Define `mlk_polyvec[_mulcache]` as typedef for array #895

hanno-becker · 2025-03-20T05:08:28Z

Resolves Define mlkem_polyvec as array rather than struct wrapper #893

This commit changes the definition of mlk_polyvec from a struct
wrapper struct { mlk_poly vec[MLKEM_K] } to a typedef for
mlk_poly[MLKEM_K]. Similarly, it introduces mlk_polymat
as a typedef for mlk_poly[MLKEM_K * MLKEM_K].

This ensures that when working with polynomial matrices,
the polynomial entries are consecutive in memory.
This needs to be assumed for the correctness of gen_matrix().

From all we know, is a theoretical concern that without this
change, compilers would be free to introduce padding into
mlk_polyvec, breaking gen_matrix(). Yet, the behaviour of
gen_matrix is strictly speaking UB, so even if a compiler
does not use padding in mlk_polyvec, it is better to change it.

Defining mlk_polymat as a typedef for mlk_poly[MLKEM_K * MLKEM_K]
and not mlk_polyvec[MLKEM_K] is deliberate: In the latter case, one
would not be able to pass a mlk_polymat to a function expecting
const polymat because of how qualifiers are (not) propagated,
forcing cumbersome manual casts.

oqs-bot

Intel Xeon 4th gen (c7i)

Benchmark suite	Current: `60e019b`	Previous: `8dba13b`	Ratio
`ML-KEM-512 keypair`	`9511` cycles	`9477` cycles	`1.00`
`ML-KEM-512 encaps`	`11045` cycles	`11217` cycles	`0.98`
`ML-KEM-512 decaps`	`15295` cycles	`15166` cycles	`1.01`
`ML-KEM-768 keypair`	`16300` cycles	`16303` cycles	`1.00`
`ML-KEM-768 encaps`	`17698` cycles	`17790` cycles	`0.99`
`ML-KEM-768 decaps`	`23325` cycles	`23444` cycles	`0.99`
`ML-KEM-1024 keypair`	`21974` cycles	`22001` cycles	`1.00`
`ML-KEM-1024 encaps`	`24062` cycles	`24012` cycles	`1.00`
`ML-KEM-1024 decaps`	`31755` cycles	`31587` cycles	`1.01`

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

oqs-bot

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

Benchmark suite	Current: `60e019b`	Previous: `8dba13b`	Ratio
`ML-KEM-512 keypair`	`28969` cycles	`30164` cycles	`0.96`
`ML-KEM-512 encaps`	`34849` cycles	`35677` cycles	`0.98`
`ML-KEM-512 decaps`	`44489` cycles	`45084` cycles	`0.99`
`ML-KEM-768 keypair`	`47644` cycles	`48448` cycles	`0.98`
`ML-KEM-768 encaps`	`55592` cycles	`55690` cycles	`1.00`
`ML-KEM-768 decaps`	`67719` cycles	`68876` cycles	`0.98`
`ML-KEM-1024 keypair`	`73246` cycles	`74942` cycles	`0.98`
`ML-KEM-1024 encaps`	`85261` cycles	`83410` cycles	`1.02`
`ML-KEM-1024 decaps`	`99188` cycles	`99614` cycles	`1.00`

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

oqs-bot

AMD EPYC 4th gen (c7a)

Benchmark suite	Current: `60e019b`	Previous: `8dba13b`	Ratio
`ML-KEM-512 keypair`	`11734` cycles	`11702` cycles	`1.00`
`ML-KEM-512 encaps`	`13449` cycles	`13439` cycles	`1.00`
`ML-KEM-512 decaps`	`18310` cycles	`18310` cycles	`1`
`ML-KEM-768 keypair`	`20213` cycles	`20125` cycles	`1.00`
`ML-KEM-768 encaps`	`21188` cycles	`21112` cycles	`1.00`
`ML-KEM-768 decaps`	`28361` cycles	`28224` cycles	`1.00`
`ML-KEM-1024 keypair`	`27021` cycles	`27067` cycles	`1.00`
`ML-KEM-1024 encaps`	`29171` cycles	`29123` cycles	`1.00`
`ML-KEM-1024 decaps`	`38640` cycles	`38639` cycles	`1.00`

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

oqs-bot

AMD EPYC 3rd gen (c6a)

Benchmark suite	Current: `60e019b`	Previous: `8dba13b`	Ratio
`ML-KEM-512 keypair`	`17199` cycles	`17264` cycles	`1.00`
`ML-KEM-512 encaps`	`18986` cycles	`19057` cycles	`1.00`
`ML-KEM-512 decaps`	`24460` cycles	`24513` cycles	`1.00`
`ML-KEM-768 keypair`	`29571` cycles	`29552` cycles	`1.00`
`ML-KEM-768 encaps`	`30807` cycles	`30875` cycles	`1.00`
`ML-KEM-768 decaps`	`38658` cycles	`38640` cycles	`1.00`
`ML-KEM-1024 keypair`	`42933` cycles	`42962` cycles	`1.00`
`ML-KEM-1024 encaps`	`45574` cycles	`45604` cycles	`1.00`
`ML-KEM-1024 decaps`	`55945` cycles	`55827` cycles	`1.00`

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

oqs-bot

Intel Xeon 3rd gen (c6i)

Benchmark suite	Current: `60e019b`	Previous: `8dba13b`	Ratio
`ML-KEM-512 keypair`	`16160` cycles	`16149` cycles	`1.00`
`ML-KEM-512 encaps`	`18378` cycles	`18247` cycles	`1.01`
`ML-KEM-512 decaps`	`24861` cycles	`24723` cycles	`1.01`
`ML-KEM-768 keypair`	`27847` cycles	`27803` cycles	`1.00`
`ML-KEM-768 encaps`	`29470` cycles	`29387` cycles	`1.00`
`ML-KEM-768 decaps`	`38888` cycles	`38836` cycles	`1.00`
`ML-KEM-1024 keypair`	`37635` cycles	`37578` cycles	`1.00`
`ML-KEM-1024 encaps`	`40616` cycles	`40554` cycles	`1.00`
`ML-KEM-1024 decaps`	`53209` cycles	`53168` cycles	`1.00`

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

oqs-bot

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

Benchmark suite	Current: `60e019b`	Previous: `8dba13b`	Ratio
`ML-KEM-512 keypair`	`35983` cycles	`36358` cycles	`0.99`
`ML-KEM-512 encaps`	`42702` cycles	`42959` cycles	`0.99`
`ML-KEM-512 decaps`	`55624` cycles	`55963` cycles	`0.99`
`ML-KEM-768 keypair`	`59251` cycles	`59249` cycles	`1.00`
`ML-KEM-768 encaps`	`67882` cycles	`67706` cycles	`1.00`
`ML-KEM-768 decaps`	`84973` cycles	`84932` cycles	`1.00`
`ML-KEM-1024 keypair`	`87551` cycles	`87560` cycles	`1.00`
`ML-KEM-1024 encaps`	`99961` cycles	`98317` cycles	`1.02`
`ML-KEM-1024 decaps`	`121161` cycles	`120104` cycles	`1.01`

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

oqs-bot

Arm Cortex-A55 (Snapdragon 888) benchmarks

Benchmark suite	Current: `60e019b`	Previous: `8dba13b`	Ratio
`ML-KEM-512 keypair`	`59376` cycles	`59330` cycles	`1.00`
`ML-KEM-512 encaps`	`66964` cycles	`66920` cycles	`1.00`
`ML-KEM-512 decaps`	`85989` cycles	`86014` cycles	`1.00`
`ML-KEM-768 keypair`	`101050` cycles	`101087` cycles	`1.00`
`ML-KEM-768 encaps`	`112066` cycles	`112317` cycles	`1.00`
`ML-KEM-768 decaps`	`139354` cycles	`139358` cycles	`1.00`
`ML-KEM-1024 keypair`	`153416` cycles	`153471` cycles	`1.00`
`ML-KEM-1024 encaps`	`173222` cycles	`170285` cycles	`1.02`
`ML-KEM-1024 decaps`	`207524` cycles	`207958` cycles	`1.00`

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

oqs-bot

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

Benchmark suite	Current: `60e019b`	Previous: `8dba13b`	Ratio
`ML-KEM-512 keypair`	`39028` cycles	`39026` cycles	`1.00`
`ML-KEM-512 encaps`	`47207` cycles	`47097` cycles	`1.00`
`ML-KEM-512 decaps`	`60983` cycles	`60745` cycles	`1.00`
`ML-KEM-768 keypair`	`62786` cycles	`63315` cycles	`0.99`
`ML-KEM-768 encaps`	`73731` cycles	`73946` cycles	`1.00`
`ML-KEM-768 decaps`	`91275` cycles	`91861` cycles	`0.99`
`ML-KEM-1024 keypair`	`94638` cycles	`94031` cycles	`1.01`
`ML-KEM-1024 encaps`	`108342` cycles	`107041` cycles	`1.01`
`ML-KEM-1024 decaps`	`131630` cycles	`129724` cycles	`1.01`

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

oqs-bot

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

Benchmark suite	Current: `60e019b`	Previous: `8dba13b`	Ratio
`ML-KEM-512 keypair`	`46352` cycles	`47466` cycles	`0.98`
`ML-KEM-512 encaps`	`54720` cycles	`55344` cycles	`0.99`
`ML-KEM-512 decaps`	`70044` cycles	`71556` cycles	`0.98`
`ML-KEM-768 keypair`	`75161` cycles	`77538` cycles	`0.97`
`ML-KEM-768 encaps`	`86346` cycles	`87559` cycles	`0.99`
`ML-KEM-768 decaps`	`106296` cycles	`108276` cycles	`0.98`
`ML-KEM-1024 keypair`	`111110` cycles	`113314` cycles	`0.98`
`ML-KEM-1024 encaps`	`125180` cycles	`126306` cycles	`0.99`
`ML-KEM-1024 decaps`	`150981` cycles	`152929` cycles	`0.99`

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

oqs-bot

Graviton4

Benchmark suite	Current: `60e019b`	Previous: `8dba13b`	Ratio
`ML-KEM-512 keypair`	`17795` cycles	`17771` cycles	`1.00`
`ML-KEM-512 encaps`	`21020` cycles	`21026` cycles	`1.00`
`ML-KEM-512 decaps`	`27681` cycles	`27696` cycles	`1.00`
`ML-KEM-768 keypair`	`30634` cycles	`30710` cycles	`1.00`
`ML-KEM-768 encaps`	`33565` cycles	`33580` cycles	`1.00`
`ML-KEM-768 decaps`	`43128` cycles	`43195` cycles	`1.00`
`ML-KEM-1024 keypair`	`44316` cycles	`44315` cycles	`1.00`
`ML-KEM-1024 encaps`	`49592` cycles	`49591` cycles	`1.00`
`ML-KEM-1024 decaps`	`62582` cycles	`62497` cycles	`1.00`

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

oqs-bot

Graviton2

Benchmark suite	Current: `60e019b`	Previous: `8dba13b`	Ratio
`ML-KEM-512 keypair`	`29402` cycles	`29444` cycles	`1.00`
`ML-KEM-512 encaps`	`34606` cycles	`34680` cycles	`1.00`
`ML-KEM-512 decaps`	`45218` cycles	`45279` cycles	`1.00`
`ML-KEM-768 keypair`	`50045` cycles	`50139` cycles	`1.00`
`ML-KEM-768 encaps`	`55297` cycles	`55224` cycles	`1.00`
`ML-KEM-768 decaps`	`70206` cycles	`70096` cycles	`1.00`
`ML-KEM-1024 keypair`	`73102` cycles	`73098` cycles	`1.00`
`ML-KEM-1024 encaps`	`81496` cycles	`81514` cycles	`1.00`
`ML-KEM-1024 decaps`	`101335` cycles	`101471` cycles	`1.00`

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

oqs-bot

Graviton3

Benchmark suite	Current: `60e019b`	Previous: `8dba13b`	Ratio
`ML-KEM-512 keypair`	`18875` cycles	`18877` cycles	`1.00`
`ML-KEM-512 encaps`	`22363` cycles	`22358` cycles	`1.00`
`ML-KEM-512 decaps`	`29607` cycles	`29566` cycles	`1.00`
`ML-KEM-768 keypair`	`32301` cycles	`32250` cycles	`1.00`
`ML-KEM-768 encaps`	`35744` cycles	`35598` cycles	`1.00`
`ML-KEM-768 decaps`	`46049` cycles	`45965` cycles	`1.00`
`ML-KEM-1024 keypair`	`46410` cycles	`46487` cycles	`1.00`
`ML-KEM-1024 encaps`	`52060` cycles	`52080` cycles	`1.00`
`ML-KEM-1024 decaps`	`65864` cycles	`65962` cycles	`1.00`

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

oqs-bot

Graviton4 (no-opt)

Benchmark suite	Current: `60e019b`	Previous: `8dba13b`	Ratio
`ML-KEM-512 keypair`	`35986` cycles	`35918` cycles	`1.00`
`ML-KEM-512 encaps`	`41033` cycles	`40842` cycles	`1.00`
`ML-KEM-512 decaps`	`52119` cycles	`52306` cycles	`1.00`
`ML-KEM-768 keypair`	`59553` cycles	`60759` cycles	`0.98`
`ML-KEM-768 encaps`	`66794` cycles	`67536` cycles	`0.99`
`ML-KEM-768 decaps`	`81139` cycles	`81425` cycles	`1.00`
`ML-KEM-1024 keypair`	`88514` cycles	`89028` cycles	`0.99`
`ML-KEM-1024 encaps`	`98571` cycles	`98877` cycles	`1.00`
`ML-KEM-1024 decaps`	`117454` cycles	`117716` cycles	`1.00`

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

oqs-bot

Arm Cortex-A76 (Raspberry Pi 5) benchmarks

Benchmark suite	Current: `60e019b`	Previous: `8dba13b`	Ratio
`ML-KEM-512 keypair`	`29402` cycles	`29440` cycles	`1.00`
`ML-KEM-512 encaps`	`34609` cycles	`34675` cycles	`1.00`
`ML-KEM-512 decaps`	`45191` cycles	`45264` cycles	`1.00`
`ML-KEM-768 keypair`	`50052` cycles	`50138` cycles	`1.00`
`ML-KEM-768 encaps`	`55297` cycles	`55220` cycles	`1.00`
`ML-KEM-768 decaps`	`70175` cycles	`70075` cycles	`1.00`
`ML-KEM-1024 keypair`	`73091` cycles	`73079` cycles	`1.00`
`ML-KEM-1024 encaps`	`81485` cycles	`81517` cycles	`1.00`
`ML-KEM-1024 decaps`	`101331` cycles	`101473` cycles	`1.00`

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

oqs-bot

Graviton2 (no-opt)

Benchmark suite	Current: `60e019b`	Previous: `8dba13b`	Ratio
`ML-KEM-512 keypair`	`61132` cycles	`59431` cycles	`1.03`
`ML-KEM-512 encaps`	`67969` cycles	`68011` cycles	`1.00`
`ML-KEM-512 decaps`	`86728` cycles	`86649` cycles	`1.00`
`ML-KEM-768 keypair`	`98894` cycles	`98924` cycles	`1.00`
`ML-KEM-768 encaps`	`111164` cycles	`110402` cycles	`1.01`
`ML-KEM-768 decaps`	`134569` cycles	`134811` cycles	`1.00`
`ML-KEM-1024 keypair`	`148703` cycles	`148822` cycles	`1.00`
`ML-KEM-1024 encaps`	`164019` cycles	`163875` cycles	`1.00`
`ML-KEM-1024 decaps`	`195825` cycles	`195498` cycles	`1.00`

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

oqs-bot

Graviton3 (no-opt)

Benchmark suite	Current: `60e019b`	Previous: `8dba13b`	Ratio
`ML-KEM-512 keypair`	`38872` cycles	`39685` cycles	`0.98`
`ML-KEM-512 encaps`	`44719` cycles	`44864` cycles	`1.00`
`ML-KEM-512 decaps`	`56426` cycles	`56842` cycles	`0.99`
`ML-KEM-768 keypair`	`64262` cycles	`64182` cycles	`1.00`
`ML-KEM-768 encaps`	`72260` cycles	`72669` cycles	`0.99`
`ML-KEM-768 decaps`	`87830` cycles	`88056` cycles	`1.00`
`ML-KEM-1024 keypair`	`95851` cycles	`96095` cycles	`1.00`
`ML-KEM-1024 encaps`	`106786` cycles	`106211` cycles	`1.01`
`ML-KEM-1024 decaps`	`127065` cycles	`127057` cycles	`1.00`

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

oqs-bot

SpacemiT K1 8 (Banana Pi F3) benchmarks

Benchmark suite	Current: `60e019b`	Previous: `8dba13b`	Ratio
`ML-KEM-512 keypair`	`225587` cycles	`225593` cycles	`1.00`
`ML-KEM-512 encaps`	`271816` cycles	`271821` cycles	`1.00`
`ML-KEM-512 decaps`	`346351` cycles	`346326` cycles	`1.00`
`ML-KEM-768 keypair`	`374190` cycles	`374216` cycles	`1.00`
`ML-KEM-768 encaps`	`433744` cycles	`433736` cycles	`1.00`
`ML-KEM-768 decaps`	`532470` cycles	`532650` cycles	`1.00`
`ML-KEM-1024 keypair`	`554460` cycles	`554463` cycles	`1.00`
`ML-KEM-1024 encaps`	`632572` cycles	`632521` cycles	`1.00`
`ML-KEM-1024 decaps`	`754426` cycles	`754358` cycles	`1.00`

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

hanno-becker · 2025-03-20T07:17:19Z

I am rather inclined not to do this. There is a pain point here in that T (*)[N] does not cast to const T (*)[N]. See https://gustedt.wordpress.com/2011/02/12/const-and-arrays/. Concretely, if you have a function void foo(const T[N][M]), and a local T x[N][M], then calling foo(x) is not valid C. You would need to explicitly cast foo((const T(*)[M]) x) (@tautschnig @remi-delmas-3000 Would such cast actually be UB?) For mlk_polyvec alone, this only surfaces twice when calling matvec_mul, so it's not too bad.

However, it's a fair question why one would not also make mlk_poly a typedef, which the above issue renders entirely impractical. So the only option we have is turning mlk_polyvec into an array, but not mlk_poly. That's good enough for gen_matrix, but a bit inconsistent.

If we don't do this, gen_matrix exhibits UB, albeit one which is highly unlikely to cause any practical problem.

oqs-bot

Mac Mini (M1, 2020) benchmarks

Benchmark suite	Current: `60e019b`	Previous: `8dba13b`	Ratio
`ML-KEM-512 keypair`	`12254` cycles	`12254` cycles	`1`
`ML-KEM-512 encaps`	`14838` cycles	`14846` cycles	`1.00`
`ML-KEM-512 decaps`	`19715` cycles	`19721` cycles	`1.00`
`ML-KEM-768 keypair`	`21051` cycles	`21051` cycles	`1`
`ML-KEM-768 encaps`	`23677` cycles	`23679` cycles	`1.00`
`ML-KEM-768 decaps`	`30514` cycles	`30515` cycles	`1.00`
`ML-KEM-1024 keypair`	`29978` cycles	`29978` cycles	`1`
`ML-KEM-1024 encaps`	`34478` cycles	`34478` cycles	`1`
`ML-KEM-1024 decaps`	`43428` cycles	`43442` cycles	`1.00`

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

hanno-becker · 2025-03-20T07:43:00Z

@mkannwischer Do you have an opinion here?

Another option would be to define mlk_polyvec as a typedef, but avoid mlk_polyvec[MLKEM_K] and instead use mlk_poly[MLKEM_K * MLKEM_K] throughout. There aren't many instances where matrices are used anyway.

oqs-bot

Arm Cortex-A72 (Raspberry Pi 4) benchmarks

Benchmark suite	Current: `60e019b`	Previous: `8dba13b`	Ratio
`ML-KEM-512 keypair`	`53183` cycles	`53196` cycles	`1.00`
`ML-KEM-512 encaps`	`61013` cycles	`61239` cycles	`1.00`
`ML-KEM-512 decaps`	`78268` cycles	`77578` cycles	`1.01`
`ML-KEM-768 keypair`	`89696` cycles	`89929` cycles	`1.00`
`ML-KEM-768 encaps`	`97419` cycles	`98860` cycles	`0.99`
`ML-KEM-768 decaps`	`122199` cycles	`123436` cycles	`0.99`
`ML-KEM-1024 keypair`	`135525` cycles	`134458` cycles	`1.01`
`ML-KEM-1024 encaps`	`148040` cycles	`147089` cycles	`1.01`
`ML-KEM-1024 decaps`	`181769` cycles	`180649` cycles	`1.01`

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

This commit changes the definition of `mlk_polyvec` from a struct wrapper `struct { mlk_poly vec[MLKEM_K] }` to a typedef for `mlk_poly[MLKEM_K]`. Similarly, it introduces `mlk_polymat` as a typedef for `mlk_poly[MLKEM_K * MLKEM_K]`. This ensures that when working with polynomial matrices -- that is, arrays of `mlk_polyvec`s -- the polynomial entries are consecutive in memory. This needs to be assumed for the correctness of `gen_matrix()`. From all we know, is a theoretical concern that without this change, compilers would be free to introduce padding into `mlk_polyvec`, breaking `gen_matrix()`. Yet, the behaviour of `gen_matrix` is strictly speaking UB, so even if a compiler does not use padding in `mlk_polyvec`, it is better to change it. Defining `mlk_polymat` as a typedef for `mlk_poly[MLKEM_K * MLKEM_K]` and not `mlk_polyvec[MLKEM_K]` is deliberate: In the latter case, one would not be able to pass a `mlk_polymat` to a function expecting `const polymat`, forcing cumbersome manual casts. Signed-off-by: Hanno Becker <[email protected]>

mkannwischer

LGTM. Thanks @hanno-becker for resolving this!
I have no concerns about this change.

hanno-becker requested a review from a team as a code owner March 20, 2025 05:08

hanno-becker force-pushed the polyvec_arr branch 2 times, most recently from 579738a to e7893df Compare March 20, 2025 05:12

hanno-becker marked this pull request as draft March 20, 2025 05:15

hanno-becker force-pushed the polyvec_arr branch 7 times, most recently from b713230 to 1aadee9 Compare March 20, 2025 06:49

hanno-becker added the benchmark this PR should be benchmarked in CI label Mar 20, 2025

oqs-bot reviewed Mar 20, 2025

View reviewed changes

hanno-becker changed the title ~~Define mlk_polyvec[_mulcache] as typedef for array~~ [TEST] Define mlk_polyvec[_mulcache] as typedef for array Mar 20, 2025

oqs-bot reviewed Mar 20, 2025

View reviewed changes

hanno-becker force-pushed the polyvec_arr branch 5 times, most recently from 0574488 to 3b28c14 Compare March 23, 2025 14:10

hanno-becker changed the title ~~[TEST] Define mlk_polyvec[_mulcache] as typedef for array~~ Define mlk_polyvec[_mulcache] as typedef for array Mar 23, 2025

hanno-becker force-pushed the polyvec_arr branch 5 times, most recently from 2a17dbe to 60e019b Compare March 23, 2025 18:37

hanno-becker marked this pull request as ready for review March 23, 2025 19:32

hanno-becker added benchmark this PR should be benchmarked in CI and removed benchmark this PR should be benchmarked in CI labels Mar 23, 2025

oqs-bot reviewed Mar 23, 2025

View reviewed changes

hanno-becker force-pushed the polyvec_arr branch from 60e019b to 00d3f1c Compare March 25, 2025 09:51

mkannwischer approved these changes Mar 25, 2025

View reviewed changes

mkannwischer merged commit 118f091 into main Mar 25, 2025
200 checks passed

mkannwischer deleted the polyvec_arr branch March 25, 2025 12:06

hanno-becker mentioned this pull request Mar 25, 2025

CBMC: Refine bounds for input and output of base multiplication #906

Draft

Define mlk_polyvec[_mulcache] as typedef for array #895

Define mlk_polyvec[_mulcache] as typedef for array #895

Uh oh!

Conversation

hanno-becker commented Mar 20, 2025 • 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

Intel Xeon 4th gen (c7i)

Uh oh!

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

Uh oh!

Choose a reason for hiding this comment

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

Uh oh!

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

Uh oh!

Choose a reason for hiding this comment

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

AMD EPYC 3rd gen (c6a)

Uh oh!

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

Uh oh!

Choose a reason for hiding this comment

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 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

Arm Cortex-A55 (Snapdragon 888) 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 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

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

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

Graviton2

Uh oh!

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

Uh oh!

Choose a reason for hiding this comment

Graviton3

Uh oh!

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

Uh oh!

Choose a reason for hiding this comment

Graviton4 (no-opt)

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

Define `mlk_polyvec[_mulcache]` as typedef for array #895

Define `mlk_polyvec[_mulcache]` as typedef for array #895

hanno-becker commented Mar 20, 2025 •

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

hanno-becker commented Mar 20, 2025 •

edited

Loading

oqs-bot left a comment •

edited

Loading

hanno-becker commented Mar 20, 2025 •

edited

Loading