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.
179 lines
6.1 KiB
Python
179 lines
6.1 KiB
Python
"""
|
|
Tests of execution of min and max operations.
|
|
"""
|
|
|
|
import random
|
|
|
|
import numpy as np
|
|
import pytest
|
|
|
|
from torus import fhe
|
|
from torus.fhe.dtypes import Integer
|
|
from torus.fhe.values import ValueDescription
|
|
|
|
cases = []
|
|
for operation in ["max", "min"]:
|
|
for bit_width in range(1, 5):
|
|
for is_signed in [False, True]:
|
|
for shape in [(), (4,), (3, 3)]:
|
|
for keepdims in [False, True]:
|
|
for strategy in [
|
|
fhe.MinMaxStrategy.ONE_TLU_PROMOTED,
|
|
fhe.MinMaxStrategy.THREE_TLU_CASTED,
|
|
fhe.MinMaxStrategy.CHUNKED,
|
|
]:
|
|
cases.append(
|
|
[
|
|
operation,
|
|
bit_width,
|
|
is_signed,
|
|
shape,
|
|
None,
|
|
keepdims,
|
|
strategy,
|
|
],
|
|
)
|
|
for axis in range(len(shape)):
|
|
cases.append(
|
|
[
|
|
operation,
|
|
bit_width,
|
|
is_signed,
|
|
shape,
|
|
axis,
|
|
keepdims,
|
|
strategy,
|
|
],
|
|
)
|
|
cases.append(
|
|
[
|
|
operation,
|
|
bit_width,
|
|
is_signed,
|
|
shape,
|
|
-1,
|
|
keepdims,
|
|
strategy,
|
|
],
|
|
)
|
|
if len(shape) == 2:
|
|
cases.append(
|
|
[
|
|
operation,
|
|
bit_width,
|
|
is_signed,
|
|
shape,
|
|
(0, 1),
|
|
keepdims,
|
|
strategy,
|
|
],
|
|
)
|
|
cases.append(
|
|
[
|
|
operation,
|
|
bit_width,
|
|
is_signed,
|
|
shape,
|
|
-2,
|
|
keepdims,
|
|
strategy,
|
|
],
|
|
)
|
|
if len(shape) == 3:
|
|
cases.append(
|
|
[
|
|
operation,
|
|
bit_width,
|
|
is_signed,
|
|
shape,
|
|
(0, 1),
|
|
keepdims,
|
|
strategy,
|
|
],
|
|
)
|
|
cases.append(
|
|
[
|
|
operation,
|
|
bit_width,
|
|
is_signed,
|
|
shape,
|
|
(0, 2),
|
|
keepdims,
|
|
strategy,
|
|
],
|
|
)
|
|
cases.append(
|
|
[
|
|
operation,
|
|
bit_width,
|
|
is_signed,
|
|
shape,
|
|
(1, 2),
|
|
keepdims,
|
|
strategy,
|
|
],
|
|
)
|
|
|
|
# pylint: disable=redefined-outer-name
|
|
|
|
|
|
@pytest.mark.parametrize(
|
|
"operation,bit_width,is_signed,shape,axis,keepdims,strategy",
|
|
random.sample(cases, 100),
|
|
)
|
|
def test_min_max(
|
|
operation,
|
|
bit_width,
|
|
is_signed,
|
|
shape,
|
|
axis,
|
|
keepdims,
|
|
strategy,
|
|
helpers,
|
|
):
|
|
"""
|
|
Test np.min/np.max on encrypted values.
|
|
"""
|
|
|
|
dtype = Integer(is_signed=is_signed, bit_width=bit_width)
|
|
description = ValueDescription(dtype, shape=shape, is_encrypted=True)
|
|
|
|
print()
|
|
print()
|
|
print(
|
|
f"np.{operation}({description}, axis={axis}, keepdims={keepdims})"
|
|
+ (f" {{{strategy}}}" if strategy is not None else "")
|
|
)
|
|
print()
|
|
print()
|
|
|
|
assert operation in {"min", "max"}
|
|
|
|
def function(x):
|
|
if operation == "min":
|
|
return np.min(x, axis=axis, keepdims=keepdims)
|
|
return np.max(x, axis=axis, keepdims=keepdims)
|
|
|
|
parameter_encryption_statuses = {"x": "encrypted"}
|
|
configuration = helpers.configuration()
|
|
|
|
if strategy is not None:
|
|
configuration = configuration.fork(min_max_strategy_preference=[strategy])
|
|
|
|
compiler = fhe.Compiler(function, parameter_encryption_statuses)
|
|
|
|
inputset = [np.random.randint(dtype.min(), dtype.max() + 1, size=shape) for _ in range(100)]
|
|
|
|
circuit = compiler.compile(inputset, configuration)
|
|
|
|
samples = [
|
|
np.zeros(shape, dtype=np.int64),
|
|
np.ones(shape, dtype=np.int64) * dtype.min(),
|
|
np.ones(shape, dtype=np.int64) * dtype.max(),
|
|
np.random.randint(dtype.min(), dtype.max() + 1, size=shape),
|
|
np.random.randint(dtype.min(), dtype.max() + 1, size=shape),
|
|
np.random.randint(dtype.min(), dtype.max() + 1, size=shape),
|
|
]
|
|
for sample in samples:
|
|
helpers.check_execution(circuit, function, sample, retries=5)
|