Agnes Leroy
e5742e63e9
chore(gpu): add compute-sanitizer run on H100
2025-11-21 09:15:04 +01:00
Agnes Leroy
bc0a4fb0d6
chore(gpu): add classical pbs tests to HL API, complement missing tests, classical in long run
2025-11-21 09:15:04 +01:00
Enzo Di Maria
c6709a82c0
refactor(gpu): match_value to backend with multiple streams
2025-11-20 17:11:15 +01:00
Agnes Leroy
0f5e538f9a
chore(gpu): add better panic when calling old compression functions
2025-11-20 17:06:06 +01:00
David Testé
58378b7972
chore(bench): add dedicated targets for aes cuda benchmarks
2025-11-20 16:58:06 +01:00
Arthur Meyre
9b5df143cb
chore: update dp_ks_pbs128_packingks test to take metaparameters
2025-11-20 16:03:50 +01:00
Arthur Meyre
cbac102a0d
chore: update dp_ks_ms test to take metaparameters
2025-11-20 16:03:50 +01:00
Arthur Meyre
0da331590c
chore: update cpk_ks_ms test to take metaparameters
...
- also added a seemingly missing sanity check for cpk_ks_ms
2025-11-20 16:03:50 +01:00
Arthur Meyre
e5a15b33d9
chore: fix incorrect number in seed_bytes for rerand noise check test
...
- has no impact on correctness of test, in addition the size of the array
would have been the same due to integer division
2025-11-20 16:03:50 +01:00
Arthur Meyre
d68645c493
chore: update br_rerand_dp_ks_ms test to take metaparameters
2025-11-20 16:03:50 +01:00
Arthur Meyre
a72de66744
chore: update br_dp_packing_ms test to take metaparameters
2025-11-20 16:03:50 +01:00
Arthur Meyre
1b924fa872
chore: update br_dp_ks_ms test to take metaparameters
2025-11-20 16:03:50 +01:00
Arthur Meyre
560f595620
chore: remove temporary test param
2025-11-20 16:03:50 +01:00
Mayeul@Zama
f4665886ee
fix(integer): fix StaticUnsignedBigInt cast into u128
2025-11-20 15:40:39 +01:00
Beka Barbakadze
80cacbd079
feat(gpu): add boolean bitops in cuda backend
2025-11-20 14:56:21 +01:00
tmontaigu
1b17cfc0f8
chore(c-api): add missing docs for some C function
...
Add explanations on the fact that config and config buider
do not need to be freed/
2025-11-20 14:46:54 +01:00
David Testé
a5c248566d
chore(ci): run parameters_check workflow if files change
2025-11-20 12:36:21 +01:00
David Testé
eb2cf19d10
chore(ci): install rust toolchain for parameters checks
...
This is needed since our AWS EC2 image doesn't embed Rust toolchain.
2025-11-20 11:01:37 +01:00
David Testé
4928efcca8
chore(ci): remove check on secret availability for params checks
...
This is needed only when the triggering event is pull_request and PR comes from an external contributor.
2025-11-20 11:01:37 +01:00
Nicolas Sarlin
edb435bd46
chore: update msrv to 1.91.1
2025-11-20 09:29:37 +01:00
David Testé
071e70c037
chore(bench): fix benchmark id pattern for aes and aes256
2025-11-19 17:23:05 +01:00
Arthur Meyre
19c2146b2d
chore: remove redundant log2/powf in params_to_file.rs
2025-11-19 12:59:26 +01:00
David Testé
c5bac7994a
chore(ci): fix typos in newly checked ci folder
2025-11-19 11:51:02 +01:00
David Testé
bdef5b358e
chore(ci): ignore only specific files in ci folder
...
The whole ci/ folder was ignored, that would cause issue regarding search capability in some IDE.
2025-11-19 11:51:02 +01:00
Arthur Meyre
4229523c50
chore: add artificial intelligence contribution guidelines
2025-11-19 10:43:34 +01:00
Nicolas Sarlin
ac6178fd35
chore(zk): add batched mode for verification pairings
2025-11-19 09:24:13 +01:00
David Testé
893c321e60
chore(ci): update repositories before any other apt operation
...
This is done to avoid failure on dependencies' installation.
2025-11-18 14:13:02 +01:00
David Testé
51139e6e82
chore(ci): use builds result to pass branch protection
...
Instance teardown might fail regardless the result of build
commands. To make CI green in pull-request we need to check the
result of the build commands instead of the whole workflow result.
Doing so ensure a PR could pass branch protection rules even if
the teardown fails.
2025-11-18 11:44:56 +01:00
David Testé
5ce068642e
chore(ci): replace workflow name in cpu weekly bench
2025-11-18 11:42:10 +01:00
David Testé
5c2fe133e1
chore(ci): use explicit boolean checking on scheduled cpu bench
...
It seems that boolean output set by user is not coerced into
boolean type by the runner. Rather it checks if the string
associated ("true" or "false") is not empty.
2025-11-18 11:42:10 +01:00
David Testé
d12bbd848c
chore(ci): fix op_flavor generation for weekly cpu bench
...
Sub-workflow is supposed to handle comma-separated values not an array.
2025-11-18 11:42:10 +01:00
David Testé
c30e16db9c
chore(ci): move workflows runners to aws ec2
...
Done to reduce CI running costs.
2025-11-18 10:44:50 +01:00
Mayeul@Zama
f9268b889f
chore(bench): revert print bench id
...
This reverts commit ef07963767 .
2025-11-17 11:23:50 +01:00
David Testé
1568f7c532
chore(ci): use aws ec2 as runner provider for cargo builds
2025-11-17 10:19:26 +01:00
Arthur Meyre
89bf6a3331
chore: add tool to update AP params' msg and carry moduli
rtl_bin-v2.9
2025-11-17 09:41:53 +01:00
Arthur Meyre
82cebb9b26
test(shortint): add compression atomic pattern for noise checks
...
- noise checks and pfail based on expected noise have been added
- compatible with KS PBS and KS32 PBS
2025-11-17 09:41:53 +01:00
Arthur Meyre
ba5f4850b9
chore: update naming of noise simulation primitive to avoid clashes
...
- makes it clearer from which parameters some noise simulation primitives
are built from
2025-11-17 09:41:53 +01:00
Arthur Meyre
7197b85ec9
chore: lift some restrictions on confidence interval function
...
- if a value is computed it will be correct
- if the value is not finite (NaN or infinity) we panic with a message to
the user indicating what course of action they can take
- ideally we would want to use a scientific crate written in rust, xsf-rust
seemed promising but the dependency on clang + libclang is proving more
annoying than not, given we would need a single function from xsf (and it's
hard to translate all the required pieces) we keep a sort of status quo
- statrs issue : https://github.com/statrs-dev/statrs/issues/361
2025-11-17 09:41:53 +01:00
Enzo Di Maria
54c8c5e020
chore(gpu): no crash with aes benches if oom error
2025-11-14 17:02:33 +01:00
David Testé
164fc26025
chore(ci): add placeholders for documentation benchmarks
...
This is done to be able to execute CI in further development.
2025-11-14 16:48:49 +01:00
David Testé
ad818ee117
chore(ci): add placeholder for cargo_build_common.yml
...
This is done to be able to execute CI in further development.
Also, we won't have to temporary lift the branch protection rules
to be able to merge since this upcoming development is a rework
cargo_build.yml workflow.
2025-11-14 16:48:49 +01:00
Agnes Leroy
df73c36cbf
fix(gpu): fix decomposition algorithm not matching the theory
2025-11-14 16:36:35 +01:00
David Testé
a33c12d5a9
chore(ci): fix zizmor findings in workflows
2025-11-14 15:24:10 +01:00
David Testé
522a612ad4
chore(ci): update zizmor and use zizmor-action in workflow
2025-11-14 15:24:10 +01:00
David Testé
f8c998f0da
chore(ci): avoid unwanted cancellation in csprng tests
2025-11-14 15:18:04 +01:00
Arthur Meyre
84c80c529d
chore: remove redundant clones
...
co-authored-by: Himess <95512809+Himess@users.noreply.github.com >
2025-11-14 14:14:39 +01:00
Arthur Meyre
c3c892708a
chore: fix comment confusing comment in decomposer.rs
...
- function is documented and the comment did not match, the behavior is
checked in a test
2025-11-14 14:14:24 +01:00
Agnes Leroy
4f9f4982f6
fix(gpu): fix memory leak in rerand
2025-11-14 14:00:01 +01:00
Arthur Meyre
d75844dea5
fix(core): fix decomposition algorithm not matching the theory
...
- problem arose from a shift being done on an unsigned value which did not
keep the signed characteristics of the represented signed value
- introduce an arithmetic_shift on the UnsignedInteger trait with a blanket
implementation
- add the edge case which revelead the issue
- the asm has been verified to only change for the shift operation being
applied, meaning no performance regression will occurr
2025-11-14 13:52:17 +01:00
David Testé
ef07963767
chore(bench): print bench id before running the benchmark
...
Done to circumvent criterion limitation regarding automatic
truncation of long benchmark ID.
Using a println() call we ensure the complete name is displayed
before benchmark execution to ease manual parsing and debugging.
2025-11-14 13:45:04 +01:00