PPHlo API reference#

PPHlo is short for (SPU High level ops), it’s the assembly language of SPU.

PPHlo is built on MLIR infrastructure, the concrete ops definition could be found here.

Op List#

`pphlo.pad` (::mlir::pphlo::PadOp)#

Pad operator

Pads the edges of operand with the padding_value and according to the passed configuration.

See https://www.tensorflow.org/xla/operation_semantics#pad.

Traits: AlwaysSpeculatableImplTrait, InferTensorType, SameOperandsAndResultElementType

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:#

Attribute	MLIR Type	Description
`edge_padding_low`	::mlir::DenseIntElementsAttr	64-bit signless integer elements attribute
`edge_padding_high`	::mlir::DenseIntElementsAttr	64-bit signless integer elements attribute
`interior_padding`	::mlir::DenseIntElementsAttr	64-bit signless integer elements attribute

Operands:#

Operand	Description
`operand`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`padding_value`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.abs` (::mlir::pphlo::AbsOp)#

Absolute value operator

Returns abs(operand) element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_unary_functions.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape, SameOperandsAndResultType

Interfaces: ConditionallySpeculatable, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`operand`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.add` (::mlir::pphlo::AddOp)#

Addition operator

Returns lhs + rhs element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_binary_arithmetic_operations.

Traits: AlwaysSpeculatableImplTrait, Commutative, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`lhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`rhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.and` (::mlir::pphlo::AndOp)#

Logical and

Returns logical_and(lhs, rhs) element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_binary_arithmetic_operations.

Traits: AlwaysSpeculatableImplTrait, Commutative, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`lhs`	statically shaped tensor of public integer type or secret integer type values
`rhs`	statically shaped tensor of public integer type or secret integer type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.argmax` (::mlir::pphlo::ArgMaxOp)#

ArgMax operator

Returns the max value and index in a window.

Traits: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:#

Attribute	MLIR Type	Description
`window_dimensions`	::mlir::DenseIntElementsAttr	64-bit signless integer elements attribute
`window_strides`	::mlir::DenseIntElementsAttr	64-bit signless integer elements attribute
`base_dilations`	::mlir::DenseIntElementsAttr	64-bit signless integer elements attribute
`window_dilations`	::mlir::DenseIntElementsAttr	64-bit signless integer elements attribute
`padding`	::mlir::DenseIntElementsAttr	64-bit signless integer elements attribute
`onehot_index`	::mlir::BoolAttr	bool attribute

Operands:#

Operand	Description
`input`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values
«unnamed»	statically shaped tensor of public integer type or secret integer type values

`pphlo.bitcast_convert` (::mlir::pphlo::BitcastConvertOp)#

BitcastConvert operator

Similar to a ‘tf.bitcast’ in TensorFlow, performs an element-wise bitcast operation from a data shape to a target shape. The dimensions must match, and the conversion is an element-wise one. Bitcast is implemented as a low-level cast, so machines with different floating-point representations will give different results.

See https://www.tensorflow.org/xla/operation_semantics#bitcastconverttype.

Traits: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`operand`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.broadcast` (::mlir::pphlo::BroadcastOp)#

Broadcast a tensor into the given shape by adding dimensions.

Broadcasts the operand tensor to a higher rank. This is not the limited form of broadcasting exposed as the XLA client broadcast op, but rather the more powerful “InDim” broadcasting, which is closer to the HLO broadcast op and exposed in the XLA client BroadcastInDim method.

broadcast_dimensions maps the operand dimension number to the target shape dimension number. It must have the same size as the rank of the operand. The mapped dimensions must either be the same size or the dimension being broadcast from must be size 1 (degenerate broadcasting).

For a scalar (0D tensor) operand, broadcast_dimensions must be empty. The The scalar value will be broadcast to every element in the target shape.

See https://www.tensorflow.org/xla/broadcasting.

Traits: AlwaysSpeculatableImplTrait, SameOperandsAndResultElementType

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:#

Attribute	MLIR Type	Description
`broadcast_dimensions`	::mlir::DenseIntElementsAttr	64-bit signless integer elements attribute

Operands:#

Operand	Description
`operand`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.case` (::mlir::pphlo::CaseOp)#

Switch-Case operator

Returns the result of executing branches[index]. If index is < 0 or >= N, then branches[N-1] is executed as the default branch.

The type of the returned values of each branch must be the same and equal to the types of the values returned by the operation.

Note that if index is public only one of the branches will be executed depending on the value of index, otherwise all branches will be executed and select results based on index value.

Traits: RecursiveMemoryEffects, SingleBlockImplicitTerminator

Operands:#

Operand	Description
`index`	statically shaped tensor of public integer type or secret integer type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.ceil` (::mlir::pphlo::CeilOp)#

Ceil operator

Returns Ceil(operand) element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_unary_functions.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape, SameOperandsAndResultType

Interfaces: ConditionallySpeculatable, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`operand`	statically shaped tensor of public fixed-point type or secret fixed-point type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of public fixed-point type or secret fixed-point type values

`pphlo.clamp` (::mlir::pphlo::ClampOp)#

Clamp operator

Clamps an operand to within the range between a minimum and maximum value.

Note: All three arrays must be the same shape. Alternatively, as a restricted form of broadcasting, min and/or max can be a scalar (0D tensor) of the element type of the tensor operand.

See https://www.tensorflow.org/xla/operation_semantics#clamp.

Traits: AlwaysSpeculatableImplTrait, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`min`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`operand`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`max`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.concatenate` (::mlir::pphlo::ConcatenateOp)#

Concatenate op

Concatenates a set of tensors along the specified dimension.

See https://www.tensorflow.org/xla/operation_semantics#concatenate.

Traits: AlwaysSpeculatableImplTrait, SameOperandsAndResultElementType

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:#

Attribute	MLIR Type	Description
`dimension`	::mlir::IntegerAttr	64-bit signless integer attribute

Operands:#

Operand	Description
`val`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.constant` (::mlir::pphlo::ConstantOp)#

Constant operator

Create a constant value. Traits: AlwaysSpeculatableImplTrait, ConstantLike

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:#

Attribute	MLIR Type	Description
`value`	::mlir::ElementsAttr	constant vector/tensor attribute

Results:#

Result	Description
`output`	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.convert` (::mlir::pphlo::ConvertOp)#

Convert operator

Performs element-wise conversion of values from one type to another, e.g.fxp to int.

See https://www.tensorflow.org/xla/operation_semantics#convertelementtype.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`operand`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.convolution` (::mlir::pphlo::ConvolutionOp)#

Convolution operator

Syntax:

operation ::= `pphlo.convolution` `(`operands`)`
              `dim_numbers` `=` custom<ConvolutionDimensions>($dimension_numbers) `,`
              `window` `=` `{` custom<WindowAttributes>($window_strides) `}`
              attr-dict `:` functional-type(operands, results)

Computes a convolution of the kind used in neural networks.

See https://www.tensorflow.org/xla/operation_semantics#conv_convolution.

Traits: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:#

Attribute	MLIR Type	Description
`window_strides`	::mlir::DenseIntElementsAttr	64-bit signless integer elements attribute
`dimension_numbers`	::mlir::pphlo::ConvDimensionNumbersAttr	Structure of dimension information for conv op
`feature_group_count`	::mlir::IntegerAttr	64-bit signless integer attribute
`batch_group_count`	::mlir::IntegerAttr	64-bit signless integer attribute

Operands:#

Operand	Description
`lhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`rhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.dbg_print` (::mlir::pphlo::DbgPrintOp)#

Operands:#

Operand	Description
`operand`	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.divide` (::mlir::pphlo::DivOp)#

Division operator

Returns lhs / rhs element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_binary_arithmetic_operations.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`lhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`rhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.dot_general` (::mlir::pphlo::DotGeneralOp)#

General Dot operator

Performs general dot products between vectors, vector/matrix and matrix/matrix multiplication.

See https://www.tensorflow.org/xla/operation_semantics#dotgeneral.

Traits: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:#

Attribute	MLIR Type	Description
`dot_dimension_numbers`	::mlir::pphlo::DotDimensionNumbersAttr	Attribute that models the dimension information for dot.

Operands:#

Operand	Description
`lhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`rhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.dot` (::mlir::pphlo::DotOp)#

Dot operator

Performs dot products between vectors, vector/matrix and matrix/matrix multiplication.

See https://www.tensorflow.org/xla/operation_semantics#dot.

Traits: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`lhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`rhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.dynamic-slice` (::mlir::pphlo::DynamicSliceOp)#

Dynamic Slice operator

Extracts a sub-array from the input array at dynamic start_indices.

See https://www.tensorflow.org/xla/operation_semantics#dynamicslice.

Traits: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:#

Attribute	MLIR Type	Description
`slice_sizes`	::mlir::DenseIntElementsAttr	64-bit signless integer elements attribute

Operands:#

Operand	Description
`operand`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`start_indices`	0D tensor of public integer type or secret integer type values

Results:#

Result	Description
`result`	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.dynamic-update-slice` (::mlir::pphlo::DynamicUpdateSliceOp)#

Dynamic Update Slice operator

DynamicUpdateSlice generates a result which is the value of the input array operand, with a slice update overwritten at start_indices.

See https://www.tensorflow.org/xla/operation_semantics#dynamicupdateslice.

Traits: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`operand`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`update`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`start_indices`	0D tensor of public integer type or secret integer type values

Results:#

Result	Description
`result`	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.epsilon` (::mlir::pphlo::EpsilonOp)#

Epsilon operator

Return a value that representation runtime epsilon. Traits: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Results:#

Result	Description
`output`	statically shaped tensor of public fixed-point type or secret fixed-point type values

`pphlo.equal` (::mlir::pphlo::EqualOp)#

Equal comparison operator

Returns lhs == rhs element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_comparison_operations.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`lhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`rhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of public integer type or secret integer type values

`pphlo.exponential` (::mlir::pphlo::ExpOp)#

Exponential operator

Returns e ^ (operand) element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_unary_functions.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape, SameOperandsAndResultType

Interfaces: ConditionallySpeculatable, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`operand`	statically shaped tensor of public fixed-point type or secret fixed-point type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of public fixed-point type or secret fixed-point type values

`pphlo.exponential_minus_one` (::mlir::pphlo::Expm1Op)#

Exponential minus one operator

Returns e^(operand) - 1 element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_unary_functions.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape, SameOperandsAndResultType

Interfaces: ConditionallySpeculatable, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`operand`	statically shaped tensor of public fixed-point type or secret fixed-point type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of public fixed-point type or secret fixed-point type values

`pphlo.floor` (::mlir::pphlo::FloorOp)#

Floor operator

Returns Floor(operand) element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_unary_functions.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape, SameOperandsAndResultType

Interfaces: ConditionallySpeculatable, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`operand`	statically shaped tensor of public fixed-point type or secret fixed-point type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of public fixed-point type or secret fixed-point type values

`pphlo.gather` (::mlir::pphlo::GatherOp)#

Gather operator

Stitches together several slices of operand from offsets specified in start_indices (each slice at a potentially different runtime offset).

See https://www.tensorflow.org/xla/operation_semantics#gather.

Traits: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:#

Attribute	MLIR Type	Description
`dimension_numbers`	::mlir::pphlo::GatherDimensionNumbersAttr	Attribute that models the dimension information for gather
`slice_sizes`	::mlir::DenseIntElementsAttr	64-bit signless integer elements attribute
`indices_are_sorted`	::mlir::BoolAttr	bool attribute

Operands:#

Operand	Description
`operand`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`start_indices`	statically shaped tensor of public integer type or secret integer type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.greater_equal` (::mlir::pphlo::GreaterEqualOp)#

greater_equal comparison operator

Returns lhs >= rhs element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_comparison_operations.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`lhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`rhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of public integer type or secret integer type values

`pphlo.greater` (::mlir::pphlo::GreaterOp)#

greater comparison operator

elementwise lhs > rhs. Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`lhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`rhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of public integer type or secret integer type values

`pphlo.if` (::mlir::pphlo::IfOp)#

If operator

Returns the result of executing either a true or false function depending on the result of a condition function.

See https://www.tensorflow.org/xla/operation_semantics#conditional.

Traits: RecursiveMemoryEffects, SingleBlockImplicitTerminator

Operands:#

Operand	Description
`condition`	statically shaped tensor of public integer type or secret integer type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.iota` (::mlir::pphlo::IotaOp)#

Iota operator

Creates a rank 1 array of values starting at zero and incrementing by one.

See https://www.tensorflow.org/xla/operation_semantics#iota

Traits: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:#

Attribute	MLIR Type	Description
`iota_dimension`	::mlir::IntegerAttr	64-bit signless integer attribute

Results:#

Result	Description
`output`	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.less_equal` (::mlir::pphlo::LessEqualOp)#

less_equal comparison operator

elementwise lhs <= rhs. Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`lhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`rhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of public integer type or secret integer type values

`pphlo.less` (::mlir::pphlo::LessOp)#

less comparison operator

Returns lhs < rhs element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_comparison_operations.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`lhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`rhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of public integer type or secret integer type values

`pphlo.log_plus_one` (::mlir::pphlo::Log1pOp)#

Log1p operator

Returns log(operand + 1) element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_unary_functions.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape, SameOperandsAndResultType

Interfaces: ConditionallySpeculatable, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`operand`	statically shaped tensor of public fixed-point type or secret fixed-point type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of public fixed-point type or secret fixed-point type values

`pphlo.log` (::mlir::pphlo::LogOp)#

Log operator

Returns log(operand) element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_unary_functions.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape, SameOperandsAndResultType

Interfaces: ConditionallySpeculatable, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`operand`	statically shaped tensor of public fixed-point type or secret fixed-point type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of public fixed-point type or secret fixed-point type values

`pphlo.logistic` (::mlir::pphlo::LogisticOp)#

Reciprocal operator

Returns logistic(operand) element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_unary_functions.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape, SameOperandsAndResultType

Interfaces: ConditionallySpeculatable, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`operand`	statically shaped tensor of public fixed-point type or secret fixed-point type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of public fixed-point type or secret fixed-point type values

`pphlo.maximum` (::mlir::pphlo::MaxOp)#

Maximum operator

Returns max(lhs, rhs) element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_binary_arithmetic_operations.

Traits: AlwaysSpeculatableImplTrait, Commutative, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`lhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`rhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.maxpool_scatter` (::mlir::pphlo::MaxPoolScatterOp)#

MaxPool Scatter operator

Generates a result which is the value of the input array operand, with several slices (at indices specified by scatter_indices) updated with the values in updates using update_computation.

See https://www.tensorflow.org/xla/operation_semantics#scatter.

Traits: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:#

Attribute	MLIR Type	Description
`window_dimensions`	::mlir::DenseIntElementsAttr	64-bit signless integer elements attribute
`window_strides`	::mlir::DenseIntElementsAttr	64-bit signless integer elements attribute
`padding`	::mlir::DenseIntElementsAttr	64-bit signless integer elements attribute

Operands:#

Operand	Description
`scatter_indices`	statically shaped tensor of public integer type or secret integer type values
`update`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.minimum` (::mlir::pphlo::MinOp)#

Minimum operator

Returns min(lhs, rhs) element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_binary_arithmetic_operations.

Traits: AlwaysSpeculatableImplTrait, Commutative, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`lhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`rhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.multiply` (::mlir::pphlo::MulOp)#

Multiplication operator

Returns lhs * rhs element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_binary_arithmetic_operations.

Traits: AlwaysSpeculatableImplTrait, Commutative, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`lhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`rhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.negate` (::mlir::pphlo::NegOp)#

Negation operator

Returns - operand element - wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_unary_functions.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape, SameOperandsAndResultType

Interfaces: ConditionallySpeculatable, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`operand`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.not_equal` (::mlir::pphlo::NotEqualOp)#

Not-equal comparison operator

Returns lhs != rhs element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_comparison_operations.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`lhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`rhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of public integer type or secret integer type values

`pphlo.not` (::mlir::pphlo::NotOp)#

Not operator

Returns !operand element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_unary_functions.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape, SameOperandsAndResultType

Interfaces: ConditionallySpeculatable, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`operand`	statically shaped tensor of public integer type or secret integer type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of public integer type or secret integer type values

`pphlo.or` (::mlir::pphlo::OrOp)#

Logical or

Returns logical_or(lhs, rhs) element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_binary_arithmetic_operations.

Traits: AlwaysSpeculatableImplTrait, Commutative, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`lhs`	statically shaped tensor of public integer type or secret integer type values
`rhs`	statically shaped tensor of public integer type or secret integer type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.power` (::mlir::pphlo::PowOp)#

Power operator

Returns lhs ^ rhs element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_binary_arithmetic_operations.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`lhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`rhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.prefer_a` (::mlir::pphlo::PreferAOp)#

Prefer AShare operator

Convert input to AShare if possible.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape, SameOperandsAndResultType

Interfaces: ConditionallySpeculatable, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`operand`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.reciprocal` (::mlir::pphlo::ReciprocalOp)#

Reciprocal operator

Returns 1.0 / operand element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_unary_functions.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape, SameOperandsAndResultType

Interfaces: ConditionallySpeculatable, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`operand`	statically shaped tensor of public fixed-point type or secret fixed-point type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of public fixed-point type or secret fixed-point type values

`pphlo.reduce` (::mlir::pphlo::ReduceOp)#

Reduce operator

Returns the result of executing a reduction function on one or more arrays in parallel.

See https://www.tensorflow.org/xla/operation_semantics#reduce.

Traits: RecursiveMemoryEffects, SameVariadicOperandSize, SingleBlockImplicitTerminator

Attributes:#

Attribute	MLIR Type	Description
`dimensions`	::mlir::DenseIntElementsAttr	64-bit signless integer elements attribute

Operands:#

Operand	Description
`inputs`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`init_values`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.reduce_window` (::mlir::pphlo::ReduceWindowOp)#

ReduceWindow operator

Returns the result of executing a reduction function over all elements in each window of one or more arrays.

See https://www.tensorflow.org/xla/operation_semantics#reducewindow.

Traits: RecursiveMemoryEffects, SameVariadicOperandSize, SingleBlockImplicitTerminator

Attributes:#

Attribute	MLIR Type	Description
`window_dimensions`	::mlir::DenseIntElementsAttr	64-bit signless integer elements attribute
`window_strides`	::mlir::DenseIntElementsAttr	64-bit signless integer elements attribute
`base_dilations`	::mlir::DenseIntElementsAttr	64-bit signless integer elements attribute
`window_dilations`	::mlir::DenseIntElementsAttr	64-bit signless integer elements attribute
`padding`	::mlir::DenseIntElementsAttr	64-bit signless integer elements attribute

Operands:#

Operand	Description
`inputs`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`init_values`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.remainder` (::mlir::pphlo::RemOp)#

Remainder operator

Returns lhs % rhs element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_binary_arithmetic_operations.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`lhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`rhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.reshape` (::mlir::pphlo::ReshapeOp)#

Reshape operator

Reshapes the dimensions of operand into a new configuration.

See https://www.tensorflow.org/xla/operation_semantics#reshape.

Traits: AlwaysSpeculatableImplTrait, SameOperandsAndResultElementType

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`operand`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.return` (::mlir::pphlo::ReturnOp)#

The `pphlo.return` operation terminates a region and returns values.

Traits: AlwaysSpeculatableImplTrait, Terminator

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`results`	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.reverse` (::mlir::pphlo::ReverseOp)#

Reverse operator

Reverses the specified dimensions of operand according to the given dimensions.

See https://www.tensorflow.org/xla/operation_semantics#rev_reverse.

Traits: AlwaysSpeculatableImplTrait, SameOperandsAndResultType

Interfaces: ConditionallySpeculatable, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:#

Attribute	MLIR Type	Description
`dimensions`	::mlir::DenseIntElementsAttr	64-bit signless integer elements attribute

Operands:#

Operand	Description
`operand`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.rng` (::mlir::pphlo::RngOp)#

RNG with uniform distribution.

Constructs an output of a given shape with random numbers generated following the uniform distribution over the interval [a,b). The parameters and output element type have to be an integral type or a fixed point type, and the types have to be consistent.

See https://www.tensorflow.org/xla/operation_semantics#rnguniform.

Traits: SameOperandsAndResultElementType

Operands:#

Operand	Description
`a`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`b`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.round_nearest_afz` (::mlir::pphlo::RoundOp)#

Round operator, ties away from zero

Returns Round(operand) element-wise, rounding to nearest integer with half-way cases rounding away from zero.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_unary_functions.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape, SameOperandsAndResultType

Interfaces: ConditionallySpeculatable, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`operand`	statically shaped tensor of public fixed-point type or secret fixed-point type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of public fixed-point type or secret fixed-point type values

`pphlo.rsqrt` (::mlir::pphlo::RsqrtOp)#

Reciprocal of square-root operator

Returns rsqrt(operand) element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_unary_functions.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape, SameOperandsAndResultType

Interfaces: ConditionallySpeculatable, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`operand`	statically shaped tensor of public fixed-point type or secret fixed-point type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of public fixed-point type or secret fixed-point type values

`pphlo.select_and_scatter` (::mlir::pphlo::SelectAndScatterOp)#

SelectAndScatter operator

Runs a windowed selection select function over operand with shape window_dimensions and stride window_strides. This will produce an amount of selected locations whose shape matches source. These are then scattered to the output which is initialized with init_value. Multiple scattered elements which land in the same output location are combined using the scatter function.

See https://www.tensorflow.org/xla/operation_semantics#selectandscatter.

Traits: RecursiveMemoryEffects

Attributes:#

Attribute	MLIR Type	Description
`window_dimensions`	::mlir::DenseIntElementsAttr	64-bit signless integer elements attribute
`window_strides`	::mlir::DenseIntElementsAttr	64-bit signless integer elements attribute
`padding`	::mlir::DenseIntElementsAttr	64-bit signless integer elements attribute

Operands:#

Operand	Description
`operand`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`source`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`init_value`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.select` (::mlir::pphlo::SelectOp)#

Select operator

Constructs an output tensor from the elements of on_true and on_false based on the values of pred. All three operands must be of the same shape with the exception of pred, which may also be a scalar in which case it is broadcasted.

See https://www.tensorflow.org/xla/operation_semantics#select.

Traits: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`pred`	statically shaped tensor of public integer type or secret integer type values
`on_true`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`on_false`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.shift_left` (::mlir::pphlo::ShiftLeftOp)#

Shift Left operator

Returns lhs << rhs element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_binary_arithmetic_operations.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`lhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`rhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.shift_right_arithmetic` (::mlir::pphlo::ShiftRightArithmeticOp)#

Shift right arithmetic operator

Returns arithmetic lhs >> rhs element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_binary_arithmetic_operations.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`lhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`rhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.shift_right_logical` (::mlir::pphlo::ShiftRightLogicalOp)#

Shift right logical operator

Returns logical lhs >> rhs element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_binary_arithmetic_operations.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`lhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`rhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.sign` (::mlir::pphlo::SignOp)#

Sign operator

Returns sign(operand) element-wise, where

sign(x) = -1  : x < 0
        = 0   : x = 0
        = 1   : x > 0

See https://www.tensorflow.org/xla/operation_semantics#element-wise_unary_functions.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape, SameOperandsAndResultType

Interfaces: ConditionallySpeculatable, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`operand`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.slice` (::mlir::pphlo::SliceOp)#

The dynamic shape version of SliceOp. Extracts a sub-array from the input array according to start_indices, limit_indices and strides. Expect start_indices/limit_indices/strides to be statically shaped and matching the rank of the input.

See https://www.tensorflow.org/xla/operation_semantics#slice

Traits: AlwaysSpeculatableImplTrait, SameOperandsAndResultElementType

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:#

Attribute	MLIR Type	Description
`start_indices`	::mlir::DenseIntElementsAttr	64-bit signless integer elements attribute
`limit_indices`	::mlir::DenseIntElementsAttr	64-bit signless integer elements attribute
`strides`	::mlir::DenseIntElementsAttr	64-bit signless integer elements attribute

Operands:#

Operand	Description
`operand`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.sort` (::mlir::pphlo::SortOp)#

Sort operator

Sorts the given operands at the given dimension with the given comparator.

See https://www.tensorflow.org/xla/operation_semantics#sort.

Traits: RecursiveMemoryEffects, SameOperandsAndResultShape

Attributes:#

Attribute	MLIR Type	Description
`dimension`	::mlir::IntegerAttr	64-bit signless integer attribute
`is_stable`	::mlir::BoolAttr	bool attribute

Operands:#

Operand	Description
`operands`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.sqrt` (::mlir::pphlo::SqrtOp)#

Square-root operator

Returns sqrt(operand) element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_unary_functions.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape, SameOperandsAndResultType

Interfaces: ConditionallySpeculatable, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`operand`	statically shaped tensor of public fixed-point type or secret fixed-point type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of public fixed-point type or secret fixed-point type values

`pphlo.subtract` (::mlir::pphlo::SubtractOp)#

Subtraction operator

Returns lhs - rhs element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_binary_arithmetic_operations.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`lhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values
`rhs`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.tanh` (::mlir::pphlo::TanhOp)#

Tanh operator

Returns tanh(operand) element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_unary_functions.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape, SameOperandsAndResultType

Interfaces: ConditionallySpeculatable, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`operand`	statically shaped tensor of public fixed-point type or secret fixed-point type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of public fixed-point type or secret fixed-point type values

`pphlo.transpose` (::mlir::pphlo::TransposeOp)#

Transpose operator

Permutes the dimensions of operand according to the given permutation.

res_dimensions[i] = operand_dimensions[permutation[i]]

See https://www.tensorflow.org/xla/operation_semantics#transpose.

Traits: AlwaysSpeculatableImplTrait, SameOperandsAndResultElementType

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:#

Attribute	MLIR Type	Description
`permutation`	::mlir::DenseIntElementsAttr	64-bit signless integer elements attribute

Operands:#

Operand	Description
`operand`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.while` (::mlir::pphlo::WhileOp)#

While operator

Returns the result of executing a body function until the cond body returns true.

See https://www.tensorflow.org/xla/operation_semantics#while.

Traits: PPHLO_PairwiseSameOperandAndResultType, RecursiveMemoryEffects, SingleBlockImplicitTerminator

Operands:#

Operand	Description
`args`	statically shaped tensor of PPHlo public type or PPHlo secret type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

`pphlo.xor` (::mlir::pphlo::XorOp)#

Logical xor

Returns logical_xor(lhs, rhs) element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_binary_arithmetic_operations.

Traits: AlwaysSpeculatableImplTrait, Commutative, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:#

Operand	Description
`lhs`	statically shaped tensor of public integer type or secret integer type values
`rhs`	statically shaped tensor of public integer type or secret integer type values

Results:#

Result	Description
«unnamed»	statically shaped tensor of PPHlo public type or PPHlo secret type values

PPHlo API reference#

Op List#

pphlo.pad (::mlir::pphlo::PadOp)#

Attributes:#

Operands:#

Results:#

pphlo.abs (::mlir::pphlo::AbsOp)#

Operands:#

Results:#

pphlo.add (::mlir::pphlo::AddOp)#

Operands:#

Results:#

pphlo.and (::mlir::pphlo::AndOp)#

Operands:#

Results:#

pphlo.argmax (::mlir::pphlo::ArgMaxOp)#

Attributes:#

Operands:#

Results:#

pphlo.bitcast_convert (::mlir::pphlo::BitcastConvertOp)#

Operands:#

Results:#

pphlo.broadcast (::mlir::pphlo::BroadcastOp)#

Attributes:#

Operands:#

Results:#

pphlo.case (::mlir::pphlo::CaseOp)#

Operands:#

Results:#

pphlo.ceil (::mlir::pphlo::CeilOp)#

Operands:#

Results:#

pphlo.clamp (::mlir::pphlo::ClampOp)#

Operands:#

Results:#

pphlo.concatenate (::mlir::pphlo::ConcatenateOp)#

Attributes:#

Operands:#

Results:#

pphlo.constant (::mlir::pphlo::ConstantOp)#

Attributes:#

Results:#

pphlo.convert (::mlir::pphlo::ConvertOp)#

Operands:#

Results:#

pphlo.convolution (::mlir::pphlo::ConvolutionOp)#

Attributes:#

Operands:#

Results:#

pphlo.dbg_print (::mlir::pphlo::DbgPrintOp)#

Operands:#

pphlo.divide (::mlir::pphlo::DivOp)#

Operands:#

Results:#

pphlo.dot_general (::mlir::pphlo::DotGeneralOp)#

Attributes:#

Operands:#

Results:#

pphlo.dot (::mlir::pphlo::DotOp)#

Operands:#

Results:#

pphlo.dynamic-slice (::mlir::pphlo::DynamicSliceOp)#

Attributes:#

Operands:#

Results:#

pphlo.dynamic-update-slice (::mlir::pphlo::DynamicUpdateSliceOp)#

Operands:#

Results:#

pphlo.epsilon (::mlir::pphlo::EpsilonOp)#

Results:#

pphlo.equal (::mlir::pphlo::EqualOp)#

Operands:#

Results:#

pphlo.exponential (::mlir::pphlo::ExpOp)#

Operands:#

Results:#

pphlo.exponential_minus_one (::mlir::pphlo::Expm1Op)#

Operands:#

Results:#

pphlo.floor (::mlir::pphlo::FloorOp)#

`pphlo.pad` (::mlir::pphlo::PadOp)#

`pphlo.abs` (::mlir::pphlo::AbsOp)#

`pphlo.add` (::mlir::pphlo::AddOp)#

`pphlo.and` (::mlir::pphlo::AndOp)#

`pphlo.argmax` (::mlir::pphlo::ArgMaxOp)#

`pphlo.bitcast_convert` (::mlir::pphlo::BitcastConvertOp)#

`pphlo.broadcast` (::mlir::pphlo::BroadcastOp)#

`pphlo.case` (::mlir::pphlo::CaseOp)#

`pphlo.ceil` (::mlir::pphlo::CeilOp)#

`pphlo.clamp` (::mlir::pphlo::ClampOp)#

`pphlo.concatenate` (::mlir::pphlo::ConcatenateOp)#

`pphlo.constant` (::mlir::pphlo::ConstantOp)#

`pphlo.convert` (::mlir::pphlo::ConvertOp)#

`pphlo.convolution` (::mlir::pphlo::ConvolutionOp)#

`pphlo.dbg_print` (::mlir::pphlo::DbgPrintOp)#

`pphlo.divide` (::mlir::pphlo::DivOp)#

`pphlo.dot_general` (::mlir::pphlo::DotGeneralOp)#

`pphlo.dot` (::mlir::pphlo::DotOp)#

`pphlo.dynamic-slice` (::mlir::pphlo::DynamicSliceOp)#

`pphlo.dynamic-update-slice` (::mlir::pphlo::DynamicUpdateSliceOp)#

`pphlo.epsilon` (::mlir::pphlo::EpsilonOp)#

`pphlo.equal` (::mlir::pphlo::EqualOp)#

`pphlo.exponential` (::mlir::pphlo::ExpOp)#

`pphlo.exponential_minus_one` (::mlir::pphlo::Expm1Op)#

`pphlo.floor` (::mlir::pphlo::FloorOp)#

`pphlo.gather` (::mlir::pphlo::GatherOp)#

`pphlo.greater_equal` (::mlir::pphlo::GreaterEqualOp)#

`pphlo.greater` (::mlir::pphlo::GreaterOp)#

`pphlo.if` (::mlir::pphlo::IfOp)#

`pphlo.iota` (::mlir::pphlo::IotaOp)#

`pphlo.less_equal` (::mlir::pphlo::LessEqualOp)#

`pphlo.less` (::mlir::pphlo::LessOp)#

`pphlo.log_plus_one` (::mlir::pphlo::Log1pOp)#

`pphlo.log` (::mlir::pphlo::LogOp)#

`pphlo.logistic` (::mlir::pphlo::LogisticOp)#

`pphlo.maximum` (::mlir::pphlo::MaxOp)#

`pphlo.maxpool_scatter` (::mlir::pphlo::MaxPoolScatterOp)#

`pphlo.minimum` (::mlir::pphlo::MinOp)#

`pphlo.multiply` (::mlir::pphlo::MulOp)#

`pphlo.negate` (::mlir::pphlo::NegOp)#

`pphlo.not_equal` (::mlir::pphlo::NotEqualOp)#

`pphlo.not` (::mlir::pphlo::NotOp)#

`pphlo.or` (::mlir::pphlo::OrOp)#

`pphlo.power` (::mlir::pphlo::PowOp)#

`pphlo.prefer_a` (::mlir::pphlo::PreferAOp)#

`pphlo.reciprocal` (::mlir::pphlo::ReciprocalOp)#

`pphlo.reduce` (::mlir::pphlo::ReduceOp)#