public class LibTensorAgg extends Object
Constructor and Description |
---|
LibTensorAgg() |
Modifier and Type | Method and Description |
---|---|
static void |
aggregateBinaryTensor(BasicTensorBlock in,
BasicTensorBlock aggVal,
AggregateOperator aop)
Core incremental tensor aggregate (ak+) as used for uack+ and acrk+.
|
static void |
aggregateUnaryTensor(BasicTensorBlock in,
BasicTensorBlock out,
AggregateUnaryOperator uaop)
Aggregate a tensor-block with the given unary operator.
|
static boolean |
isSupportedUnaryAggregateOperator(AggregateUnaryOperator op)
Determines whether the unary operator is supported.
|
static boolean |
satisfiesMultiThreadingConstraints(BasicTensorBlock in,
int k)
Check if a aggregation fulfills the constraints to be split to multiple threads.
|
public static boolean satisfiesMultiThreadingConstraints(BasicTensorBlock in, int k)
in
- the tensor block to be aggregatedk
- the number of threadspublic static void aggregateUnaryTensor(BasicTensorBlock in, BasicTensorBlock out, AggregateUnaryOperator uaop)
in
- the input tensor blockout
- the output tensor block containing the aggregated resultuaop
- the unary operation to applypublic static void aggregateBinaryTensor(BasicTensorBlock in, BasicTensorBlock aggVal, AggregateOperator aop)
in
- partial aggregationaggVal
- partial aggregation, also output (in will be added to this)aop
- aggregation operatorpublic static boolean isSupportedUnaryAggregateOperator(AggregateUnaryOperator op)
op
- the unary operator to checkCopyright © 2020 The Apache Software Foundation. All rights reserved.