[SPARK-10491] [MLLIB] move RowMatrix.dspr to BLAS
jira: https://issues.apache.org/jira/browse/SPARK-10491 We implemented dspr with sparse vector support in `RowMatrix`. This method is also used in WeightedLeastSquares and other places. It would be useful to move it to `linalg.BLAS`. Let me know if new UT needed. Author: Yuhao Yang <hhbyyh@gmail.com> Closes #8663 from hhbyyh/movedspr.
This commit is contained in:
Yuhao Yang
Xiangrui Meng
parent
09b7e7c198
commit
c35fdcb7e9
|
|
@ -88,7 +88,7 @@ private[ml] class WeightedLeastSquares(
|
|||
if (fitIntercept) {
|
||||
// shift centers
|
||||
// A^T A - aBar aBar^T
|
||||
RowMatrix.dspr(-1.0, aBar, aaValues)
|
||||
BLAS.spr(-1.0, aBar, aaValues)
|
||||
// A^T b - bBar aBar
|
||||
BLAS.axpy(-bBar, aBar, abBar)
|
||||
}
|
||||
|
|
@ -203,7 +203,7 @@ private[ml] object WeightedLeastSquares {
|
|||
bbSum += w * b * b
|
||||
BLAS.axpy(w, a, aSum)
|
||||
BLAS.axpy(w * b, a, abSum)
|
||||
RowMatrix.dspr(w, a, aaSum.values)
|
||||
BLAS.spr(w, a, aaSum)
|
||||
this
|
||||
}
|
||||
|
||||
|
|
|
|||
|
|
@ -236,6 +236,50 @@ private[spark] object BLAS extends Serializable with Logging {
|
|||
_nativeBLAS
|
||||
}
|
||||
|
||||
/**
|
||||
* Adds alpha * x * x.t to a matrix in-place. This is the same as BLAS's ?SPR.
|
||||
*
|
||||
* @param U the upper triangular part of the matrix in a [[DenseVector]](column major)
|
||||
*/
|
||||
def spr(alpha: Double, v: Vector, U: DenseVector): Unit = {
|
||||
spr(alpha, v, U.values)
|
||||
}
|
||||
|
||||
/**
|
||||
* Adds alpha * x * x.t to a matrix in-place. This is the same as BLAS's ?SPR.
|
||||
*
|
||||
* @param U the upper triangular part of the matrix packed in an array (column major)
|
||||
*/
|
||||
def spr(alpha: Double, v: Vector, U: Array[Double]): Unit = {
|
||||
val n = v.size
|
||||
v match {
|
||||
case DenseVector(values) =>
|
||||
NativeBLAS.dspr("U", n, alpha, values, 1, U)
|
||||
case SparseVector(size, indices, values) =>
|
||||
val nnz = indices.length
|
||||
var colStartIdx = 0
|
||||
var prevCol = 0
|
||||
var col = 0
|
||||
var j = 0
|
||||
var i = 0
|
||||
var av = 0.0
|
||||
while (j < nnz) {
|
||||
col = indices(j)
|
||||
// Skip empty columns.
|
||||
colStartIdx += (col - prevCol) * (col + prevCol + 1) / 2
|
||||
col = indices(j)
|
||||
av = alpha * values(j)
|
||||
i = 0
|
||||
while (i <= j) {
|
||||
U(colStartIdx + indices(i)) += av * values(i)
|
||||
i += 1
|
||||
}
|
||||
j += 1
|
||||
prevCol = col
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* A := alpha * x * x^T^ + A
|
||||
* @param alpha a real scalar that will be multiplied to x * x^T^.
|
||||
|
|
|
|||
|
|
@ -24,7 +24,6 @@ import scala.collection.mutable.ListBuffer
|
|||
import breeze.linalg.{DenseMatrix => BDM, DenseVector => BDV, SparseVector => BSV, axpy => brzAxpy,
|
||||
svd => brzSvd, MatrixSingularException, inv}
|
||||
import breeze.numerics.{sqrt => brzSqrt}
|
||||
import com.github.fommil.netlib.BLAS.{getInstance => blas}
|
||||
|
||||
import org.apache.spark.Logging
|
||||
import org.apache.spark.SparkContext._
|
||||
|
|
@ -123,7 +122,7 @@ class RowMatrix @Since("1.0.0") (
|
|||
// Compute the upper triangular part of the gram matrix.
|
||||
val GU = rows.treeAggregate(new BDV[Double](new Array[Double](nt)))(
|
||||
seqOp = (U, v) => {
|
||||
RowMatrix.dspr(1.0, v, U.data)
|
||||
BLAS.spr(1.0, v, U.data)
|
||||
U
|
||||
}, combOp = (U1, U2) => U1 += U2)
|
||||
|
||||
|
|
@ -673,43 +672,6 @@ class RowMatrix @Since("1.0.0") (
|
|||
@Experimental
|
||||
object RowMatrix {
|
||||
|
||||
/**
|
||||
* Adds alpha * x * x.t to a matrix in-place. This is the same as BLAS's DSPR.
|
||||
*
|
||||
* @param U the upper triangular part of the matrix packed in an array (column major)
|
||||
*/
|
||||
// TODO: SPARK-10491 - move this method to linalg.BLAS
|
||||
private[spark] def dspr(alpha: Double, v: Vector, U: Array[Double]): Unit = {
|
||||
// TODO: Find a better home (breeze?) for this method.
|
||||
val n = v.size
|
||||
v match {
|
||||
case DenseVector(values) =>
|
||||
blas.dspr("U", n, alpha, values, 1, U)
|
||||
case SparseVector(size, indices, values) =>
|
||||
val nnz = indices.length
|
||||
var colStartIdx = 0
|
||||
var prevCol = 0
|
||||
var col = 0
|
||||
var j = 0
|
||||
var i = 0
|
||||
var av = 0.0
|
||||
while (j < nnz) {
|
||||
col = indices(j)
|
||||
// Skip empty columns.
|
||||
colStartIdx += (col - prevCol) * (col + prevCol + 1) / 2
|
||||
col = indices(j)
|
||||
av = alpha * values(j)
|
||||
i = 0
|
||||
while (i <= j) {
|
||||
U(colStartIdx + indices(i)) += av * values(i)
|
||||
i += 1
|
||||
}
|
||||
j += 1
|
||||
prevCol = col
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Fills a full square matrix from its upper triangular part.
|
||||
*/
|
||||
|
|
|
|||
|
|
@ -126,6 +126,31 @@ class BLASSuite extends SparkFunSuite {
|
|||
}
|
||||
}
|
||||
|
||||
test("spr") {
|
||||
// test dense vector
|
||||
val alpha = 0.1
|
||||
val x = new DenseVector(Array(1.0, 2, 2.1, 4))
|
||||
val U = new DenseVector(Array(1.0, 2, 2, 3, 3, 3, 4, 4, 4, 4))
|
||||
val expected = new DenseVector(Array(1.1, 2.2, 2.4, 3.21, 3.42, 3.441, 4.4, 4.8, 4.84, 5.6))
|
||||
|
||||
spr(alpha, x, U)
|
||||
assert(U ~== expected absTol 1e-9)
|
||||
|
||||
val matrix33 = new DenseVector(Array(1.0, 2, 3, 4, 5))
|
||||
withClue("Size of vector must match the rank of matrix") {
|
||||
intercept[Exception] {
|
||||
spr(alpha, x, matrix33)
|
||||
}
|
||||
}
|
||||
|
||||
// test sparse vector
|
||||
val sv = new SparseVector(4, Array(0, 3), Array(1.0, 2))
|
||||
val U2 = new DenseVector(Array(1.0, 2, 2, 3, 3, 3, 4, 4, 4, 4))
|
||||
spr(0.1, sv, U2)
|
||||
val expectedSparse = new DenseVector(Array(1.1, 2.0, 2.0, 3.0, 3.0, 3.0, 4.2, 4.0, 4.0, 4.4))
|
||||
assert(U2 ~== expectedSparse absTol 1e-15)
|
||||
}
|
||||
|
||||
test("syr") {
|
||||
val dA = new DenseMatrix(4, 4,
|
||||
Array(0.0, 1.2, 2.2, 3.1, 1.2, 3.2, 5.3, 4.6, 2.2, 5.3, 1.8, 3.0, 3.1, 4.6, 3.0, 0.8))
|
||||
|
|
|
|||
Loading…
Reference in a new issue