| Class | Description |
|---|---|
| Append | |
| AppendG |
TODO Additional compiler enhancements:
1) Partial Shuffle Elimination - Any full or aligned blocks could be directly output from the mappers
to the result index.
|
| AppendGAlignedSP | |
| AppendM | |
| AppendR | |
| Binary |
Lop to perform binary operation.
|
| BinaryM |
Lop to perform binary operation.
|
| BinaryScalar |
Lop to perform binary scalar operations.
|
| BinaryUAggChain | |
| CentralMoment |
Lop to perform cross product operation
|
| Checkpoint |
Lop for checkpoint operations.
|
| Compression | |
| CoVariance |
Lop to compute covariance between two 1D matrices
|
| CSVReBlock |
Lop to convert CSV data into SystemDS data format
|
| Ctable |
Lop to perform ternary operation.
|
| CumulativeOffsetBinary | |
| CumulativePartialAggregate | |
| Data |
Lop to represent data objects.
|
| DataGen |
Defines a LOP that generates data.
|
| DeCompression | |
| DnnTransform | |
| Federated | |
| FunctionCallCP | |
| GroupedAggregate |
Lop to perform grouped aggregates
|
| GroupedAggregateM |
Lop to perform mr map-side grouped aggregates
(restriction: sum, w/o weights, ngroups), groups broadcasted
|
| LeftIndex | |
| Lop |
Base class for all Lops.
|
| LopProperties | |
| MapMult | |
| MapMultChain | |
| MatMultCP | |
| MMCJ |
Lop to perform cross product operation
|
| MMRJ |
Lop to perform cross product operation
|
| MMTSJ |
Lop to perform transpose-identity operation (t(X)%*%X or X%*%t(X)),
used to represent CP and MR instruction but in case of MR there is
an additional Aggregate at the reducers.
|
| MMZip |
Lop to perform zip matrix multiplication
|
| Nary |
Lop to perform an operation on a variable number of operands.
|
| OutputParameters |
class to maintain output parameters for a lop.
|
| ParameterizedBuiltin |
Defines a LOP for functions.
|
| PartialAggregate |
Lop to perform a partial aggregation.
|
| PickByCount | |
| PMapMult | |
| PMMJ | |
| ReBlock |
Lop to perform reblock operation
|
| RightIndex | |
| SortKeys | |
| SpoofFused | |
| Sql | |
| Ternary |
Lop to perform Sum of a matrix with another matrix multiplied by Scalar.
|
| TernaryAggregate | |
| Transform | |
| UAggOuterChain |
TODO Currently this lop only support the right hand side in distributed cache.
|
| Unary |
Lop to perform following operations: with one operand -- NOT(A), ABS(A),
SQRT(A), LOG(A) with two operands where one of them is a scalar -- H=H*i,
H=H*5, EXP(A,2), LOG(A,2)
|
| UnaryCP | |
| WeightedCrossEntropy | |
| WeightedCrossEntropyR | |
| WeightedDivMM | |
| WeightedDivMMR | |
| WeightedSigmoid | |
| WeightedSigmoidR | |
| WeightedSquaredLoss | |
| WeightedSquaredLossR | |
| WeightedUnaryMM | |
| WeightedUnaryMMR |
| Exception | Description |
|---|---|
| LopsException |
Copyright © 2021 The Apache Software Foundation. All rights reserved.