github/concrete
1
1
Fork 0
mirror of https://github.com/zama-ai/concrete.git synced 2026-04-17 03:00:54 -04:00
Code Issues Packages Projects Releases Wiki Activity

fix(frontend-python): add missing error messages on high bit-width encrypted dot and matmul

Browse Source
This commit is contained in:
Umut
2023-05-16 14:09:58 +02:00
parent 52cc0c108a
commit 6d16d34b22
2 changed files with 112 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

57
frontends/concrete-python/concrete/fhe/mlir/context.py
View File

@@ -757,13 +757,39 @@ class Context:
def dot(self, resulting_type: ConversionType, x: Conversion, y: Conversion) -> Conversion:
if x.is_clear and y.is_clear:
highlights = {
highlights: Dict[Node, Union[str, List[str]]] = {
x.origin: "lhs is clear",
y.origin: "rhs is clear" if x.origin is not y.origin else "operand is clear",
self.converting: "but clear-clear dot products are not supported",
}
self.error(highlights)
if (x.is_encrypted and y.is_encrypted) and (
x.bit_width > MAXIMUM_TLU_BIT_WIDTH or y.bit_width > MAXIMUM_TLU_BIT_WIDTH
):
highlights = {
self.converting: [
f"but only up to {MAXIMUM_TLU_BIT_WIDTH}-bit "
f"encrypted dot products are supported"
],
}
for operand in [x, y]:
if operand.bit_width > MAXIMUM_TLU_BIT_WIDTH:
highlights[operand.origin] = [
f"this {operand.bit_width}-bit value "
f"is used as an operand to an encrypted dot products"
]
if operand.bit_width != operand.original_bit_width:
highlights[operand.origin].append( # type: ignore
"("
f"note that it's assigned {operand.bit_width}-bits "
f"during compilation because of its relation with other operations"
")"
)
self.error(highlights)
assert self.is_bit_width_compatible(resulting_type, x, y)
if x.is_scalar or y.is_scalar:
@@ -1298,19 +1324,44 @@ class Context:
def matmul(self, resulting_type: ConversionType, x: Conversion, y: Conversion) -> Conversion:
if x.is_clear and y.is_clear:
highlights = {
highlights: Dict[Node, Union[str, List[str]]] = {
x.origin: "lhs is clear",
y.origin: "rhs is clear" if x.origin is not y.origin else "operand is clear",
self.converting: "but clear-clear matrix multiplications are not supported",
}
self.error(highlights)
if (x.is_encrypted and y.is_encrypted) and (
x.bit_width > MAXIMUM_TLU_BIT_WIDTH or y.bit_width > MAXIMUM_TLU_BIT_WIDTH
):
highlights = {
self.converting: [
f"but only up to {MAXIMUM_TLU_BIT_WIDTH}-bit "
f"encrypted matrix multiplications are supported"
],
}
for operand in [x, y]:
if operand.bit_width > MAXIMUM_TLU_BIT_WIDTH:
highlights[operand.origin] = [
f"this {operand.bit_width}-bit value "
f"is used as an operand to an encrypted matrix multiplication"
]
if operand.bit_width != operand.original_bit_width:
highlights[operand.origin].append( # type: ignore
"("
f"note that it's assigned {operand.bit_width}-bits "
f"during compilation because of its relation with other operations"
")"
)
self.error(highlights)
assert self.is_bit_width_compatible(resulting_type, x, y)
if resulting_type.shape == ():
if x.is_clear:
x, y = y, x
operation = fhelinalg.DotEint if x.is_encrypted and y.is_encrypted else fhelinalg.Dot
elif x.is_encrypted and y.is_encrypted:
operation = fhelinalg.MatMulEintEintOp

58
frontends/concrete-python/tests/mlir/test_converter.py
View File

@@ -615,6 +615,64 @@ Function you are trying to compile cannot be compiled
(note that it's assigned 21-bits during compilation because of its relation with other operations)
%2 = multiply(%0, %1) # EncryptedScalar<uint21> ∈ [2000000, 2000000]
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ but only up to 16-bit encrypted multiplications are supported
return %2
""", # noqa: E501
),
pytest.param(
lambda x, y: np.dot(x, y),
{"x": "encrypted", "y": "encrypted"},
[
(
[100_000, 200_000],
[200_000, 100_000],
)
],
RuntimeError,
"""
Function you are trying to compile cannot be compiled
%0 = x # EncryptedTensor<uint18, shape=(2,)> ∈ [100000, 200000]
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ this 36-bit value is used as an operand to an encrypted dot products
(note that it's assigned 36-bits during compilation because of its relation with other operations)
%1 = y # EncryptedTensor<uint18, shape=(2,)> ∈ [100000, 200000]
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ this 36-bit value is used as an operand to an encrypted dot products
(note that it's assigned 36-bits during compilation because of its relation with other operations)
%2 = dot(%0, %1) # EncryptedScalar<uint36> ∈ [40000000000, 40000000000]
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ but only up to 16-bit encrypted dot products are supported
return %2
""", # noqa: E501
),
pytest.param(
lambda x, y: x @ y,
{"x": "encrypted", "y": "encrypted"},
[
(
[
[100_000, 200_000],
[200_000, 100_000],
],
[
[100_000, 200_000],
[200_000, 100_000],
],
)
],
RuntimeError,
"""
Function you are trying to compile cannot be compiled
%0 = x # EncryptedTensor<uint18, shape=(2, 2)> ∈ [100000, 200000]
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ this 36-bit value is used as an operand to an encrypted matrix multiplication
(note that it's assigned 36-bits during compilation because of its relation with other operations)
%1 = y # EncryptedTensor<uint18, shape=(2, 2)> ∈ [100000, 200000]
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ this 36-bit value is used as an operand to an encrypted matrix multiplication
(note that it's assigned 36-bits during compilation because of its relation with other operations)
%2 = matmul(%0, %1) # EncryptedTensor<uint36, shape=(2, 2)> ∈ [40000000000, 50000000000]
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ but only up to 16-bit encrypted matrix multiplications are supported
return %2
""", # noqa: E501