mirror of
https://github.com/luxfhe/torus.git
synced 2026-10-30 17:07:26 +00:00
zeekay
b832ba2c39
Forked from upstream concrete (BSD-3-Clause-Clear, © 2024 ZAMA) at HEAD
ecb729bb. Modifications copyright © 2026 Lux Industries Inc.
Rebranding scope:
- Python module concrete → torus (across frontends/torus-python)
- Rust module concrete-* → torus-* (across frontends/torus-rust)
- MLIR dialect concretelang → toruslang (across compilers/torus-compiler)
- Compiler optimizer concrete-optimizer → torus-optimizer
- Backend concrete-cpu / concrete-cuda → torus-cpu / torus-cuda
- Wire protocol concrete-protocol → torus-protocol
- Parameter-curves concrete-security-curves-* → torus-security-curves-*
- Public symbols concretelib → toruslib, ConcreteLang → TorusLang
- URLs github.com/zama-ai/concrete → github.com/luxfhe/torus
- URLs docs.zama.ai/concrete → docs.luxfhe.com/torus
Legal compliance:
- LICENSE.txt retains original Copyright © 2024 ZAMA verbatim (§1-§2)
- LICENSE.txt adds modifications copyright © 2026 Lux Industries Inc.
- §3 endorsement clause: Torus is marketed exclusively as a Lux product;
no Zama name used in README, docs, or marketing
- NOTICE file provides factual attribution per legal best practice
(not endorsement; required by upstream BSD-3-Clause-Clear)
README + NOTICE rewritten from scratch with Lux brand:
- Quick-start in Python (torus.fhe)
- Architecture diagram showing the gen-1 vtbl integration path
(luxcpp/gpu → lux-private/gpu-kernels for CUDA/Metal/WebGPU)
- Scheme coverage table: TFHE (luxfi/fhe), CKKS/BGV/BFV (luxfi/lattice)
- Lux FHE family table: Torus / TorusVM / TorusEVM / TorusML / TorusNet
Known limitations (phase-2A scope):
- PyPI package torus-fhe not yet published; the torus Python module
is renamed structurally but pip-installable distribution is a
follow-up (Phase 2A.1)
- MLIR dialect rename is text-level; the compiler-internal LLVM
rebuild (concrete-compiler-internal-llvm-project submodule) is
structural and works; tablegen files renamed concrete→torus but the
generated .inc files in the source tree will regenerate at build
time
- backends/torus-cuda/implementation/ is an upstream submodule
pointer to the tfhe-rs source (not fetched in the staging clone);
the CUDA backend body is delivered out-of-tree
- ~100 test-fixture variable names like concrete_encoded_sample
retain "concrete" as a math term (vs symbolic); these are inside
tests/, not the public API; cleanup is a follow-up commit
- 295 Python `from concrete import ...` residuals in luxfhe/ml-sdk
consumer remain; that repo's import update is Phase 2B
This commit is the seed for the luxfhe/torus repo. Push as the first
commit to a fresh GitHub repo; do not re-import upstream Zama git
history.
365 lines
7.2 KiB
Python
365 lines
7.2 KiB
Python
"""
|
|
Tests of execution of matmul operation.
|
|
"""
|
|
|
|
import numpy as np
|
|
import pytest
|
|
|
|
from torus import fhe
|
|
|
|
|
|
@pytest.mark.parametrize(
|
|
"lhs_shape,rhs_shape,bounds",
|
|
[
|
|
pytest.param(
|
|
(3, 2),
|
|
(2, 3),
|
|
(0, 3),
|
|
),
|
|
pytest.param(
|
|
(3, 2),
|
|
(2, 3),
|
|
(0, 127),
|
|
),
|
|
pytest.param(
|
|
(1, 2),
|
|
(2, 1),
|
|
(0, 3),
|
|
),
|
|
pytest.param(
|
|
(3, 3),
|
|
(3, 3),
|
|
(0, 3),
|
|
),
|
|
pytest.param(
|
|
(2, 1),
|
|
(1, 2),
|
|
(0, 7),
|
|
),
|
|
pytest.param(
|
|
(2,),
|
|
(2,),
|
|
(0, 7),
|
|
),
|
|
pytest.param(
|
|
(5, 5),
|
|
(5,),
|
|
(0, 3),
|
|
),
|
|
pytest.param(
|
|
(5,),
|
|
(5, 5),
|
|
(0, 3),
|
|
),
|
|
pytest.param(
|
|
(5,),
|
|
(5, 5),
|
|
(-127, 127),
|
|
),
|
|
pytest.param(
|
|
(5,),
|
|
(5, 3),
|
|
(0, 3),
|
|
),
|
|
pytest.param(
|
|
(5, 3),
|
|
(3,),
|
|
(0, 3),
|
|
),
|
|
pytest.param(
|
|
(5,),
|
|
(4, 5, 3),
|
|
(-5, 5),
|
|
),
|
|
pytest.param(
|
|
(4, 5, 3),
|
|
(3,),
|
|
(0, 5),
|
|
),
|
|
pytest.param(
|
|
(5,),
|
|
(2, 4, 5, 3),
|
|
(0, 5),
|
|
),
|
|
pytest.param(
|
|
(2, 4, 5, 3),
|
|
(3,),
|
|
(-1, 5),
|
|
),
|
|
pytest.param(
|
|
(5, 4, 3),
|
|
(3, 2),
|
|
(0, 5),
|
|
),
|
|
pytest.param(
|
|
(4, 3),
|
|
(5, 3, 2),
|
|
(0, 5),
|
|
),
|
|
pytest.param(
|
|
(2, 5, 4, 3),
|
|
(3, 2),
|
|
(0, 5),
|
|
),
|
|
pytest.param(
|
|
(5, 4, 3),
|
|
(1, 3, 2),
|
|
(0, 5),
|
|
),
|
|
pytest.param(
|
|
(1, 4, 3),
|
|
(5, 3, 2),
|
|
(0, 5),
|
|
),
|
|
pytest.param(
|
|
(5, 4, 3),
|
|
(2, 1, 3, 2),
|
|
(0, 5),
|
|
),
|
|
pytest.param(
|
|
(2, 1, 4, 3),
|
|
(5, 3, 2),
|
|
(0, 5),
|
|
),
|
|
],
|
|
)
|
|
def test_constant_matmul(lhs_shape, rhs_shape, bounds, helpers):
|
|
"""
|
|
Test matmul where one of the operators is a constant.
|
|
"""
|
|
|
|
configuration = helpers.configuration()
|
|
|
|
minimum, maximum = bounds
|
|
|
|
lhs_cst = list(np.random.randint(minimum, maximum, size=lhs_shape))
|
|
rhs_cst = list(np.random.randint(minimum, maximum, size=rhs_shape))
|
|
|
|
@fhe.compiler({"x": "encrypted"})
|
|
def lhs_operator(x):
|
|
return x @ rhs_cst
|
|
|
|
@fhe.compiler({"x": "encrypted"})
|
|
def rhs_operator(x):
|
|
return lhs_cst @ x
|
|
|
|
@fhe.compiler({"x": "encrypted"})
|
|
def lhs_function(x):
|
|
return np.matmul(x, rhs_cst)
|
|
|
|
@fhe.compiler({"x": "encrypted"})
|
|
def rhs_function(x):
|
|
return np.matmul(lhs_cst, x)
|
|
|
|
lhs_inputset = [np.random.randint(minimum, maximum, size=lhs_shape) for i in range(100)]
|
|
rhs_inputset = [np.random.randint(minimum, maximum, size=rhs_shape) for i in range(100)]
|
|
|
|
lhs_operator_circuit = lhs_operator.compile(lhs_inputset, configuration)
|
|
rhs_operator_circuit = rhs_operator.compile(rhs_inputset, configuration)
|
|
lhs_function_circuit = lhs_function.compile(lhs_inputset, configuration)
|
|
rhs_function_circuit = rhs_function.compile(rhs_inputset, configuration)
|
|
|
|
lhs_sample = lhs_inputset[-1]
|
|
rhs_sample = rhs_inputset[-1]
|
|
|
|
helpers.check_execution(lhs_operator_circuit, lhs_operator, lhs_sample)
|
|
helpers.check_execution(rhs_operator_circuit, rhs_operator, rhs_sample)
|
|
helpers.check_execution(lhs_function_circuit, lhs_function, lhs_sample)
|
|
helpers.check_execution(rhs_function_circuit, rhs_function, rhs_sample)
|
|
|
|
|
|
test_matmul_shape_and_bounds = [
|
|
(
|
|
(3, 2),
|
|
(2, 3),
|
|
(0, 3),
|
|
),
|
|
(
|
|
(3, 2),
|
|
(2, 3),
|
|
(0, 127),
|
|
),
|
|
(
|
|
(1, 2),
|
|
(2, 1),
|
|
(0, 3),
|
|
),
|
|
(
|
|
(3, 3),
|
|
(3, 3),
|
|
(0, 3),
|
|
),
|
|
(
|
|
(2, 1),
|
|
(1, 2),
|
|
(0, 7),
|
|
),
|
|
(
|
|
(2,),
|
|
(2,),
|
|
(0, 7),
|
|
),
|
|
(
|
|
(5, 5),
|
|
(5,),
|
|
(0, 3),
|
|
),
|
|
(
|
|
(5,),
|
|
(5, 5),
|
|
(0, 3),
|
|
),
|
|
(
|
|
(5,),
|
|
(5, 5),
|
|
(-63, 63),
|
|
),
|
|
(
|
|
(2,),
|
|
(2, 7),
|
|
(-63, 0),
|
|
),
|
|
(
|
|
(5,),
|
|
(5, 3),
|
|
(0, 3),
|
|
),
|
|
(
|
|
(5, 3),
|
|
(3,),
|
|
(0, 3),
|
|
),
|
|
(
|
|
(5,),
|
|
(4, 5, 3),
|
|
(-5, 5),
|
|
),
|
|
(
|
|
(4, 5, 3),
|
|
(3,),
|
|
(0, 5),
|
|
),
|
|
(
|
|
(5,),
|
|
(2, 4, 5, 3),
|
|
(0, 5),
|
|
),
|
|
(
|
|
(2, 4, 5, 3),
|
|
(3,),
|
|
(-1, 5),
|
|
),
|
|
(
|
|
(5, 4, 3),
|
|
(3, 2),
|
|
(0, 5),
|
|
),
|
|
(
|
|
(4, 3),
|
|
(5, 3, 2),
|
|
(0, 5),
|
|
),
|
|
(
|
|
(2, 5, 4, 3),
|
|
(3, 2),
|
|
(0, 5),
|
|
),
|
|
(
|
|
(5, 4, 3),
|
|
(1, 3, 2),
|
|
(0, 5),
|
|
),
|
|
(
|
|
(1, 4, 3),
|
|
(5, 3, 2),
|
|
(0, 5),
|
|
),
|
|
(
|
|
(5, 4, 3),
|
|
(2, 1, 3, 2),
|
|
(0, 5),
|
|
),
|
|
(
|
|
(2, 1, 4, 3),
|
|
(5, 3, 2),
|
|
(0, 5),
|
|
),
|
|
]
|
|
|
|
|
|
@pytest.mark.parametrize(
|
|
"lhs_shape,rhs_shape,bounds,clear_rhs",
|
|
[
|
|
(lhs_shape, rhs_shape, bounds, clear)
|
|
for lhs_shape, rhs_shape, bounds in test_matmul_shape_and_bounds
|
|
for clear in [False, True]
|
|
],
|
|
)
|
|
def test_matmul(lhs_shape, rhs_shape, bounds, clear_rhs, helpers):
|
|
"""
|
|
Test matmul.
|
|
"""
|
|
|
|
configuration = helpers.configuration()
|
|
|
|
minimum, maximum = bounds
|
|
|
|
@fhe.compiler({"x": "encrypted", "y": "clear"})
|
|
def clear(x, y):
|
|
return x @ y
|
|
|
|
@fhe.compiler({"x": "encrypted", "y": "encrypted"})
|
|
def encrypted(x, y):
|
|
return np.matmul(x, y)
|
|
|
|
implementation = clear if clear_rhs else encrypted
|
|
|
|
inputset = [
|
|
(
|
|
np.random.randint(minimum, maximum, size=lhs_shape),
|
|
np.random.randint(minimum, maximum, size=rhs_shape),
|
|
)
|
|
for _ in range(100)
|
|
]
|
|
circuit = implementation.compile(inputset, configuration)
|
|
|
|
sample = list(inputset[-1])
|
|
|
|
helpers.check_execution(circuit, implementation, sample, retries=3)
|
|
|
|
|
|
@pytest.mark.parametrize("bit_width", [4, 10])
|
|
@pytest.mark.parametrize("signed", [True, False])
|
|
def test_zero_matmul(bit_width, signed, helpers):
|
|
"""
|
|
Test matmul where one of the operators is all zeros.
|
|
"""
|
|
|
|
configuration = helpers.configuration()
|
|
|
|
lhs_shape = (2, 1)
|
|
rhs_shape = (1, 2)
|
|
|
|
bounds = (-(2 ** (bit_width - 1)), 2 ** (bit_width - 1) - 1) if signed else (0, 2**bit_width)
|
|
minimum, maximum = bounds
|
|
|
|
@fhe.compiler({"x": "encrypted", "y": "encrypted"})
|
|
def function(x, y):
|
|
return x * y
|
|
|
|
inputset = [
|
|
(
|
|
np.random.randint(minimum, maximum, size=lhs_shape),
|
|
np.zeros(rhs_shape, dtype=np.int64),
|
|
)
|
|
for i in range(100)
|
|
]
|
|
circuit = function.compile(inputset, configuration)
|
|
|
|
sample = [
|
|
np.random.randint(minimum, maximum, size=lhs_shape),
|
|
np.zeros(rhs_shape, dtype=np.int64),
|
|
]
|
|
|
|
helpers.check_execution(circuit, function, sample, retries=3)
|