✅	2s	3s	-33%
`mlk_poly_tobytes_c`	✅	2s	2s	+0%
`mlk_poly_tobytes_native`	✅	2s	1s	+100%
`mlk_poly_tomont`	✅	2s	4s	-50%
`mlk_poly_tomont_c`	✅	2s	2s	+0%
`mlk_poly_tomsg`	✅	2s	6s	-67%
`mlk_polymat_permute_bitrev_to_custom`	✅	2s	3s	-33%
`mlk_polyvec_compress_du`	✅	2s	3s	-33%
`mlk_polyvec_decompress_du`	✅	2s	3s	-33%
`mlk_polyvec_mulcache_compute`	✅	2s	3s	-33%
`mlk_polyvec_ntt`	✅	2s	2s	+0%
`mlk_polyvec_tomont`	✅	2s	2s	+0%
`mlk_scalar_compress_d4`	✅	2s	3s	-33%
`mlk_scalar_compress_d5`	✅	2s	3s	-33%
`mlk_sha3_512`	✅	2s	2s	+0%
`mlk_shake128_squeezeblocks`	✅	2s	2s	+0%
`mlk_shake128x4_squeezeblocks`	✅	2s	2s	+0%
`mlk_shake256`	✅	2s	3s	-33%
`ntt_native_x86_64`	✅	2s	2s	+0%
`poly_compress_d10_native_x86_64`	✅	2s	2s	+0%
`poly_compress_d11_native_x86_64`	✅	2s	3s	-33%
`poly_getnoise_eta1122_4x_native`	✅	2s	2s	+0%
`poly_invntt_tomont_native`	✅	2s	4s	-50%
`poly_tobytes_native_aarch64`	✅	2s	3s	-33%
`poly_tobytes_native_x86_64`	✅	2s	3s	-33%
`poly_tomont_native_aarch64`	✅	2s	1s	+100%
`poly_tomont_native_x86_64`	✅	2s	3s	-33%
`polyvec_basemul_acc_montgomery_cached_k3_native_aarch64`	✅	2s	1s	+100%
`keccak_f1600_x4_native_aarch64_v8a_scalar_hybrid`	✅	1s	1s	+0%
`keccak_f1600_x4_native_aarch64_v8a_v84a_scalar_hybrid`	✅	1s	2s	-50%
`keccak_f1600_x4_native_avx2`	✅	1s	2s	-50%
`keccakf1600x4_extract_bytes_native`	✅	1s	1s	+0%
`kem_check_sk`	✅	1s	5s	-80%
`mlk_check_pct`	✅	1s	1s	+0%
`mlk_ct_get_optblocker_u8`	✅	1s	1s	+0%
`mlk_ct_memcmp`	✅	1s	1s	+0%
`mlk_ct_sel_int16`	✅	1s	2s	-50%
`mlk_keccakf1600_extract_bytes`	✅	1s	4s	-75%
`mlk_keccakf1600_xor_bytes (big endian)`	✅	1s	3s	-67%
`mlk_poly_compress_d11_c`	✅	1s	2s	-50%
`mlk_poly_compress_du`	✅	1s	1s	+0%
`mlk_poly_compress_dv`	✅	1s	2s	-50%
`mlk_poly_decompress_d11_native`	✅	1s	1s	+0%
`mlk_poly_decompress_dv`	✅	1s	2s	-50%
`mlk_poly_tomont_native`	✅	1s	2s	-50%
`mlk_rej_uniform`	✅	1s	2s	-50%
`mlk_scalar_compress_d10`	✅	1s	3s	-67%
`mlk_scalar_decompress_d4`	✅	1s	2s	-50%
`mlk_scalar_signed_to_unsigned_q`	✅	1s	2s	-50%
`mlk_shake128x4_absorb_once`	✅	1s	2s	-50%
`mlk_value_barrier_u32`	✅	1s	1s	+0%
`ntt_native_aarch64`	✅	1s	2s	-50%
`poly_reduce_native_aarch64`	✅	1s	1s	+0%
`poly_reduce_native_x86_64`	✅	1s	3s	-67%
`polyvec_basemul_acc_montgomery_cached_k4_native_x86_64`	✅	1s	3s	-67%
`rej_uniform_native`	✅	1s	3s	-67%
`sys_check_capability`	✅	1s	1s	+0%

oqs-bot · 2026-07-09T07:32:52Z

CBMC Results (ML-KEM-512)

Full Results (191 proofs)

Proof	Status	Current	Previous	Change
`TOTAL`	✅	1507s	1605s	-6.1%
`mlk_indcpa_enc`	✅	237s	259s	-8%
`mlk_indcpa_keypair_derand`	✅	194s	215s	-10%
`mlk_poly_rej_uniform`	✅	130s	143s	-9%
`mlk_rej_uniform_c`	✅	114s	123s	-7%
`rej_uniform_native_x86_64`	✅	57s	63s	-10%
`mlk_polyvec_basemul_acc_montgomery_cached_c`	✅	51s	53s	-4%
`mlk_poly_reduce_native`	✅	36s	34s	+6%
`poly_ntt_native`	✅	36s	40s	-10%
`rej_uniform_native_aarch64`	✅	35s	38s	-8%
`mlk_ntt_layer`	✅	27s	34s	-21%
`mlk_keccak_squeezeblocks_x4`	✅	24s	24s	+0%
`keccakf1600x4_permute_native_x4`	✅	17s	17s	+0%
`mlk_fqmul`	✅	16s	16s	+0%
`mlk_indcpa_dec`	✅	14s	16s	-12%
`mlk_poly_decompress_d10_native`	✅	14s	14s	+0%
`mlk_polyvec_add`	✅	14s	12s	+17%
`mlk_poly_decompress_d4_native`	✅	11s	15s	-27%
`mlk_poly_frommsg`	✅	9s	10s	-10%
`polyvec_basemul_acc_montgomery_cached_native`	✅	9s	8s	+12%
`mlk_keccak_squeeze_once`	✅	8s	7s	+14%
`mlk_keccak_squeezeblocks`	✅	8s	8s	+0%
`mlk_ntt_butterfly_block`	✅	8s	8s	+0%
`mlk_poly_frombytes_native`	✅	8s	9s	-11%
`mlk_poly_sub`	✅	8s	7s	+14%
`poly_decompress_d10_native_x86_64`	✅	7s	7s	+0%
`mlk_keccak_absorb_once_x4`	✅	6s	6s	+0%
`mlk_poly_add`	✅	6s	5s	+20%
`mlk_poly_rej_uniform_x4`	✅	6s	7s	-14%
`kem_check_sk`	✅	5s	2s	+150%
`mlk_keccak_absorb_once`	✅	5s	5s	+0%
`mlk_matvec_mul`	✅	5s	2s	+150%
`mlk_poly_ntt`	✅	5s	8s	-38%
`mlk_polyvec_tomont`	✅	5s	2s	+150%
`poly_compress_d4_native_x86_64`	✅	5s	2s	+150%
`poly_decompress_d4_native_x86_64`	✅	5s	4s	+25%
`poly_reduce_native_x86_64`	✅	5s	2s	+150%
`mlk_check_pct`	✅	4s	2s	+100%
`mlk_ct_cmask_nonzero_u16`	✅	4s	2s	+100%
`mlk_ct_memcmp`	✅	4s	2s	+100%
`mlk_ct_sel_uint8`	✅	4s	1s	+300%
`mlk_invntt_layer`	✅	4s	6s	-33%
`mlk_poly_decompress_d4_c`	✅	4s	4s	+0%
`mlk_poly_frombytes_c`	✅	4s	5s	-20%
`mlk_poly_getnoise_eta2`	✅	4s	1s	+300%
`mlk_poly_invntt_tomont_c`	✅	4s	2s	+100%
`mlk_poly_mulcache_compute`	✅	4s	3s	+33%
`mlk_poly_tobytes_c`	✅	4s	2s	+100%
`mlk_poly_tomont_native`	✅	4s	4s	+0%
`mlk_polyvec_permute_bitrev_to_custom_native`	✅	4s	4s	+0%
`mlk_polyvec_tobytes`	✅	4s	2s	+100%
`mlk_scalar_compress_d10`	✅	4s	2s	+100%
`mlk_shake256x4`	✅	4s	2s	+100%
`poly_frombytes_native_x86_64`	✅	4s	3s	+33%
`polyvec_basemul_acc_montgomery_cached_k2_native_aarch64`	✅	4s	4s	+0%
`intt_native_aarch64`	✅	3s	2s	+50%
`keccak_f1600_x4_native_aarch64_v8a_v84a_scalar_hybrid`	✅	3s	2s	+50%
`keccakf1600_permute_native`	✅	3s	3s	+0%
`keccakf1600x4_extract_bytes_native`	✅	3s	2s	+50%
`keccakf1600x4_xor_bytes_native`	✅	3s	3s	+0%
`mlk_ct_cmask_neg_i16`	✅	3s	3s	+0%
`mlk_ct_sel_int16`	✅	3s	3s	+0%
`mlk_enc_getnoise_eta1_eta2`	✅	3s	4s	-25%
`mlk_gen_matrix`	✅	3s	3s	+0%
`mlk_keccakf1600_extract_bytes (big endian)`	✅	3s	1s	+200%
`mlk_keccakf1600_permute_c`	✅	3s	6s	-50%
`mlk_keccakf1600x4_extract_bytes_c`	✅	3s	3s	+0%
`mlk_keccakf1600x4_xor_bytes`	✅	3s	3s	+0%
`mlk_keypair_getnoise_eta1`	✅	3s	2s	+50%
`mlk_montgomery_reduce`	✅	3s	1s	+200%
`mlk_poly_cbd_eta1`	✅	3s	2s	+50%
`mlk_poly_compress_d10`	✅	3s	1s	+200%
`mlk_poly_compress_d11_native`	✅	3s	3s	+0%
`mlk_poly_compress_d4`	✅	3s	3s	+0%
`mlk_poly_compress_dv`	✅	3s	4s	-25%
`mlk_poly_decompress_d10_c`	✅	3s	3s	+0%
`mlk_poly_decompress_d5`	✅	3s	2s	+50%
`mlk_poly_decompress_du`	✅	3s	2s	+50%
`mlk_poly_getnoise_eta1_4x_native`	✅	3s	4s	-25%
`mlk_poly_ntt_c`	✅	3s	2s	+50%
`mlk_poly_reduce_c`	✅	3s	2s	+50%
`mlk_poly_tobytes`	✅	3s	1s	+200%
`mlk_poly_tomsg`	✅	3s	2s	+50%
`mlk_polymat_permute_bitrev_to_custom`	✅	3s	4s	-25%
`mlk_polyvec_decompress_du`	✅	3s	2s	+50%
`mlk_polyvec_mulcache_compute`	✅	3s	3s	+0%
`mlk_rej_uniform`	✅	3s	2s	+50%
`mlk_scalar_compress_d4`	✅	3s	2s	+50%
`mlk_shake128_absorb_once`	✅	3s	1s	+200%
`mlk_value_barrier_i32`	✅	3s	2s	+50%
`mlk_value_barrier_u32`	✅	3s	2s	+50%
`ntt_native_x86_64`	✅	3s	5s	-40%
`poly_compress_d11_native_x86_64`	✅	3s	2s	+50%
`poly_decompress_d11_native_x86_64`	✅	3s	2s	+50%
`poly_tobytes_native_x86_64`	✅	3s	3s	+0%
`poly_tomont_native_x86_64`	✅	3s	3s	+0%
`polyvec_basemul_acc_montgomery_cached_k2_native_x86_64`	✅	3s	3s	+0%
`polyvec_basemul_acc_montgomery_cached_k3_native_x86_64`	✅	3s	1s	+200%
`intt_native_x86_64`	✅	2s	2s	+0%
`keccak_f1600_x4_native_aarch64_v84a`	✅	2s	1s	+100%
`keccak_f1600_x4_native_aarch64_v8a_scalar_hybrid`	✅	2s	2s	+0%
`keccak_f1600_x4_native_avx2`	✅	2s	1s	+100%
`kem_check_pk`	✅	2s	4s	-50%
`kem_dec`	✅	2s	3s	-33%
`kem_enc`	✅	2s	4s	-50%
`kem_enc_derand`	✅	2s	4s	-50%
`kem_keypair`	✅	2s	4s	-50%
`kem_keypair_derand`	✅	2s	2s	+0%
`mlk_ct_cmov_zero`	✅	2s	2s	+0%
`mlk_ct_get_optblocker_u8`	✅	2s	2s	+0%
`mlk_keccakf1600_extract_bytes`	✅	2s	2s	+0%
`mlk_keccakf1600_xor_bytes (big endian)`	✅	2s	3s	-33%
`mlk_keccakf1600x4_extract_bytes`	✅	2s	3s	-33%
`mlk_keccakf1600x4_permute`	✅	2s	1s	+100%
`mlk_keccakf1600x4_xor_bytes_c`	✅	2s	2s	+0%
`mlk_poly_compress_d10_native`	✅	2s	2s	+0%
`mlk_poly_compress_d4_c`	✅	2s	4s	-50%
`mlk_poly_compress_d5`	✅	2s	2s	+0%
`mlk_poly_compress_d5_native`	✅	2s	2s	+0%
`mlk_poly_decompress_d11`	✅	2s	1s	+100%
`mlk_poly_decompress_d11_c`	✅	2s	3s	-33%
`mlk_poly_decompress_d4`	✅	2s	2s	+0%
`mlk_poly_decompress_d5_c`	✅	2s	2s	+0%
`mlk_poly_frombytes`	✅	2s	3s	-33%
`mlk_poly_getnoise_eta1122_4x`	✅	2s	2s	+0%
`mlk_poly_getnoise_eta1_4x`	✅	2s	3s	-33%
`mlk_poly_invntt_tomont`	✅	2s	2s	+0%
`mlk_poly_mulcache_compute_native`	✅	2s	3s	-33%
`mlk_poly_tobytes_native`	✅	2s	3s	-33%
`mlk_poly_tomont_c`	✅	2s	1s	+100%
`mlk_polyvec_basemul_acc_montgomery_cached`	✅	2s	2s	+0%
`mlk_polyvec_compress_du`	✅	2s	3s	-33%
`mlk_polyvec_frombytes`	✅	2s	6s	-67%
`mlk_polyvec_invntt_tomont`	✅	2s	2s	+0%
`mlk_polyvec_permute_bitrev_to_custom`	✅	2s	1s	+100%
`mlk_polyvec_reduce`	✅	2s	4s	-50%
`mlk_scalar_compress_d1`	✅	2s	4s	-50%
`mlk_scalar_compress_d11`	✅	2s	2s	+0%
`mlk_scalar_decompress_d4`	✅	2s	2s	+0%
`mlk_scalar_signed_to_unsigned_q`	✅	2s	1s	+100%
`mlk_sha3_256`	✅	2s	3s	-33%
`mlk_sha3_512`	✅	2s	5s	-60%
`mlk_shake128_squeezeblocks`	✅	2s	2s	+0%
`mlk_shake128x4_absorb_once`	✅	2s	3s	-33%
`mlk_shake128x4_squeezeblocks`	✅	2s	1s	+100%
`mlk_value_barrier_u8`	✅	2s	2s	+0%
`nttunpack_native_x86_64`	✅	2s	1s	+100%
`poly_compress_d10_native_x86_64`	✅	2s	4s	-50%
`poly_compress_d5_native_x86_64`	✅	2s	2s	+0%
`poly_decompress_d5_native_x86_64`	✅	2s	1s	+100%
`poly_getnoise_eta1122_4x_native`	✅	2s	3s	-33%
`poly_invntt_tomont_native`	✅	2s	2s	+0%
`poly_mulcache_compute_native_aarch64`	✅	2s	3s	-33%
`poly_mulcache_compute_native_x86_64`	✅	2s	2s	+0%
`poly_reduce_native_aarch64`	✅	2s	4s	-50%
`poly_tobytes_native_aarch64`	✅	2s	3s	-33%
`polyvec_basemul_acc_montgomery_cached_k3_native_aarch64`	✅	2s	3s	-33%
`polyvec_basemul_acc_montgomery_cached_k4_native_aarch64`	✅	2s	1s	+100%
`polyvec_basemul_acc_montgomery_cached_k4_native_x86_64`	✅	2s	1s	+100%
`sys_check_capability`	✅	2s	3s	-33%
`keccak_f1600_x1_native_aarch64`	✅	1s	2s	-50%
`keccak_f1600_x1_native_aarch64_v84a`	✅	1s	3s	-67%
`mlk_barrett_reduce`	✅	1s	1s	+0%
`mlk_ct_cmask_nonzero_u8`	✅	1s	2s	-50%
`mlk_ct_get_optblocker_i32`	✅	1s	4s	-75%
`mlk_ct_get_optblocker_u32`	✅	1s	1s	+0%
`mlk_gen_matrix_serial`	✅	1s	2s	-50%
`mlk_keccakf1600_permute`	✅	1s	2s	-50%
`mlk_keccakf1600_xor_bytes`	✅	1s	4s	-75%
`mlk_poly_cbd_eta2`	✅	1s	3s	-67%
`mlk_poly_compress_d10_c`	✅	1s	3s	-67%
`mlk_poly_compress_d11`	✅	1s	2s	-50%
`mlk_poly_compress_d11_c`	✅	1s	1s	+0%
`mlk_poly_compress_d4_native`	✅	1s	2s	-50%
`mlk_poly_compress_d5_c`	✅	1s	3s	-67%
`mlk_poly_compress_du`	✅	1s	1s	+0%
`mlk_poly_decompress_d10`	✅	1s	2s	-50%
`mlk_poly_decompress_d11_native`	✅	1s	2s	-50%
`mlk_poly_decompress_d5_native`	✅	1s	2s	-50%
`mlk_poly_decompress_dv`	✅	1s	2s	-50%
`mlk_poly_mulcache_compute_c`	✅	1s	2s	-50%
`mlk_poly_reduce`	✅	1s	1s	+0%
`mlk_poly_tomont`	✅	1s	1s	+0%
`mlk_polyvec_ntt`	✅	1s	2s	-50%
`mlk_scalar_compress_d5`	✅	1s	2s	-50%
`mlk_scalar_decompress_d10`	✅	1s	2s	-50%
`mlk_scalar_decompress_d11`	✅	1s	2s	-50%
`mlk_scalar_decompress_d5`	✅	1s	3s	-67%
`mlk_shake256`	✅	1s	1s	+0%
`ntt_native_aarch64`	✅	1s	3s	-67%
`poly_tomont_native_aarch64`	✅	1s	2s	-50%
`rej_uniform_native`	✅	1s	2s	-50%

oqs-bot · 2026-07-09T07:37:00Z

CBMC Results (ML-KEM-1024)

Full Results (191 proofs)

Proof	Status	Current	Previous	Change
`TOTAL`	✅	1809s	1648s	+9.8%
`mlk_indcpa_enc`	✅	385s	350s	+10%
`mlk_indcpa_keypair_derand`	✅	172s	166s	+4%
`mlk_poly_rej_uniform`	✅	160s	129s	+24%
`mlk_rej_uniform_c`	✅	142s	123s	+15%
`mlk_polyvec_basemul_acc_montgomery_cached_c`	✅	77s	73s	+5%
`rej_uniform_native_x86_64`	✅	61s	56s	+9%
`poly_ntt_native`	✅	45s	36s	+25%
`mlk_poly_reduce_native`	✅	39s	33s	+18%
`rej_uniform_native_aarch64`	✅	39s	34s	+15%
`mlk_ntt_layer`	✅	37s	28s	+32%
`mlk_keccak_squeezeblocks_x4`	✅	27s	25s	+8%
`polyvec_basemul_acc_montgomery_cached_native`	✅	27s	26s	+4%
`mlk_polyvec_add`	✅	19s	17s	+12%
`keccakf1600x4_permute_native_x4`	✅	17s	18s	-6%
`mlk_fqmul`	✅	17s	17s	+0%
`mlk_poly_decompress_d11_native`	✅	17s	14s	+21%
`mlk_poly_decompress_d5_native`	✅	16s	14s	+14%
`mlk_indcpa_dec`	✅	12s	9s	+33%
`mlk_poly_frommsg`	✅	12s	10s	+20%
`mlk_polyvec_ntt`	✅	12s	10s	+20%
`mlk_keccak_squeezeblocks`	✅	10s	8s	+25%
`mlk_poly_frombytes_native`	✅	9s	8s	+12%
`mlk_poly_ntt`	✅	9s	5s	+80%
`mlk_keccak_squeeze_once`	✅	8s	8s	+0%
`mlk_ntt_butterfly_block`	✅	8s	8s	+0%
`mlk_poly_sub`	✅	8s	8s	+0%
`mlk_invntt_layer`	✅	7s	3s	+133%
`mlk_poly_rej_uniform_x4`	✅	7s	6s	+17%
`mlk_keccak_absorb_once_x4`	✅	6s	7s	-14%
`mlk_poly_tomont_c`	✅	6s	2s	+200%
`mlk_polymat_permute_bitrev_to_custom`	✅	6s	3s	+100%
`kem_check_pk`	✅	5s	3s	+67%
`mlk_gen_matrix_serial`	✅	5s	3s	+67%
`mlk_keccakf1600_permute_c`	✅	5s	6s	-17%
`mlk_keccakf1600x4_xor_bytes_c`	✅	5s	2s	+150%
`mlk_poly_add`	✅	5s	5s	+0%
`mlk_poly_frombytes_c`	✅	5s	5s	+0%
`mlk_polyvec_tobytes`	✅	5s	2s	+150%
`mlk_sha3_256`	✅	5s	2s	+150%
`poly_frombytes_native_x86_64`	✅	5s	3s	+67%
`keccak_f1600_x4_native_aarch64_v8a_scalar_hybrid`	✅	4s	1s	+300%
`keccakf1600x4_extract_bytes_native`	✅	4s	3s	+33%
`keccakf1600x4_xor_bytes_native`	✅	4s	2s	+100%
`mlk_gen_matrix`	✅	4s	4s	+0%
`mlk_poly_cbd_eta1`	✅	4s	1s	+300%
`mlk_poly_compress_d11_c`	✅	4s	3s	+33%
`mlk_poly_compress_d4_native`	✅	4s	1s	+300%
`mlk_poly_invntt_tomont_c`	✅	4s	3s	+33%
`mlk_poly_mulcache_compute_c`	✅	4s	4s	+0%
`mlk_polyvec_frombytes`	✅	4s	2s	+100%
`mlk_polyvec_invntt_tomont`	✅	4s	2s	+100%
`mlk_shake128_absorb_once`	✅	4s	3s	+33%
`poly_decompress_d11_native_x86_64`	✅	4s	5s	-20%
`poly_mulcache_compute_native_x86_64`	✅	4s	4s	+0%
`poly_reduce_native_aarch64`	✅	4s	1s	+300%
`polyvec_basemul_acc_montgomery_cached_k2_native_x86_64`	✅	4s	3s	+33%
`intt_native_aarch64`	✅	3s	2s	+50%
`intt_native_x86_64`	✅	3s	2s	+50%
`keccak_f1600_x1_native_aarch64_v84a`	✅	3s	2s	+50%
`keccak_f1600_x4_native_aarch64_v84a`	✅	3s	2s	+50%
`kem_dec`	✅	3s	3s	+0%
`kem_keypair`	✅	3s	2s	+50%
`mlk_ct_cmask_neg_i16`	✅	3s	4s	-25%
`mlk_ct_cmask_nonzero_u16`	✅	3s	2s	+50%
`mlk_ct_cmov_zero`	✅	3s	2s	+50%
`mlk_ct_memcmp`	✅	3s	3s	+0%
`mlk_keccak_absorb_once`	✅	3s	4s	-25%
`mlk_keccakf1600_extract_bytes`	✅	3s	1s	+200%
`mlk_keccakf1600_extract_bytes (big endian)`	✅	3s	2s	+50%
`mlk_keccakf1600x4_permute`	✅	3s	1s	+200%
`mlk_matvec_mul`	✅	3s	4s	-25%
`mlk_montgomery_reduce`	✅	3s	3s	+0%
`mlk_poly_compress_d10_c`	✅	3s	3s	+0%
`mlk_poly_compress_d10_native`	✅	3s	2s	+50%
`mlk_poly_compress_d4_c`	✅	3s	4s	-25%
`mlk_poly_compress_dv`	✅	3s	2s	+50%
`mlk_poly_decompress_d11`	✅	3s	3s	+0%
`mlk_poly_frombytes`	✅	3s	3s	+0%
`mlk_poly_getnoise_eta1_4x`	✅	3s	3s	+0%
`mlk_poly_getnoise_eta1_4x_native`	✅	3s	4s	-25%
`mlk_poly_mulcache_compute`	✅	3s	3s	+0%
`mlk_poly_reduce_c`	✅	3s	1s	+200%
`mlk_polyvec_mulcache_compute`	✅	3s	2s	+50%
`mlk_polyvec_permute_bitrev_to_custom_native`	✅	3s	1s	+200%
`mlk_rej_uniform`	✅	3s	1s	+200%
`mlk_scalar_compress_d11`	✅	3s	2s	+50%
`mlk_scalar_decompress_d4`	✅	3s	2s	+50%
`mlk_scalar_decompress_d5`	✅	3s	2s	+50%
`mlk_scalar_signed_to_unsigned_q`	✅	3s	1s	+200%
`mlk_shake128x4_squeezeblocks`	✅	3s	3s	+0%
`mlk_shake256x4`	✅	3s	3s	+0%
`ntt_native_x86_64`	✅	3s	2s	+50%
`poly_compress_d10_native_x86_64`	✅	3s	1s	+200%
`poly_compress_d4_native_x86_64`	✅	3s	3s	+0%
`poly_decompress_d5_native_x86_64`	✅	3s	4s	-25%
`poly_getnoise_eta1122_4x_native`	✅	3s	1s	+200%
`poly_invntt_tomont_native`	✅	3s	1s	+200%
`poly_mulcache_compute_native_aarch64`	✅	3s	4s	-25%
`polyvec_basemul_acc_montgomery_cached_k3_native_aarch64`	✅	3s	4s	-25%
`polyvec_basemul_acc_montgomery_cached_k4_native_x86_64`	✅	3s	3s	+0%
`keccak_f1600_x1_native_aarch64`	✅	2s	1s	+100%
`keccak_f1600_x4_native_avx2`	✅	2s	1s	+100%
`kem_check_sk`	✅	2s	3s	-33%
`kem_enc_derand`	✅	2s	3s	-33%
`kem_keypair_derand`	✅	2s	2s	+0%
`mlk_check_pct`	✅	2s	5s	-60%
`mlk_ct_cmask_nonzero_u8`	✅	2s	2s	+0%
`mlk_ct_get_optblocker_i32`	✅	2s	2s	+0%
`mlk_ct_get_optblocker_u32`	✅	2s	4s	-50%
`mlk_ct_sel_uint8`	✅	2s	2s	+0%
`mlk_enc_getnoise_eta1_eta2`	✅	2s	5s	-60%
`mlk_keccakf1600_permute`	✅	2s	2s	+0%
`mlk_keccakf1600_xor_bytes`	✅	2s	4s	-50%
`mlk_keccakf1600_xor_bytes (big endian)`	✅	2s	2s	+0%
`mlk_keccakf1600x4_extract_bytes_c`	✅	2s	2s	+0%
`mlk_keccakf1600x4_xor_bytes`	✅	2s	4s	-50%
`mlk_keypair_getnoise_eta1`	✅	2s	1s	+100%
`mlk_poly_compress_d11`	✅	2s	2s	+0%
`mlk_poly_compress_d11_native`	✅	2s	3s	-33%
`mlk_poly_compress_d4`	✅	2s	3s	-33%
`mlk_poly_compress_d5_c`	✅	2s	2s	+0%
`mlk_poly_compress_d5_native`	✅	2s	3s	-33%
`mlk_poly_decompress_d10`	✅	2s	2s	+0%
`mlk_poly_decompress_d4_c`	✅	2s	4s	-50%
`mlk_poly_decompress_d4_native`	✅	2s	2s	+0%
`mlk_poly_decompress_d5`	✅	2s	1s	+100%
`mlk_poly_decompress_d5_c`	✅	2s	1s	+100%
`mlk_poly_decompress_du`	✅	2s	2s	+0%
`mlk_poly_decompress_dv`	✅	2s	2s	+0%
`mlk_poly_getnoise_eta1122_4x`	✅	2s	3s	-33%
`mlk_poly_getnoise_eta2`	✅	2s	3s	-33%
`mlk_poly_mulcache_compute_native`	✅	2s	2s	+0%
`mlk_poly_tobytes`	✅	2s	2s	+0%
`mlk_poly_tobytes_native`	✅	2s	2s	+0%
`mlk_poly_tomont`	✅	2s	1s	+100%
`mlk_poly_tomont_native`	✅	2s	2s	+0%
`mlk_poly_tomsg`	✅	2s	4s	-50%
`mlk_polyvec_compress_du`	✅	2s	1s	+100%
`mlk_polyvec_decompress_du`	✅	2s	2s	+0%
`mlk_polyvec_permute_bitrev_to_custom`	✅	2s	5s	-60%
`mlk_scalar_compress_d1`	✅	2s	2s	+0%
`mlk_scalar_compress_d4`	✅	2s	2s	+0%
`mlk_scalar_compress_d5`	✅	2s	2s	+0%
`mlk_scalar_decompress_d10`	✅	2s	1s	+100%
`mlk_scalar_decompress_d11`	✅	2s	2s	+0%
`mlk_shake128_squeezeblocks`	✅	2s	1s	+100%
`mlk_shake128x4_absorb_once`	✅	2s	2s	+0%
`mlk_value_barrier_i32`	✅	2s	2s	+0%
`mlk_value_barrier_u32`	✅	2s	2s	+0%
`mlk_value_barrier_u8`	✅	2s	3s	-33%
`ntt_native_aarch64`	✅	2s	1s	+100%
`nttunpack_native_x86_64`	✅	2s	2s	+0%
`poly_compress_d5_native_x86_64`	✅	2s	1s	+100%
`poly_reduce_native_x86_64`	✅	2s	5s	-60%
`poly_tobytes_native_x86_64`	✅	2s	1s	+100%
`poly_tomont_native_aarch64`	✅	2s	3s	-33%
`poly_tomont_native_x86_64`	✅	2s	3s	-33%
`polyvec_basemul_acc_montgomery_cached_k2_native_aarch64`	✅	2s	3s	-33%
`polyvec_basemul_acc_montgomery_cached_k4_native_aarch64`	✅	2s	5s	-60%
`sys_check_capability`	✅	2s	1s	+100%
`keccak_f1600_x4_native_aarch64_v8a_v84a_scalar_hybrid`	✅	1s	1s	+0%
`keccakf1600_permute_native`	✅	1s	3s	-67%
`kem_enc`	✅	1s	1s	+0%
`mlk_barrett_reduce`	✅	1s	2s	-50%
`mlk_ct_get_optblocker_u8`	✅	1s	1s	+0%
`mlk_ct_sel_int16`	✅	1s	3s	-67%
`mlk_keccakf1600x4_extract_bytes`	✅	1s	2s	-50%
`mlk_poly_cbd_eta2`	✅	1s	4s	-75%
`mlk_poly_compress_d10`	✅	1s	3s	-67%
`mlk_poly_compress_d5`	✅	1s	2s	-50%
`mlk_poly_compress_du`	✅	1s	3s	-67%
`mlk_poly_decompress_d10_c`	✅	1s	3s	-67%
`mlk_poly_decompress_d10_native`	✅	1s	2s	-50%
`mlk_poly_decompress_d11_c`	✅	1s	1s	+0%
`mlk_poly_decompress_d4`	✅	1s	2s	-50%
`mlk_poly_invntt_tomont`	✅	1s	2s	-50%
`mlk_poly_ntt_c`	✅	1s	2s	-50%
`mlk_poly_reduce`	✅	1s	2s	-50%
`mlk_poly_tobytes_c`	✅	1s	3s	-67%
`mlk_polyvec_basemul_acc_montgomery_cached`	✅	1s	2s	-50%
`mlk_polyvec_reduce`	✅	1s	2s	-50%
`mlk_polyvec_tomont`	✅	1s	2s	-50%
`mlk_scalar_compress_d10`	✅	1s	1s	+0%
`mlk_sha3_512`	✅	1s	1s	+0%
`mlk_shake256`	✅	1s	4s	-75%
`poly_compress_d11_native_x86_64`	✅	1s	3s	-67%
`poly_decompress_d10_native_x86_64`	✅	1s	5s	-80%
`poly_decompress_d4_native_x86_64`	✅	1s	1s	+0%
`poly_tobytes_native_aarch64`	✅	1s	3s	-67%
`polyvec_basemul_acc_montgomery_cached_k3_native_x86_64`	✅	1s	1s	+0%
`rej_uniform_native`	✅	1s	4s	-75%

oqs-bot

Arm Cortex-M55 (NUCLEO-N657X0-Q) benchmarks

Details

Benchmark suite	Current: `0e34d09`	Previous: `cfd1912`	Ratio
`ML-KEM-512 keypair`	`727950` cycles	`7850380` cycles	`0.09272799533271`
`ML-KEM-512 encaps`	`856828` cycles	`8587300` cycles	`0.09977851012541777`
`ML-KEM-512 decaps`	`1090494` cycles	`10952844` cycles	`0.09956263414324171`
`ML-KEM-768 keypair`	`1196905` cycles	`12711760` cycles	`0.09415730001195743`
`ML-KEM-768 encaps`	`1390354` cycles	`14138090` cycles	`0.09834100645843957`
`ML-KEM-768 decaps`	`1705536` cycles	`17499057` cycles	`0.09746445194161034`
`ML-KEM-1024 keypair`	`1766592` cycles	`23228659` cycles	`0.07605225940937874`
`ML-KEM-1024 encaps`	`1975479` cycles	`25701892` cycles	`0.07686122873755753`
`ML-KEM-1024 decaps`	`2370613` cycles	`30573558` cycles	`0.07753801503900855`

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

oqs-bot

Arm Cortex-A55 (Snapdragon 888) benchmarks

Details

Benchmark suite	Current: `0e34d09`	Previous: `cfd1912`	Ratio
`ML-KEM-512 keypair`	`59810` cycles	`59849` cycles	`1.00`
`ML-KEM-512 encaps`	`67491` cycles	`67486` cycles	`1.00`
`ML-KEM-512 decaps`	`86491` cycles	`86496` cycles	`1.00`
`ML-KEM-768 keypair`	`97501` cycles	`97575` cycles	`1.00`
`ML-KEM-768 encaps`	`110987` cycles	`111254` cycles	`1.00`
`ML-KEM-768 decaps`	`138794` cycles	`138464` cycles	`1.00`
`ML-KEM-1024 keypair`	`154817` cycles	`154791` cycles	`1.00`
`ML-KEM-1024 encaps`	`172378` cycles	`172787` cycles	`1.00`
`ML-KEM-1024 decaps`	`208461` cycles	`208533` 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: `0e34d09`	Previous: `cfd1912`	Ratio
`ML-KEM-512 keypair`	`51926` cycles	`52898` cycles	`0.98`
`ML-KEM-512 encaps`	`60239` cycles	`60920` cycles	`0.99`
`ML-KEM-512 decaps`	`76346` cycles	`78181` cycles	`0.98`
`ML-KEM-768 keypair`	`89783` cycles	`88876` cycles	`1.01`
`ML-KEM-768 encaps`	`98665` cycles	`97634` cycles	`1.01`
`ML-KEM-768 decaps`	`122699` cycles	`121328` cycles	`1.01`
`ML-KEM-1024 keypair`	`133541` cycles	`133260` cycles	`1.00`
`ML-KEM-1024 encaps`	`145458` cycles	`146057` cycles	`1.00`
`ML-KEM-1024 decaps`	`178502` cycles	`178659` cycles	`1.00`

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

hanno-becker · 2026-07-10T08:52:41Z

-  return (int)((*((const uint64_t *)a)) - (*((const uint64_t *)b)));
+  const uint64_t va = *((const uint64_t *)a);
+  const uint64_t vb = *((const uint64_t *)b);
+
+  return (va > vb) - (va < vb);


Why was this change made? Some comment would be useful.

hanno-becker · 2026-07-10T08:54:01Z

+        zephyr,memory-region = "DTCM";
+    };
+
    mlk_axisram: memory@34080000 {


Is this still needed?

hanno-becker

Can you split the cycle counter changes and the memory layout changes in two separate commits?

bremoran requested a review from a team as a code owner July 9, 2026 06:56

mkannwischer added the benchmark this PR should be benchmarked in CI label Jul 9, 2026

oqs-bot reviewed Jul 9, 2026

View reviewed changes

hanno-becker reviewed Jul 10, 2026

View reviewed changes

Uh oh!

Conversation

bremoran commented Jul 9, 2026

Uh oh!

oqs-bot left a comment

Choose a reason for hiding this comment

Mac Mini (M1, 2020) benchmarks

Uh oh!

oqs-bot left a comment

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

Choose a reason for hiding this comment

Arm Cortex-A76 (Raspberry Pi 5) benchmarks

Uh oh!

oqs-bot left a comment

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

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

Choose a reason for hiding this comment

AMD EPYC 4th gen (c7a)

Uh oh!

oqs-bot left a comment

Choose a reason for hiding this comment

Graviton5

Uh oh!

oqs-bot left a comment

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

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

Uh oh!

oqs-bot left a comment

Choose a reason for hiding this comment

Graviton3

Uh oh!

oqs-bot left a comment

Choose a reason for hiding this comment

Graviton4

Uh oh!

oqs-bot left a comment

Choose a reason for hiding this comment

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

Uh oh!

oqs-bot left a comment

Choose a reason for hiding this comment

Graviton5 (no-opt)

Uh oh!

oqs-bot left a comment

Choose a reason for hiding this comment

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

Uh oh!

oqs-bot left a comment

Choose a reason for hiding this comment

Graviton4 (no-opt)

Uh oh!

oqs-bot left a comment

Choose a reason for hiding this comment

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

Uh oh!

oqs-bot left a comment

Choose a reason for hiding this comment

Graviton2

Uh oh!