mirror of
https://github.com/zama-ai/concrete.git
synced 2026-02-09 03:55:04 -05:00
feat(compiler): HLFHELinalg.mul_eint_int
This commit is contained in:
Quentin Bourgerie
committed by
Andi Drebes
Andi Drebes
parent
64e327209f
commit
f72d51d98d
@@ -175,4 +175,52 @@ def SubIntEintOp : HLFHELinalg_Op<"sub_int_eint", [TensorBroadcastingRules, Tens
|
||||
];
|
||||
}
|
||||
|
||||
def MulEintIntOp : HLFHELinalg_Op<"mul_eint_int", [TensorBroadcastingRules, TensorBinaryEintInt]> {
|
||||
let summary = "Returns a tensor that contains the multiplication of a tensor of encrypted integers and a tensor of clear integers.";
|
||||
|
||||
let description = [{
|
||||
Performs a multiplication following the broadcasting rules between a tensor of encrypted integers and a tensor of clear integers.
|
||||
The width of the clear integers should be less or equals than the witdh of encrypted integers.
|
||||
|
||||
Examples:
|
||||
```mlir
|
||||
// Returns the term to term multiplication of `%a0` with `%a1`
|
||||
"HLFHELinalg.mul_eint_int"(%a0, %a1) : (tensor<4x!HLFHE.eint<4>>, tensor<4xi5>) -> tensor<4x!HLFHE.eint<4>>
|
||||
|
||||
// Returns the term to term multiplication of `%a0` with `%a1`, where dimensions equal to one are stretched.
|
||||
"HLFHELinalg.mul_eint_int"(%a0, %a1) : (tensor<4x1x4x!HLFHE.eint<4>>, tensor<1x4x4xi5>) -> tensor<4x4x4x!HLFHE.eint<4>>
|
||||
|
||||
// Returns the multiplication of a 3x3 matrix of encrypted integers and a 3x1 matrix (a column) of integers.
|
||||
//
|
||||
// [1,2,3] [1] [1,2,3]
|
||||
// [4,5,6] * [2] = [8,10,18]
|
||||
// [7,8,9] [3] [21,24,27]
|
||||
//
|
||||
// The dimension #1 of operand #2 is stretched as it is equals to 1.
|
||||
"HLFHELinalg.mul_eint_int(%a0, %a1)" : (tensor<3x4x!HLFHE.eint<4>>, tensor<3x1xi5>) -> tensor<3x3x!HLFHE.eint<4>>
|
||||
|
||||
// Returns the multiplication of a 3x3 matrix of encrypted integers and a 1x3 matrix (a line) of integers.
|
||||
//
|
||||
// [1,2,3] [2,4,6]
|
||||
// [4,5,6] * [1,2,3] = [5,7,9]
|
||||
// [7,8,9] [8,10,12]
|
||||
//
|
||||
// The dimension #2 of operand #2 is stretched as it is equals to 1.
|
||||
"HLFHELinalg.mul_eint_int(%a0, %a1)" : (tensor<3x4x!HLFHE.eint<4>>, tensor<1x3xi5>) -> tensor<3x3x!HLFHE.eint<4>>
|
||||
|
||||
// Same behavior than the previous one, but as the dimension #2 is missing of operand #2.
|
||||
"HLFHELinalg.mul_eint_int(%a0, %a1)" : (tensor<3x4x!HLFHE.eint<4>>, tensor<3xi5>) -> tensor<4x4x4x!HLFHE.eint<4>>
|
||||
|
||||
```
|
||||
}];
|
||||
|
||||
let arguments = (ins
|
||||
Type<And<[TensorOf<[EncryptedIntegerType]>.predicate, HasStaticShapePred]>>:$lhs,
|
||||
Type<And<[TensorOf<[AnyInteger]>.predicate, HasStaticShapePred]>>:$rhs
|
||||
);
|
||||
|
||||
let results = (outs Type<And<[TensorOf<[EncryptedIntegerType]>.predicate, HasStaticShapePred]>>);
|
||||
|
||||
}
|
||||
|
||||
#endif
|
||||
|
||||
@@ -76,4 +76,46 @@ func @main(%a0: tensor<2x3x4x!HLFHE.eint<2>>, %a1: tensor<2x3x4x!HLFHE.eint<3>>)
|
||||
// expected-error @+1 {{'HLFHELinalg.add_eint' op should have the width of encrypted equals, got 3 expect 2}}
|
||||
%1 = "HLFHELinalg.add_eint"(%a0, %a1) : (tensor<2x3x4x!HLFHE.eint<2>>, tensor<2x3x4x!HLFHE.eint<3>>) -> tensor<2x3x4x!HLFHE.eint<2>>
|
||||
return %1 : tensor<2x3x4x!HLFHE.eint<2>>
|
||||
}
|
||||
}
|
||||
|
||||
// -----
|
||||
|
||||
/////////////////////////////////////////////////
|
||||
// HLFHELinalg.mul_eint_int
|
||||
/////////////////////////////////////////////////
|
||||
|
||||
// Incompatible dimension of operands
|
||||
func @main(%a0: tensor<2x2x3x4x!HLFHE.eint<2>>, %a1: tensor<2x2x2x4xi3>) -> tensor<2x2x3x4x!HLFHE.eint<2>> {
|
||||
// expected-error @+1 {{'HLFHELinalg.mul_eint_int' op has the dimension #2 of the operand #1 incompatible with other operands, got 2 expect 1 or 3}}
|
||||
%1 = "HLFHELinalg.mul_eint_int"(%a0, %a1) : (tensor<2x2x3x4x!HLFHE.eint<2>>, tensor<2x2x2x4xi3>) -> tensor<2x2x3x4x!HLFHE.eint<2>>
|
||||
return %1 : tensor<2x2x3x4x!HLFHE.eint<2>>
|
||||
}
|
||||
|
||||
// -----
|
||||
|
||||
// Incompatible dimension of result
|
||||
func @main(%a0: tensor<2x2x3x4x!HLFHE.eint<2>>, %a1: tensor<2x2x2x4xi3>) -> tensor<2x10x3x4x!HLFHE.eint<2>> {
|
||||
// expected-error @+1 {{'HLFHELinalg.mul_eint_int' op has the dimension #3 of the result incompatible with operands dimension, got 10 expect 2}}
|
||||
%1 = "HLFHELinalg.mul_eint_int"(%a0, %a1) : (tensor<2x2x3x4x!HLFHE.eint<2>>, tensor<2x2x2x4xi3>) -> tensor<2x10x3x4x!HLFHE.eint<2>>
|
||||
return %1 : tensor<2x10x3x4x!HLFHE.eint<2>>
|
||||
}
|
||||
|
||||
// -----
|
||||
|
||||
// Incompatible number of dimension between operands and result
|
||||
func @main(%a0: tensor<2x2x3x4x!HLFHE.eint<2>>, %a1: tensor<2x2x2x4xi3>) -> tensor<2x3x4x!HLFHE.eint<2>> {
|
||||
// expected-error @+1 {{'HLFHELinalg.mul_eint_int' op should have the number of dimensions of the result equal to the highest number of dimensions of operands, got 3 expect 4}}
|
||||
%1 = "HLFHELinalg.mul_eint_int"(%a0, %a1) : (tensor<2x2x3x4x!HLFHE.eint<2>>, tensor<2x2x2x4xi3>) -> tensor<2x3x4x!HLFHE.eint<2>>
|
||||
return %1 : tensor<2x3x4x!HLFHE.eint<2>>
|
||||
}
|
||||
|
||||
// -----
|
||||
|
||||
// Incompatible width between clear and encrypted witdh
|
||||
func @main(%a0: tensor<2x3x4x!HLFHE.eint<2>>, %a1: tensor<2x3x4xi4>) -> tensor<2x3x4x!HLFHE.eint<2>> {
|
||||
// expected-error @+1 {{'HLFHELinalg.mul_eint_int' op should have the width of integer values less or equals than the width of encrypted values + 1}}
|
||||
%1 = "HLFHELinalg.mul_eint_int"(%a0, %a1) : (tensor<2x3x4x!HLFHE.eint<2>>, tensor<2x3x4xi4>) -> tensor<2x3x4x!HLFHE.eint<2>>
|
||||
return %1 : tensor<2x3x4x!HLFHE.eint<2>>
|
||||
}
|
||||
|
||||
// -----
|
||||
@@ -161,4 +161,59 @@ func @sub_int_eint_broadcast_1(%a0: tensor<3x4x1xi3>, %a1: tensor<1x4x5x!HLFHE.e
|
||||
func @sub_int_eint_broadcast_2(%a0: tensor<3x4xi3>, %a1: tensor<4x!HLFHE.eint<2>>) -> tensor<3x4x!HLFHE.eint<2>> {
|
||||
%1 ="HLFHELinalg.sub_int_eint"(%a0, %a1) : (tensor<3x4xi3>, tensor<4x!HLFHE.eint<2>>) -> tensor<3x4x!HLFHE.eint<2>>
|
||||
return %1: tensor<3x4x!HLFHE.eint<2>>
|
||||
}
|
||||
|
||||
|
||||
/////////////////////////////////////////////////
|
||||
// HLFHELinalg.mul_eint_int
|
||||
/////////////////////////////////////////////////
|
||||
|
||||
// 1D tensor
|
||||
// CHECK: func @mul_eint_int_1D(%[[a0:.*]]: tensor<4x!HLFHE.eint<2>>, %[[a1:.*]]: tensor<4xi3>) -> tensor<4x!HLFHE.eint<2>> {
|
||||
// CHECK-NEXT: %[[V0:.*]] = "HLFHELinalg.mul_eint_int"(%[[a0]], %[[a1]]) : (tensor<4x!HLFHE.eint<2>>, tensor<4xi3>) -> tensor<4x!HLFHE.eint<2>>
|
||||
// CHECK-NEXT: return %[[V0]] : tensor<4x!HLFHE.eint<2>>
|
||||
// CHECK-NEXT: }
|
||||
func @mul_eint_int_1D(%a0: tensor<4x!HLFHE.eint<2>>, %a1: tensor<4xi3>) -> tensor<4x!HLFHE.eint<2>> {
|
||||
%1 = "HLFHELinalg.mul_eint_int"(%a0, %a1) : (tensor<4x!HLFHE.eint<2>>, tensor<4xi3>) -> tensor<4x!HLFHE.eint<2>>
|
||||
return %1: tensor<4x!HLFHE.eint<2>>
|
||||
}
|
||||
|
||||
// 2D tensor
|
||||
// CHECK: func @mul_eint_int_2D(%[[a0:.*]]: tensor<2x4x!HLFHE.eint<2>>, %[[a1:.*]]: tensor<2x4xi3>) -> tensor<2x4x!HLFHE.eint<2>> {
|
||||
// CHECK-NEXT: %[[V0:.*]] = "HLFHELinalg.mul_eint_int"(%[[a0]], %[[a1]]) : (tensor<2x4x!HLFHE.eint<2>>, tensor<2x4xi3>) -> tensor<2x4x!HLFHE.eint<2>>
|
||||
// CHECK-NEXT: return %[[V0]] : tensor<2x4x!HLFHE.eint<2>>
|
||||
// CHECK-NEXT: }
|
||||
func @mul_eint_int_2D(%a0: tensor<2x4x!HLFHE.eint<2>>, %a1: tensor<2x4xi3>) -> tensor<2x4x!HLFHE.eint<2>> {
|
||||
%1 = "HLFHELinalg.mul_eint_int"(%a0, %a1) : (tensor<2x4x!HLFHE.eint<2>>, tensor<2x4xi3>) -> tensor<2x4x!HLFHE.eint<2>>
|
||||
return %1: tensor<2x4x!HLFHE.eint<2>>
|
||||
}
|
||||
|
||||
// 10D tensor
|
||||
// CHECK: func @mul_eint_int_10D(%[[a0:.*]]: tensor<1x2x3x4x5x6x7x8x9x10x!HLFHE.eint<2>>, %[[a1:.*]]: tensor<1x2x3x4x5x6x7x8x9x10xi3>) -> tensor<1x2x3x4x5x6x7x8x9x10x!HLFHE.eint<2>> {
|
||||
// CHECK-NEXT: %[[V0:.*]] = "HLFHELinalg.mul_eint_int"(%[[a0]], %[[a1]]) : (tensor<1x2x3x4x5x6x7x8x9x10x!HLFHE.eint<2>>, tensor<1x2x3x4x5x6x7x8x9x10xi3>) -> tensor<1x2x3x4x5x6x7x8x9x10x!HLFHE.eint<2>>
|
||||
// CHECK-NEXT: return %[[V0]] : tensor<1x2x3x4x5x6x7x8x9x10x!HLFHE.eint<2>>
|
||||
// CHECK-NEXT: }
|
||||
func @mul_eint_int_10D(%a0: tensor<1x2x3x4x5x6x7x8x9x10x!HLFHE.eint<2>>, %a1: tensor<1x2x3x4x5x6x7x8x9x10xi3>) -> tensor<1x2x3x4x5x6x7x8x9x10x!HLFHE.eint<2>> {
|
||||
%1 = "HLFHELinalg.mul_eint_int"(%a0, %a1) : (tensor<1x2x3x4x5x6x7x8x9x10x!HLFHE.eint<2>>, tensor<1x2x3x4x5x6x7x8x9x10xi3>) -> tensor<1x2x3x4x5x6x7x8x9x10x!HLFHE.eint<2>>
|
||||
return %1: tensor<1x2x3x4x5x6x7x8x9x10x!HLFHE.eint<2>>
|
||||
}
|
||||
|
||||
// Broadcasting with tensor with dimensions equals to one
|
||||
// CHECK: func @mul_eint_int_broadcast_1(%[[a0:.*]]: tensor<1x4x5x!HLFHE.eint<2>>, %[[a1:.*]]: tensor<3x4x1xi3>) -> tensor<3x4x5x!HLFHE.eint<2>> {
|
||||
// CHECK-NEXT: %[[V0:.*]] = "HLFHELinalg.mul_eint_int"(%[[a0]], %[[a1]]) : (tensor<1x4x5x!HLFHE.eint<2>>, tensor<3x4x1xi3>) -> tensor<3x4x5x!HLFHE.eint<2>>
|
||||
// CHECK-NEXT: return %[[V0]] : tensor<3x4x5x!HLFHE.eint<2>>
|
||||
// CHECK-NEXT: }
|
||||
func @mul_eint_int_broadcast_1(%a0: tensor<1x4x5x!HLFHE.eint<2>>, %a1: tensor<3x4x1xi3>) -> tensor<3x4x5x!HLFHE.eint<2>> {
|
||||
%1 = "HLFHELinalg.mul_eint_int"(%a0, %a1) : (tensor<1x4x5x!HLFHE.eint<2>>, tensor<3x4x1xi3>) -> tensor<3x4x5x!HLFHE.eint<2>>
|
||||
return %1: tensor<3x4x5x!HLFHE.eint<2>>
|
||||
}
|
||||
|
||||
// Broadcasting with a tensor less dimensions of another
|
||||
// CHECK: func @mul_eint_int_broadcast_2(%[[a0:.*]]: tensor<4x!HLFHE.eint<2>>, %[[a1:.*]]: tensor<3x4xi3>) -> tensor<3x4x!HLFHE.eint<2>> {
|
||||
// CHECK-NEXT: %[[V0:.*]] = "HLFHELinalg.mul_eint_int"(%[[a0]], %[[a1]]) : (tensor<4x!HLFHE.eint<2>>, tensor<3x4xi3>) -> tensor<3x4x!HLFHE.eint<2>>
|
||||
// CHECK-NEXT: return %[[V0]] : tensor<3x4x!HLFHE.eint<2>>
|
||||
// CHECK-NEXT: }
|
||||
func @mul_eint_int_broadcast_2(%a0: tensor<4x!HLFHE.eint<2>>, %a1: tensor<3x4xi3>) -> tensor<3x4x!HLFHE.eint<2>> {
|
||||
%1 ="HLFHELinalg.mul_eint_int"(%a0, %a1) : (tensor<4x!HLFHE.eint<2>>, tensor<3x4xi3>) -> tensor<3x4x!HLFHE.eint<2>>
|
||||
return %1: tensor<3x4x!HLFHE.eint<2>>
|
||||
}
|
||||
Reference in New Issue
Block a user