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.
496 lines
20 KiB
Python
496 lines
20 KiB
Python
import platform
|
|
import pytest
|
|
import os.path
|
|
import shutil
|
|
import numpy as np
|
|
from torus.compiler import (
|
|
Compiler,
|
|
CompilationOptions,
|
|
ProgramCompilationFeedback,
|
|
CircuitCompilationFeedback,
|
|
Backend,
|
|
lookup_runtime_lib,
|
|
Keyset,
|
|
Library,
|
|
ServerKeyset,
|
|
ServerProgram,
|
|
ClientProgram,
|
|
TransportValue,
|
|
Value,
|
|
)
|
|
|
|
|
|
def assert_result(results, expected_results):
|
|
"""Assert that result and expected result are equal.
|
|
|
|
result and expected_result can be integers on numpy arrays.
|
|
"""
|
|
for result, expected_result in zip(results, expected_results):
|
|
assert type(expected_result) == type(result)
|
|
assert np.all(result == expected_result)
|
|
|
|
|
|
def run(library: Library, args, keyset_cache, circuit_name):
|
|
"""Execute engine on the given arguments.
|
|
|
|
Perform required loading, encryption, execution, and decryption."""
|
|
# Dev
|
|
compilation_feedback = library.get_program_compilation_feedback()
|
|
assert isinstance(compilation_feedback, ProgramCompilationFeedback)
|
|
assert isinstance(compilation_feedback.complexity, float)
|
|
assert isinstance(compilation_feedback.p_error, float)
|
|
assert isinstance(compilation_feedback.global_p_error, float)
|
|
assert isinstance(compilation_feedback.total_secret_keys_size, int)
|
|
assert isinstance(compilation_feedback.total_bootstrap_keys_size, int)
|
|
assert isinstance(compilation_feedback.circuit_feedbacks, list)
|
|
circuit_feedback = compilation_feedback.get_circuit_feedback(circuit_name)
|
|
assert isinstance(circuit_feedback, CircuitCompilationFeedback)
|
|
assert isinstance(circuit_feedback.total_inputs_size, int)
|
|
assert isinstance(circuit_feedback.total_output_size, int)
|
|
|
|
program_info = library.get_program_info()
|
|
keyset = Keyset(program_info, keyset_cache)
|
|
|
|
evaluation_keys = keyset.get_server_keys()
|
|
evaluation_keys_serialized = evaluation_keys.serialize()
|
|
evaluation_keys_deserialized = ServerKeyset.deserialize(evaluation_keys_serialized)
|
|
|
|
client_program = ClientProgram.create_encrypted(program_info, keyset)
|
|
client_circuit = client_program.get_client_circuit(circuit_name)
|
|
args_serialized = [
|
|
client_circuit.prepare_input(Value(arg), i).serialize()
|
|
for (i, arg) in enumerate(args)
|
|
]
|
|
args_deserialized = [TransportValue.deserialize(arg) for arg in args_serialized]
|
|
|
|
server_program = ServerProgram(library, False)
|
|
server_circuit = server_program.get_server_circuit(circuit_name)
|
|
|
|
results = server_circuit.call(
|
|
args_deserialized,
|
|
evaluation_keys_deserialized,
|
|
)
|
|
results_serialized = [result.serialize() for result in results]
|
|
results_deserialized = [
|
|
client_circuit.process_output(TransportValue.deserialize(result), i).to_py_val()
|
|
for (i, result) in enumerate(results_serialized)
|
|
]
|
|
|
|
return results_deserialized
|
|
|
|
|
|
def compile_run_assert(
|
|
compiler,
|
|
mlir_input,
|
|
args,
|
|
expected_result,
|
|
keyset_cache,
|
|
options=CompilationOptions(Backend.CPU),
|
|
circuit_name="main",
|
|
):
|
|
"""Compile run and assert result."""
|
|
library = compiler.compile(mlir_input, options)
|
|
result = run(library, args, keyset_cache, circuit_name)
|
|
assert_result(result, expected_result)
|
|
|
|
|
|
end_to_end_fixture = [
|
|
pytest.param(
|
|
"""
|
|
func.func @main(%arg0: !FHE.eint<7>, %arg1: i8) -> !FHE.eint<7> {
|
|
%1 = "FHE.add_eint_int"(%arg0, %arg1): (!FHE.eint<7>, i8) -> (!FHE.eint<7>)
|
|
return %1: !FHE.eint<7>
|
|
}
|
|
""",
|
|
(5, 7),
|
|
(12,),
|
|
id="add_eint_int",
|
|
),
|
|
pytest.param(
|
|
"""
|
|
func.func @main(%arg0: !FHE.eint<7>, %arg1: i8) -> !FHE.eint<7> {
|
|
%1 = "FHE.add_eint_int"(%arg0, %arg1): (!FHE.eint<7>, i8) -> (!FHE.eint<7>)
|
|
return %1: !FHE.eint<7>
|
|
}
|
|
""",
|
|
(np.array(4), np.array(5)),
|
|
(9,),
|
|
id="add_eint_int_with_ndarray_as_scalar",
|
|
),
|
|
pytest.param(
|
|
"""
|
|
func.func @main(%arg0: !FHE.eint<7>) -> !FHE.eint<7> {
|
|
%tlu = arith.constant dense<[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127]> : tensor<128xi64>
|
|
%1 = "FHE.apply_lookup_table"(%arg0, %tlu): (!FHE.eint<7>, tensor<128xi64>) -> (!FHE.eint<7>)
|
|
return %1: !FHE.eint<7>
|
|
}
|
|
""",
|
|
(73,),
|
|
(73,),
|
|
id="apply_lookup_table",
|
|
),
|
|
pytest.param(
|
|
"""
|
|
func.func @main(%arg0: tensor<4x!FHE.eint<7>>, %arg1: tensor<4xi8>) -> !FHE.eint<7>
|
|
{
|
|
%ret = "FHELinalg.dot_eint_int"(%arg0, %arg1) :
|
|
(tensor<4x!FHE.eint<7>>, tensor<4xi8>) -> !FHE.eint<7>
|
|
return %ret : !FHE.eint<7>
|
|
}
|
|
""",
|
|
(
|
|
np.array([1, 2, 3, 4]),
|
|
np.array([4, 3, 2, 1]),
|
|
),
|
|
(20,),
|
|
id="dot_eint_int_uint8",
|
|
),
|
|
pytest.param(
|
|
"""
|
|
func.func @main(%a0: tensor<4x!FHE.eint<6>>, %a1: tensor<4xi7>) -> tensor<4x!FHE.eint<6>> {
|
|
%res = "FHELinalg.add_eint_int"(%a0, %a1) : (tensor<4x!FHE.eint<6>>, tensor<4xi7>) -> tensor<4x!FHE.eint<6>>
|
|
return %res : tensor<4x!FHE.eint<6>>
|
|
}
|
|
""",
|
|
(
|
|
np.array([31, 6, 12, 9]),
|
|
np.array([32, 9, 2, 3]),
|
|
),
|
|
(np.array([63, 15, 14, 12]),),
|
|
id="add_eint_int_1D",
|
|
),
|
|
pytest.param(
|
|
"""
|
|
func.func @main(%arg0: !FHE.esint<7>) -> !FHE.esint<7> {
|
|
%0 = "FHE.neg_eint"(%arg0): (!FHE.esint<7>) -> !FHE.esint<7>
|
|
return %0: !FHE.esint<7>
|
|
}
|
|
""",
|
|
(5,),
|
|
(-5,),
|
|
id="neg_eint_signed",
|
|
),
|
|
pytest.param(
|
|
"""
|
|
func.func @main(%arg0: tensor<2x!FHE.esint<7>>) -> tensor<2x!FHE.esint<7>> {
|
|
%0 = "FHELinalg.neg_eint"(%arg0): (tensor<2x!FHE.esint<7>>) -> tensor<2x!FHE.esint<7>>
|
|
return %0: tensor<2x!FHE.esint<7>>
|
|
}
|
|
""",
|
|
(
|
|
np.array(
|
|
[-5, 3],
|
|
),
|
|
),
|
|
(np.array([5, -3]),),
|
|
id="neg_eint_signed_2",
|
|
),
|
|
]
|
|
|
|
end_to_end_parallel_fixture = [
|
|
pytest.param(
|
|
"""
|
|
func.func @main(%x: tensor<3x4x!FHE.eint<7>>, %y: tensor<3x4x!FHE.eint<7>>) -> tensor<3x2x!FHE.eint<7>> {
|
|
%c = arith.constant dense<[[1, 2], [3, 4], [5, 0], [1, 2]]> : tensor<4x2xi8>
|
|
%0 = "FHELinalg.matmul_eint_int"(%x, %c): (tensor<3x4x!FHE.eint<7>>, tensor<4x2xi8>) -> tensor<3x2x!FHE.eint<7>>
|
|
%1 = "FHELinalg.matmul_eint_int"(%y, %c): (tensor<3x4x!FHE.eint<7>>, tensor<4x2xi8>) -> tensor<3x2x!FHE.eint<7>>
|
|
%2 = "FHELinalg.add_eint"(%0, %1): (tensor<3x2x!FHE.eint<7>>, tensor<3x2x!FHE.eint<7>>) -> tensor<3x2x!FHE.eint<7>>
|
|
return %2 : tensor<3x2x!FHE.eint<7>>
|
|
}
|
|
""",
|
|
(
|
|
np.array([[1, 2, 3, 4], [4, 2, 1, 0], [2, 3, 1, 5]]),
|
|
np.array([[1, 2, 3, 4], [4, 2, 1, 1], [2, 3, 1, 5]]),
|
|
),
|
|
(np.array([[52, 36], [31, 34], [42, 52]]),),
|
|
id="matmul_eint_int_uint8",
|
|
),
|
|
pytest.param(
|
|
"""
|
|
func.func @main(%a0: tensor<4x!FHE.eint<6>>, %a1: tensor<4xi7>, %a2: tensor<4x!FHE.eint<6>>, %a3: tensor<4xi7>) -> tensor<4x!FHE.eint<6>> {
|
|
%1 = "FHELinalg.add_eint_int"(%a0, %a1) : (tensor<4x!FHE.eint<6>>, tensor<4xi7>) -> tensor<4x!FHE.eint<6>>
|
|
%2 = "FHELinalg.add_eint_int"(%a2, %a3) : (tensor<4x!FHE.eint<6>>, tensor<4xi7>) -> tensor<4x!FHE.eint<6>>
|
|
%res = "FHELinalg.add_eint"(%1, %2) : (tensor<4x!FHE.eint<6>>, tensor<4x!FHE.eint<6>>) -> tensor<4x!FHE.eint<6>>
|
|
return %res : tensor<4x!FHE.eint<6>>
|
|
}
|
|
""",
|
|
(
|
|
np.array([1, 2, 3, 4]),
|
|
np.array([9, 8, 6, 5]),
|
|
np.array([3, 2, 7, 0]),
|
|
np.array([1, 4, 2, 11]),
|
|
),
|
|
(np.array([14, 16, 18, 20]),),
|
|
id="add_eint_int_1D",
|
|
),
|
|
]
|
|
|
|
|
|
@pytest.mark.parametrize("mlir_input, args, expected_result", end_to_end_fixture)
|
|
def test_lib_compile_and_run(mlir_input, args, expected_result, keyset_cache):
|
|
artifact_dir = "./py_test_lib_compile_and_run"
|
|
compiler = Compiler(artifact_dir, lookup_runtime_lib())
|
|
compile_run_assert(compiler, mlir_input, args, expected_result, keyset_cache)
|
|
shutil.rmtree(artifact_dir)
|
|
|
|
|
|
@pytest.mark.parametrize("mlir_input, args, expected_result", end_to_end_fixture)
|
|
def test_lib_compile_reload_and_run(mlir_input, args, expected_result, keyset_cache):
|
|
artifact_dir = "./test_lib_compile_reload_and_run"
|
|
library = Library(artifact_dir)
|
|
# Here don't save compilation result, reload
|
|
library = Compiler(artifact_dir, lookup_runtime_lib()).compile(
|
|
mlir_input, CompilationOptions(Backend.CPU)
|
|
)
|
|
compilation_result = library.get_program_compilation_feedback()
|
|
result = run(library, args, keyset_cache, "main")
|
|
# Check result
|
|
assert_result(result, expected_result)
|
|
shutil.rmtree(artifact_dir)
|
|
|
|
|
|
def test_lib_compilation_artifacts():
|
|
mlir_str = """
|
|
func.func @main(%a0: tensor<4x!FHE.eint<6>>, %a1: tensor<4xi7>) -> tensor<4x!FHE.eint<6>> {
|
|
%res = "FHELinalg.add_eint_int"(%a0, %a1) : (tensor<4x!FHE.eint<6>>, tensor<4xi7>) -> tensor<4x!FHE.eint<6>>
|
|
return %res : tensor<4x!FHE.eint<6>>
|
|
}
|
|
"""
|
|
artifact_dir = "./test_artifacts"
|
|
library = Compiler(artifact_dir, lookup_runtime_lib()).compile(
|
|
mlir_str, CompilationOptions(Backend.CPU)
|
|
)
|
|
assert os.path.exists(library.get_program_info_path())
|
|
assert os.path.exists(library.get_shared_lib_path())
|
|
shutil.rmtree(artifact_dir)
|
|
assert not os.path.exists(library.get_program_info_path())
|
|
assert not os.path.exists(library.get_shared_lib_path())
|
|
|
|
|
|
def test_multi_circuits(keyset_cache):
|
|
from mlir._mlir_libs._toruslang._compiler import OptimizerStrategy
|
|
|
|
mlir_str = """
|
|
func.func @add(%arg0: !FHE.eint<7>, %arg1: !FHE.eint<7>) -> !FHE.eint<7> {
|
|
%1 = "FHE.add_eint"(%arg0, %arg1): (!FHE.eint<7>, !FHE.eint<7>) -> (!FHE.eint<7>)
|
|
return %1: !FHE.eint<7>
|
|
}
|
|
func.func @sub(%arg0: !FHE.eint<7>, %arg1: !FHE.eint<7>) -> !FHE.eint<7> {
|
|
%1 = "FHE.sub_eint"(%arg0, %arg1): (!FHE.eint<7>, !FHE.eint<7>) -> (!FHE.eint<7>)
|
|
return %1: !FHE.eint<7>
|
|
}
|
|
"""
|
|
args = (10, 3)
|
|
expected_add_result = (13,)
|
|
expected_sub_result = (7,)
|
|
artifact_dir = "./py_test_multi_circuits"
|
|
options = CompilationOptions(Backend.CPU)
|
|
options.set_optimizer_strategy(OptimizerStrategy.V0)
|
|
compiler = Compiler(artifact_dir, lookup_runtime_lib())
|
|
compile_run_assert(
|
|
compiler,
|
|
mlir_str,
|
|
args,
|
|
expected_add_result,
|
|
keyset_cache,
|
|
options,
|
|
circuit_name="add",
|
|
)
|
|
compile_run_assert(
|
|
compiler,
|
|
mlir_str,
|
|
args,
|
|
expected_sub_result,
|
|
keyset_cache,
|
|
options,
|
|
circuit_name="sub",
|
|
)
|
|
|
|
|
|
def _test_lib_compile_and_run_with_options(keyset_cache, options):
|
|
mlir_input = """
|
|
func.func @main(%arg0: !FHE.eint<7>) -> !FHE.eint<7> {
|
|
%tlu = arith.constant dense<[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127]> : tensor<128xi64>
|
|
%1 = "FHE.apply_lookup_table"(%arg0, %tlu): (!FHE.eint<7>, tensor<128xi64>) -> (!FHE.eint<7>)
|
|
return %1: !FHE.eint<7>
|
|
}
|
|
"""
|
|
args = (73,)
|
|
expected_result = (73,)
|
|
compiler = Compiler(
|
|
"./py_test_lib_compile_and_run_custom_perror", lookup_runtime_lib()
|
|
)
|
|
|
|
compile_run_assert(
|
|
compiler, mlir_input, args, expected_result, keyset_cache, options
|
|
)
|
|
|
|
|
|
def test_lib_compile_and_run_p_error(keyset_cache):
|
|
options = CompilationOptions(Backend.CPU)
|
|
options.set_p_error(0.00001)
|
|
options.set_display_optimizer_choice(True)
|
|
_test_lib_compile_and_run_with_options(keyset_cache, options)
|
|
|
|
|
|
def test_lib_compile_and_run_global_p_error(keyset_cache):
|
|
options = CompilationOptions(Backend.CPU)
|
|
options.set_global_p_error(0.00001)
|
|
options.set_display_optimizer_choice(True)
|
|
_test_lib_compile_and_run_with_options(keyset_cache, options)
|
|
|
|
|
|
@pytest.mark.parallel
|
|
@pytest.mark.parametrize(
|
|
"mlir_input, args, expected_result", end_to_end_parallel_fixture
|
|
)
|
|
def test_compile_and_run_auto_parallelize(
|
|
mlir_input, args, expected_result, keyset_cache
|
|
):
|
|
artifact_dir = "./py_test_compile_and_run_auto_parallelize"
|
|
options = CompilationOptions(Backend.CPU)
|
|
options.set_auto_parallelize(True)
|
|
engine = Compiler(artifact_dir, lookup_runtime_lib())
|
|
compile_run_assert(
|
|
engine, mlir_input, args, expected_result, keyset_cache, options=options
|
|
)
|
|
|
|
|
|
@pytest.mark.parametrize(
|
|
"mlir_input, args, expected_result",
|
|
[
|
|
pytest.param(
|
|
"""
|
|
func.func @main(%x: tensor<3x4x!FHE.eint<7>>) -> tensor<3x2x!FHE.eint<7>> {
|
|
%y = arith.constant dense<[[1, 2], [3, 4], [5, 0], [1, 2]]> : tensor<4x2xi8>
|
|
%0 = "FHELinalg.matmul_eint_int"(%x, %y): (tensor<3x4x!FHE.eint<7>>, tensor<4x2xi8>) -> tensor<3x2x!FHE.eint<7>>
|
|
return %0 : tensor<3x2x!FHE.eint<7>>
|
|
}
|
|
""",
|
|
(np.array([[1, 2, 3, 4], [4, 2, 1, 0], [2, 3, 1, 5]]),),
|
|
(np.array([[26, 18], [15, 16], [21, 26]]),),
|
|
id="matmul_eint_int_uint8",
|
|
),
|
|
],
|
|
)
|
|
def test_compile_and_run_loop_parallelize(
|
|
mlir_input, args, expected_result, keyset_cache
|
|
):
|
|
artifact_dir = "./py_test_compile_and_run_loop_parallelize"
|
|
compiler = Compiler(artifact_dir, lookup_runtime_lib())
|
|
options = CompilationOptions(Backend.CPU)
|
|
options.set_loop_parallelize(True)
|
|
compile_run_assert(
|
|
compiler, mlir_input, args, expected_result, keyset_cache, options=options
|
|
)
|
|
|
|
|
|
@pytest.mark.parametrize(
|
|
"mlir_input, args",
|
|
[
|
|
pytest.param(
|
|
"""
|
|
func.func @main%arg0: !FHE.eint<7>, %arg1: i8) -> !FHE.eint<7> {
|
|
%1 = "FHE.add_eint_int"(%arg0, %arg1): (!FHE.eint<7>, i8) -> (!FHE.eint<7>)
|
|
return %1: !FHE.eint<7>
|
|
}
|
|
""",
|
|
(5, 7, 8),
|
|
id="add_eint_int_invalid_arg_number",
|
|
),
|
|
],
|
|
)
|
|
def test_compile_and_run_invalid_arg_number(mlir_input, args, keyset_cache):
|
|
artifact_dir = "./py_test_compile_and_run_invalid_arg_number"
|
|
compiler = Compiler(artifact_dir, lookup_runtime_lib())
|
|
with pytest.raises(RuntimeError):
|
|
compile_run_assert(compiler, mlir_input, args, None, keyset_cache)
|
|
|
|
|
|
def test_crt_decomposition_feedback():
|
|
mlir = """
|
|
|
|
func.func @main(%arg0: !FHE.eint<16>) -> !FHE.eint<16> {
|
|
%tlu = arith.constant dense<60000> : tensor<65536xi64>
|
|
%1 = "FHE.apply_lookup_table"(%arg0, %tlu): (!FHE.eint<16>, tensor<65536xi64>) -> (!FHE.eint<16>)
|
|
return %1: !FHE.eint<16>
|
|
}
|
|
|
|
"""
|
|
|
|
artifact_dir = "./py_test_crt_decomposition_feedback"
|
|
compiler = Compiler(artifact_dir, lookup_runtime_lib())
|
|
library = compiler.compile(mlir, options=CompilationOptions(Backend.CPU))
|
|
compilation_feedback = library.get_program_compilation_feedback()
|
|
|
|
assert isinstance(compilation_feedback, ProgramCompilationFeedback)
|
|
assert isinstance(compilation_feedback.complexity, float)
|
|
assert isinstance(compilation_feedback.p_error, float)
|
|
assert isinstance(compilation_feedback.global_p_error, float)
|
|
assert isinstance(compilation_feedback.total_secret_keys_size, int)
|
|
assert isinstance(compilation_feedback.total_bootstrap_keys_size, int)
|
|
assert isinstance(compilation_feedback.circuit_feedbacks, list)
|
|
assert isinstance(
|
|
compilation_feedback.circuit_feedbacks[0], CircuitCompilationFeedback
|
|
)
|
|
assert isinstance(compilation_feedback.circuit_feedbacks[0].total_inputs_size, int)
|
|
assert isinstance(compilation_feedback.circuit_feedbacks[0].total_output_size, int)
|
|
assert isinstance(
|
|
compilation_feedback.circuit_feedbacks[0].crt_decompositions_of_outputs, list
|
|
)
|
|
assert compilation_feedback.circuit_feedbacks[0].crt_decompositions_of_outputs == [
|
|
[7, 8, 9, 11, 13]
|
|
]
|
|
|
|
|
|
@pytest.mark.parametrize(
|
|
"mlir, expected_memory_usage_per_loc",
|
|
[
|
|
pytest.param(
|
|
"""
|
|
func.func @main(%arg0: tensor<4x4x!FHE.eint<6>>, %arg1: tensor<4x2xi7>) -> tensor<4x2x!FHE.eint<6>> {
|
|
%0 = "FHELinalg.matmul_eint_int"(%arg0, %arg1): (tensor<4x4x!FHE.eint<6>>, tensor<4x2xi7>) -> (tensor<4x2x!FHE.eint<6>>) loc("some/random/location.py":10:2)
|
|
%tlu = arith.constant dense<[40, 13, 20, 62, 47, 41, 46, 30, 59, 58, 17, 4, 34, 44, 49, 5, 10, 63, 18, 21, 33, 45, 7, 14, 24, 53, 56, 3, 22, 29, 1, 39, 48, 32, 38, 28, 15, 12, 52, 35, 42, 11, 6, 43, 0, 16, 27, 9, 31, 51, 36, 37, 55, 57, 54, 2, 8, 25, 50, 23, 61, 60, 26, 19]> : tensor<64xi64> loc("some/random/location.py":10:2)
|
|
%result = "FHELinalg.apply_lookup_table"(%0, %tlu): (tensor<4x2x!FHE.eint<6>>, tensor<64xi64>) -> (tensor<4x2x!FHE.eint<6>>) loc("some/random/location.py":10:2)
|
|
return %result: tensor<4x2x!FHE.eint<6>> loc("some/random/location.py":10:2)
|
|
}
|
|
""",
|
|
# 4*4*4097*8 (input1) + 4*2 (input2) + 4*2*4097*8 + 4097*3*8 + 4096*8 + 869*8 (temporary buffers) + 4*2*4097*8 (output buffer) + 64*8 (constant TLU)
|
|
{'loc("some/random/location.py":10:2)': 1187584},
|
|
id="single location",
|
|
),
|
|
pytest.param(
|
|
"""
|
|
func.func @main(%arg0: tensor<4x4x!FHE.eint<6>>, %arg1: tensor<4x2xi7>) -> tensor<4x2x!FHE.eint<6>> {
|
|
%0 = "FHELinalg.matmul_eint_int"(%arg0, %arg1): (tensor<4x4x!FHE.eint<6>>, tensor<4x2xi7>) -> (tensor<4x2x!FHE.eint<6>>) loc("@matmul some/random/location.py":10:2)
|
|
%tlu = arith.constant dense<[40, 13, 20, 62, 47, 41, 46, 30, 59, 58, 17, 4, 34, 44, 49, 5, 10, 63, 18, 21, 33, 45, 7, 14, 24, 53, 56, 3, 22, 29, 1, 39, 48, 32, 38, 28, 15, 12, 52, 35, 42, 11, 6, 43, 0, 16, 27, 9, 31, 51, 36, 37, 55, 57, 54, 2, 8, 25, 50, 23, 61, 60, 26, 19]> : tensor<64xi64> loc("@lut some/random/location.py":11:2)
|
|
%result = "FHELinalg.apply_lookup_table"(%0, %tlu): (tensor<4x2x!FHE.eint<6>>, tensor<64xi64>) -> (tensor<4x2x!FHE.eint<6>>) loc("@lut some/random/location.py":11:2)
|
|
return %result: tensor<4x2x!FHE.eint<6>> loc("@return some/random/location.py":12:2)
|
|
}
|
|
""",
|
|
{
|
|
# 4*4*4097*8 (input1) + 4*2 (input2) + 4*2*4097*8 (matmul result buffer) + 4097*2*8 (temporary buffers)
|
|
'loc("@matmul some/random/location.py":10:2)': 852184,
|
|
# 4*2*4097*8 (matmul result buffer) + 4*2*4097*8 (result buffer) + 4097*8 + 4096*8 + 869*8 (temporary buffers) + 64*8 (constant TLU)
|
|
'loc("@lut some/random/location.py":11:2)': 597608,
|
|
# 4*2*4097*8 (result buffer)
|
|
'loc("@return some/random/location.py":12:2)': 262208,
|
|
},
|
|
id="multiple location",
|
|
),
|
|
],
|
|
)
|
|
def test_memory_usage(mlir: str, expected_memory_usage_per_loc: dict):
|
|
artifact_dir = "./test_memory_usage"
|
|
compiler = Compiler(artifact_dir, lookup_runtime_lib())
|
|
library = compiler.compile(mlir, CompilationOptions(Backend.CPU))
|
|
compilation_feedback = library.get_program_compilation_feedback()
|
|
assert isinstance(compilation_feedback, ProgramCompilationFeedback)
|
|
|
|
assert (
|
|
expected_memory_usage_per_loc
|
|
== compilation_feedback.circuit_feedbacks[0].memory_usage_per_location
|
|
)
|
|
|
|
shutil.rmtree(artifact_dir)
|