org.apache.mahout.math.drm

DistributedEngine

trait DistributedEngine extends AnyRef

Abstraction of optimizer/distributed engine

Linear Supertypes
AnyRef, Any
Ordering
  1. Alphabetic
  2. By inheritance
Inherited
  1. DistributedEngine
  2. AnyRef
  3. Any
  1. Hide All
  2. Show all
Learn more about member selection
Visibility
  1. Public
  2. All

Abstract Value Members

  1. abstract def allreduceBlock[K](drm: CheckpointedDrm[K], bmf: ((Array[K], Matrix)) ⇒ Matrix, rf: (Matrix, Matrix) ⇒ Matrix): Matrix

    Optional engine-specific all reduce tensor operation.

  2. abstract def colMeans[K](drm: CheckpointedDrm[K]): Vector

    Engine-specific colMeans implementation based on a checkpoint.

  3. abstract def colSums[K](drm: CheckpointedDrm[K]): Vector

    Engine-specific colSums implementation based on a checkpoint.

  4. abstract def drm2IntKeyed[K](drmX: DrmLike[K], computeMap: Boolean = false): (DrmLike[Int], Option[DrmLike[K]])

    Convert non-int-keyed matrix to an int-keyed, computing optionally mapping from old keys to row indices in the new one.

    Convert non-int-keyed matrix to an int-keyed, computing optionally mapping from old keys to row indices in the new one. The mapping, if requested, is returned as a 1-column matrix.

  5. abstract def drmBroadcast(m: Matrix)(implicit dc: DistributedContext): BCast[Matrix]

    Broadcast support

  6. abstract def drmBroadcast(v: Vector)(implicit dc: DistributedContext): BCast[Vector]

    Broadcast support

  7. abstract def drmDfsRead(path: String, parMin: Int = 0)(implicit sc: DistributedContext): CheckpointedDrm[_]

    Load DRM from hdfs (as in Mahout DRM format).

    Load DRM from hdfs (as in Mahout DRM format). <P/>

    path

    The DFS path to load from

    parMin

    Minimum parallelism after load (equivalent to #par(min=...)).

  8. abstract def drmParallelizeEmpty(nrow: Int, ncol: Int, numPartitions: Int = 10)(implicit sc: DistributedContext): CheckpointedDrm[Int]

    This creates an empty DRM with specified number of partitions and cardinality.

  9. abstract def drmParallelizeEmptyLong(nrow: Long, ncol: Int, numPartitions: Int = 10)(implicit sc: DistributedContext): CheckpointedDrm[Long]

    Creates empty DRM with non-trivial height

  10. abstract def drmParallelizeWithRowIndices(m: Matrix, numPartitions: Int = 1)(implicit sc: DistributedContext): CheckpointedDrm[Int]

    Parallelize in-core matrix as the backend engine distributed matrix, using row ordinal indices as data set keys.

  11. abstract def drmParallelizeWithRowLabels(m: Matrix, numPartitions: Int = 1)(implicit sc: DistributedContext): CheckpointedDrm[String]

    Parallelize in-core matrix as the backend engine distributed matrix, using row labels as a data set keys.

  12. abstract def drmSampleKRows[K](drmX: DrmLike[K], numSamples: Int, replacement: Boolean = false): Matrix

  13. abstract def drmSampleRows[K](drmX: DrmLike[K], fraction: Double, replacement: Boolean = false): DrmLike[K]

    (Optional) Sampling operation.

    (Optional) Sampling operation. Consistent with Spark semantics of the same.

    K
    drmX
    fraction
    replacement
    returns

  14. abstract def indexedDatasetDFSRead(src: String, schema: Schema = DefaultIndexedDatasetReadSchema, existingRowIDs: Option[BiDictionary] = None)(implicit sc: DistributedContext): IndexedDataset

    Load IndexedDataset from text delimited format.

    Load IndexedDataset from text delimited format.

    src

    comma delimited URIs to read from

    schema

    defines format of file(s)

  15. abstract def indexedDatasetDFSReadElements(src: String, schema: Schema = ..., existingRowIDs: Option[BiDictionary] = None)(implicit sc: DistributedContext): IndexedDataset

    Load IndexedDataset from text delimited format, one element per line

    Load IndexedDataset from text delimited format, one element per line

    src

    comma delimited URIs to read from

    schema

    defines format of file(s)

  16. abstract def norm[K](drm: CheckpointedDrm[K]): Double

  17. abstract def numNonZeroElementsPerColumn[K](drm: CheckpointedDrm[K]): Vector

    Engine-specific numNonZeroElementsPerColumn implementation based on a checkpoint.

  18. abstract def toPhysical[K](plan: DrmLike[K], ch: CacheHint)(implicit arg0: ClassTag[K]): CheckpointedDrm[K]

    Second optimizer pass.

    Second optimizer pass. Translate previously rewritten logical pipeline into physical engine plan.

Concrete Value Members

  1. final def !=(arg0: AnyRef): Boolean

    Definition Classes
    AnyRef

  2. final def !=(arg0: Any): Boolean

    Definition Classes
    Any

  3. final def ##(): Int

    Definition Classes
    AnyRef → Any

  4. final def ==(arg0: AnyRef): Boolean

    Definition Classes
    AnyRef

  5. final def ==(arg0: Any): Boolean

    Definition Classes
    Any

  6. final def asInstanceOf[T0]: T0

    Definition Classes
    Any

  7. def clone(): AnyRef

    Attributes
    protected[java.lang]
    Definition Classes
    AnyRef
    Annotations
    @throws( ... )

  8. final def eq(arg0: AnyRef): Boolean

    Definition Classes
    AnyRef

  9. def equals(arg0: Any): Boolean

    Definition Classes
    AnyRef → Any

  10. def finalize(): Unit

    Attributes
    protected[java.lang]
    Definition Classes
    AnyRef
    Annotations
    @throws( classOf[java.lang.Throwable] )

  11. final def getClass(): Class[_]

    Definition Classes
    AnyRef → Any

  12. def hashCode(): Int

    Definition Classes
    AnyRef → Any

  13. final def isInstanceOf[T0]: Boolean

    Definition Classes
    Any

  14. final def ne(arg0: AnyRef): Boolean

    Definition Classes
    AnyRef

  15. final def notify(): Unit

    Definition Classes
    AnyRef

  16. final def notifyAll(): Unit

    Definition Classes
    AnyRef

  17. def optimizerRewrite[K](action: DrmLike[K])(implicit arg0: ClassTag[K]): DrmLike[K]

    First optimization pass.

    First optimization pass. Return physical plan that we can pass to exec(). This rewrite may introduce logical constructs (including engine-specific ones) that user DSL cannot even produce per se. <P>

    A particular physical engine implementation may choose to either use the default rewrites or build its own rewriting rules. <P>

  18. final def synchronized[T0](arg0: ⇒ T0): T0

    Definition Classes
    AnyRef

  19. def toString(): String

    Definition Classes
    AnyRef → Any

  20. final def wait(): Unit

    Definition Classes
    AnyRef
    Annotations
    @throws( ... )

  21. final def wait(arg0: Long, arg1: Int): Unit

    Definition Classes
    AnyRef
    Annotations
    @throws( ... )

  22. final def wait(arg0: Long): Unit

    Definition Classes
    AnyRef
    Annotations
    @throws( ... )

Inherited from AnyRef

Inherited from Any

Ungrouped