github/concrete
1
1
Fork 0
mirror of https://github.com/zama-ai/concrete.git synced 2026-02-09 03:55:04 -05:00
Code Issues Packages Projects Releases Wiki Activity
1,244 Commits 185 Branches 425 Tags
bb728959ec0c3fabb282042f955869caf755f72c
Commit Graph

1244 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Quentin Bourgerie
bb728959ec chore: Integrate concrete-optimizer CI 2023-03-07 16:46:31 +01:00
Quentin Bourgerie
c20726e487 chore: Refactor the ci in order to dispatch dependeding of file changes 2023-03-07 14:40:32 +01:00
Quentin Bourgerie
ea46bd70bb chore: Fix compiler CI 2023-03-07 09:08:29 +01:00
Quentin Bourgerie
6cd9fa6eff chore: concrete-optimizer is no more build by make 2023-03-06 17:19:39 +01:00
Quentin Bourgerie
2f9c7da6c6 chore: Fix test and build compiler with the new monorepo layout 2023-03-06 17:17:29 +01:00
David Testé
556ff1249b chore(ci): rename some worflow secrets 2023-03-06 14:01:51 +01:00
David Testé
24457bd3e8 chore(ci): change private ssh key secret name 2023-03-06 14:01:51 +01:00
David Testé
5beb13b435 chore(ci): move ci directory to to repository root 2023-03-06 14:01:51 +01:00
Quentin Bourgerie
86422cf2d5 chore: Integrate concrete-optimizer in the mono-repo 2023-03-06 13:55:27 +01:00
Quentin Bourgerie
7f4bdfd557 chore: Move concrete-optimizer to the concrete mono repo 2023-03-06 13:53:48 +01:00
Quentin Bourgerie
203393056b chore: Move to the mono-repo layout 2023-03-06 13:34:22 +01:00
Quentin Bourgerie
a279cbe88e chore: Move llvm-project in third_party 2023-03-06 10:50:21 +01:00
Quentin Bourgerie
06d3c316e7 chore: Move concrete-cpu backend to the concrete mono repository 
The main contributors of concrete-cpu are

Co-authored-by: Mayeul@Zama <mayeul.debellabre@zama.ai>
Co-authored-by: sarah <sarah.elkazdadi@zama.ai>
 2023-03-03 16:20:18 +01:00
Quentin Bourgerie
f65d6e0b44 chore: Integrate 'integrate-concrete-compiler' into the mono-repo 2023-03-03 09:26:10 +01:00
Quentin Bourgerie
f975415e86 chore: Integrate concrete-compiler to the mono-repo 2023-03-03 09:20:01 +01:00
Quentin Bourgerie
ac2f761a79 chore: Project setup 2023-03-02 10:06:11 +01:00
Quentin Bourgerie
a62b3b1d74 refactor: Integrate concrete-cpu and remove concrete-core 
Co-authored-by: Mayeul@Zama <mayeul.debellabre@zama.ai>
 2023-03-02 09:26:31 +01:00
rudy
9c784a2243 fix(ci): prepare_release, GLIBC_VER was not used 2023-03-01 11:52:45 +01:00
rudy
e399d1fed8 fix(ci): prepare_release, macos, bad wheel path 2023-02-28 11:52:29 +01:00
rudy
9c0a686a78 fix: optimizer conversion for addMaxpool2d 2023-02-28 10:04:10 +01:00
rudy
19bc9b2933 fix(ci): prepare_release, typo on PATH 2023-02-28 09:17:10 +01:00
aPere3
15f07addc7 fix(concrete-compiler): abort compilation on first error 
To prevent infinite compilation in case of potential circular
includes
 2023-02-27 10:59:28 +01:00
Mayeul@Zama
220d733e98 chore(cmake): use HPX_INCLUDE_DIRS for ConcretelangRuntime only 2023-02-27 10:56:35 +01:00
rudy
00b9ebc60a fix: benchmark use the encoding field 2023-02-27 10:54:57 +01:00
Quentin Bourgerie
7ac6a97c85 chore(ci): Fix benchamrk on main as inputs are not setup for automatic merge benchmark 2023-02-24 11:27:55 +01:00
Quentin Bourgerie
7070eff562 chore(ci): Fix benchmark workflow 2023-02-24 10:17:28 +01:00
youben11
f2c1dc6392 fix: set client parameters with correct chunked int options 
were set with default values
 2023-02-22 10:50:26 +01:00
Quentin Bourgerie
4e0f0fa5b0 fix(manp): Fixing computation of negative constant 2023-02-21 18:04:13 +01:00
youben11
2e5f92b6b8 refactor: remove BConcrete dialect 
- no more Concrete ciphertext/plaintext types: they are represented using standard MLIR types (int/tensor)
- Technically BConcrete was renamed to Concrete, and old Concrete was
  removed
- TFHE -> Concrete now takes into account the conversion of tensor of
  ciphertext into tensors of an additional dimension (LWE dim)
- Bufferization now works in Concrete
- Old Concrete optimization were moved to TFHE
- Concrete is now the dialect that lowers to CAPI calls
- TFHE -> Concrete now uses OpConversionPattern and is much cleaner in
  terms of type conversion
- Disabled tests for batching, as there was something weird about it:
  batchable operations implemented in Concrete but pass run in FHELinalg
 2023-02-21 16:16:55 +01:00
Umut
d8eafabd22 feat: implement manual manp calculation for FHE.mul_eint 2023-02-21 14:25:26 +01:00
Umut
bc69c87d62 feat: implement maxpool2d operation 2023-02-21 14:25:26 +01:00
Alexandre Péré
56bdb05be3 feat(concrete-compiler): implement a new round operator in the fhe dialect 2023-02-20 08:57:30 +01:00
Quentin Bourgerie
3175d2ec85 chore: fix typo in benchmark workflow 2023-02-17 15:17:25 +01:00
Quentin Bourgerie
8fab9a4038 chore: select benchmark name in benchmark workflow 2023-02-17 15:06:15 +01:00
Quentin Bourgerie
0307227431 tests: Add cifar_8bits benchmark 2023-02-17 14:46:27 +01:00
Quentin Bourgerie
4a3c33b60a tests: Remove unsigned prefix to have the same reference than before 2023-02-17 14:08:34 +01:00
Alexandre Péré
52ad40c9cf feat(concrete-compiler): adds a tracing op in all dialects. 2023-02-17 09:04:42 +01:00
Andi Drebes
5d3af16617 test(compiler): Add check for omp.parallel after lowering a parallelizable FHELinalg op 
This test ensures that at least one parallel region is generated for
an FHELinalg operation that is guaranteed to result in a parallel loop
when `concretecompiler` is invoked with `--parallelize`.
 2023-02-14 09:53:02 +01:00
Andi Drebes
fc02bb5cd8 fix(compiler): Copy attributes when reinstantiating scf.for ops 
The switch to reinstantiating conversion patterns for the conversion
from FHE to TFHE in commit 73fd6c5fe7
caused all attributes of `scf.for` operations to be dropped during the
conversion. This included the custom attribute `parallel`, which is
exploited further down the compilation pipeline to generate parallel
code. As a result, the performance of end-to-end benchmarks dropped
significantly.

This patch copies all attributes of `scf.for` operations upon
reinstantiation, which solves the performance regression.
 2023-02-14 09:53:02 +01:00
Antoniu Pop
3ab10c8d3f fix(GPU runtime): fix BS & KS GPU allocation and synchronization, avoid re-allocation of KSK. 2023-02-14 08:22:15 +00:00
rudy
717c8c815f fix: ci macos test 2023-02-13 09:20:25 +01:00
youben11
59d35619a8 feat: lower and exec boolean mux 2023-02-09 09:43:48 +01:00
youben11
d0ae2563fa test: add chunked eint e2e tests 2023-02-07 12:27:01 +01:00
youben11
7e60f87141 refactor: remove chunked_eint 
the pass transforming operations on chunked_eint will operate now on
eint
 2023-02-07 12:27:01 +01:00
youben11
bb87d29934 feat: support chunked integer during enc/dec/exec 2023-02-07 12:27:01 +01:00
youben11
d41d14dbb8 feat: lower FHE.add on eint64 to ops on smaller chunks 
this is a first commit to support operations on U64 by decomposing them
into smaller chunks (32 chunks of 2 bits). This commit introduce the
lowering pass that will be later populated to support other operations.
 2023-02-07 12:27:01 +01:00
aPere3
fb680340f9 feat(concrete-compiler): add new ciphertext multiplication operator 2023-02-06 11:30:31 +01:00
rudy
117e15cc05 fix: build python package during build 2023-02-06 09:36:29 +01:00
rudy
a01fab0a90 feat: Makefile, python-package target 2023-02-06 09:36:29 +01:00
rudy
37421774a0 fix: python package requires wheel 2023-02-06 09:36:29 +01:00