Commit Graph

84 Commits

Author SHA1 Message Date
Pedro Alves
b91a736cd6 fix(backend-gpu): minor fixes to the rust API 2023-07-26 10:43:43 -03:00
Pedro Alves
ae43e7ffe5 feat(backend-gpu): implement the multi-bit PBS 2023-07-11 12:25:18 -03:00
Pedro Alves
5b103acbd9 feat(backend-gpu): expose a rust API for concrete-cuda 2023-07-07 13:15:56 -03:00
Bourgerie Quentin
1b0106302b fix(optimizer/backend-cpu): Fix Cargo.lock to avoid flaky dependencies 2023-06-30 16:08:43 +02:00
Pedro Alves
a4ef7543a6 feat(backend-gpu): introduce a function to check for cooperative groups support 2023-06-27 08:48:39 -03:00
Mayeul@Zama
6877f6c9e3 fix(backend-cpu): fix clippy lint 2023-06-27 11:14:40 +02:00
Mayeul@Zama
97b13e871c feat(optimizer): introduce fft precision 2023-06-15 10:48:07 +02:00
Mayeul@Zama
8897acab3d feat(optimizer): fix noise model tests compilation 2023-06-15 10:48:07 +02:00
youben11
566f119a06 feat(backend): expose random gaussian function 2023-06-05 16:48:52 +01:00
Agnes Leroy
3561b51329 bench(backend-gpu): add multi-gpu support in pbs benchmark
Remove amortized PBS bench & bench with copy as well, it is deprecated
2023-05-24 15:32:20 +02:00
Agnes Leroy
72ee8eb572 fix(backend-gpu): decrease opt values for N >= 2048 for new low lat pbs 2023-05-23 11:34:53 +02:00
Agnes Leroy
b174f81687 fix(backend-gpu): avoid warp branching in the PBS 2023-05-17 15:01:37 +02:00
Pedro Alves
43a58e1820 refactor(backend-gpu): replace the amortized PBS by the new low latency PBS in boolean gates. 2023-05-05 18:15:06 +02:00
Pedro Alves
336709a011 refactor(backend-gpu): replace the amortized PBS by the new low latency PBS in circuit bootstrap 2023-05-05 18:15:06 +02:00
Pedro Alves
dda65a4ac6 refactor(backend-gpu): implement an alternative to the low latency pbs 2023-05-05 18:15:06 +02:00
Beka Barbakadze
d57cbfd705 fix(backend-cuda): modify bit extract to work for batch lwe ciphertexts 2023-05-05 13:01:07 +02:00
Agnes Leroy
c67c246e02 fix(ci): revert "fix(ci): fix concrete-cuda ci" and fix ci properly
Install nightly Rust instead of stable
Build concrete-cpu with nightly cargo
2023-05-04 15:28:08 +02:00
youben11
d7ecd66387 feat(backend): add concrete-cpu noise-model CAPI 2023-05-04 10:36:25 +01:00
Agnes Leroy
4a0c9bdf7d chore(backend-gpu): fix formatting 2023-04-28 16:51:45 -03:00
Agnes Leroy
1e8a5b0d14 fix(backend-gpu): fix cmux tree for tau > 1 2023-04-25 10:29:07 +02:00
Pedro Alves
ec5c2c93cd chore(backend-gpu): modifies the benchmark for the low latency pbs kernel to use the correct number of inputs 2023-04-24 20:52:15 +02:00
Mayeul@Zama
3558ba5586 feat(backend-cpu): choose fastest aes implementation at runtime 2023-04-24 15:02:22 +02:00
Pedro Alves
6d2bf674fe feat(backend-gpu): add a new column to the benchmark tool related to the throughput 2023-04-21 13:38:58 +02:00
Pedro Alves
681e3081e3 docs(backend-gpu): improve concrete-cuda docs regarding tests and benchmarks. 2023-04-18 08:47:44 -03:00
Mayeul@Zama
315c1a08e1 fix(backend-cpu): fix Security framework linking in tests 2023-04-18 09:54:42 +02:00
Mayeul@Zama
348fe028cf feat(backend-cpu): use Apple secure enclave seeder on M1 2023-04-17 18:13:18 +02:00
Pedro Alves
fa226e232f feat(backend-gpu): refactor the Wop-PBS implementation to support a wider set of parameters 2023-04-17 09:45:17 +02:00
Mayeul@Zama
84a3dd7a8d chore(backend-cpu): use zig build 2023-04-13 11:09:54 +02:00
Mayeul@Zama
53fa78cfbd chore(backend-cpu): simplify import in zig tests 2023-04-13 11:09:54 +02:00
youben11
3e0d959b12 chore(common): update repo name 2023-04-11 16:35:42 +02:00
Mayeul@Zama
e3c5eee6fe docs(backend-cpu): add readme 2023-04-11 14:55:57 +02:00
Agnes Leroy
766df20232 chore(backend/gpu): add README file for concrete-cuda 2023-04-06 16:29:01 +02:00
Agnes Leroy
b3c1446816 bench(backend/gpu): change benchmark parameters to match tfhe-rs ones 2023-04-06 14:06:14 +02:00
Agnes Leroy
aa9147b5b7 bench(backend/gpu): reduce amount of inputs benchmarks in pbs 2023-04-03 17:49:33 +02:00
Pedro Alves
fb862e37e8 bench(backend/gpu): fixes a memory leakage and refactors the verification for the grid size of the low latency kernel 2023-04-03 17:49:33 +02:00
Agnes Leroy
74ded6c48b bench(backend|gpu): fix low latency pbs bench 2023-04-03 17:49:33 +02:00
Agnes Leroy
32bcda3a93 bench(backend/cuda): restrict the number of inputs for large poly sizes 2023-04-03 17:49:33 +02:00
Pedro Alves
c236dffdd8 chore(backend/cuda): reduces memory consumption in the bootstrap benchmark 2023-04-03 17:49:33 +02:00
Agnes Leroy
d9652b8936 bench(ci): fix concrete-cuda benchmarks 2023-04-03 17:49:33 +02:00
Beka Barbakadze
adf7adffa3 feat(backend/cuda): add tests for fft in concrete-cuda. fix shared memory mode choosing bug, when compute capability was lower than 8. 2023-03-30 11:12:08 +02:00
Pedro Alves
dab31433d9 fix(backend): fix ci after 08e8012
Google benchmark is built twice due to the new bench infrastructure for
concrete-cuda, this commit fixes it by introducing
CONCRETE_CUDA_BUILD_TESTS and CONCRETE_CUDA_BUILD_BENCHMARKS options to skip
unecessary builds.
2023-03-29 14:36:35 +02:00
Pedro Alves
08e8012061 bench(backend): add a benchmark tool for concrete-cuda 2023-03-28 10:19:57 +02:00
Quentin Bourgerie
8b4dc7be82 chore(concrete-cpu): Use concrete-fft nightly when nightly features is enable to use avx512 2023-03-28 08:50:33 +02:00
Agnes Leroy
573f4756fd fix(concrete_cuda): fix buffer size type in the scratches, int is not big enough 2023-03-22 15:55:15 +01:00
Quentin Bourgerie
8349d7824e chore(concrete-cpu): Move concrete-cpu implementation 2023-03-22 14:22:11 +01:00
Quentin Bourgerie
803c302e69 chore: Integrate concrete-cpu/noise-model to the mono repo 2023-03-21 15:08:43 +01:00
Quentin Bourgerie
b18148788d chore(concrete-cpu): Remove security curves from the concrete-cpu noise-model 2023-03-21 15:05:28 +01:00
Quentin Bourgerie
b7ba9c38e8 chore(): Move concrete-cpu noise model to the mono repo layout 2023-03-21 15:00:17 +01:00
Mayeul@Zama
0936cfccd2 chore(cpu): remove GgswLevelRow container 2023-03-21 14:37:08 +01:00
Mayeul@Zama
4ae3be5c4f chore(cpu): separate LevCt 2023-03-21 14:37:08 +01:00