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.
102 lines
2.8 KiB
Python
102 lines
2.8 KiB
Python
"""
|
|
Tests execution of simulation.
|
|
"""
|
|
|
|
import numpy as np
|
|
import pytest
|
|
|
|
from torus import fhe
|
|
|
|
|
|
@pytest.mark.parametrize("p_error", [0.5, 0.01, 0.005])
|
|
def test_tlu_error_rate(p_error):
|
|
"""
|
|
Test that error rates are realistic.
|
|
"""
|
|
repetition = max(1000, int(10 / p_error))
|
|
assert repetition < 10_0000, "Test will be too long"
|
|
conf = fhe.Configuration(p_error=p_error)
|
|
|
|
precision = 3
|
|
|
|
@fhe.compiler({"x": "encrypted"})
|
|
def tlu(x):
|
|
return fhe.univariate(lambda x: x)(x)
|
|
|
|
inputset = [np.array(list(range(0, 2**precision)))]
|
|
|
|
circuit = tlu.compile(inputset, conf)
|
|
|
|
nok_fhe = 0
|
|
nok_sim = 0
|
|
size = 0
|
|
for t in inputset * repetition:
|
|
fhe_value = circuit.encrypt_run_decrypt(t)
|
|
sim_value = circuit.simulate(t)
|
|
for v, fhe_v, sim_v in zip(t, fhe_value, sim_value):
|
|
size += 1
|
|
nok_fhe += v != fhe_v
|
|
nok_sim += v != sim_v
|
|
|
|
fhe_p_error = nok_fhe / size
|
|
sim_p_error = nok_sim / size
|
|
|
|
assert p_error * size > 10 # test invalid if no more than 10 expected errors
|
|
assert fhe_p_error
|
|
assert sim_p_error
|
|
# the sample size is not big => we accept significant releative error
|
|
assert fhe_p_error == pytest.approx(p_error, rel=0.75)
|
|
assert sim_p_error == pytest.approx(p_error, rel=0.75)
|
|
assert fhe_p_error == pytest.approx(sim_p_error, rel=0.75)
|
|
|
|
|
|
@pytest.mark.parametrize("p_error", [0.01, 1e-120])
|
|
def test_approx_tlu_error_rate(p_error):
|
|
"""
|
|
Test that error rates are realistic in approximate mode.
|
|
"""
|
|
conf = fhe.Configuration(p_error=p_error)
|
|
precision = 8
|
|
lsbs_to_remove = 5
|
|
|
|
def rnd(x):
|
|
return (x + 2 ** (lsbs_to_remove - 1)) // 2**lsbs_to_remove
|
|
|
|
@fhe.compiler({"x": "encrypted"})
|
|
def tlu(x):
|
|
rounded = fhe.round_bit_pattern(
|
|
x,
|
|
lsbs_to_remove=lsbs_to_remove,
|
|
overflow_protection=False,
|
|
exactness=fhe.Exactness.APPROXIMATE,
|
|
)
|
|
return fhe.univariate(rnd)(rounded)
|
|
|
|
inputset = [np.array(list(range(0, 2**precision - 2**lsbs_to_remove)))]
|
|
|
|
circuit = tlu.compile(inputset, conf)
|
|
|
|
nok_fhe = 0
|
|
nok_sim = 0
|
|
size = 0
|
|
for t in inputset * 30:
|
|
fhe_values = circuit.encrypt_run_decrypt(t)
|
|
sim_values = circuit.simulate(t)
|
|
for x, fhe_v, sim_v in zip(t, fhe_values, sim_values):
|
|
r = rnd(x)
|
|
size += 1
|
|
nok_fhe += r != fhe_v
|
|
nok_sim += r != sim_v
|
|
print(r, fhe_v, sim_v)
|
|
|
|
fhe_p_error = nok_fhe / size
|
|
sim_p_error = nok_sim / size
|
|
|
|
# error rate is always significant (>3%) in approximate mode
|
|
assert p_error + 0.005 < fhe_p_error
|
|
assert p_error + 0.005 < sim_p_error
|
|
assert fhe_p_error < 0.2 + p_error
|
|
assert sim_p_error < 0.2 + p_error
|
|
|
|
assert fhe_p_error == pytest.approx(sim_p_error, rel=0.75)
|