Skip to content

[SPARK-7140][MLLIB] only scan the first 16 entries in Vector.hashCode #5697

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
mengxr wants to merge 3 commits into from
Closed

[SPARK-7140][MLLIB] only scan the first 16 entries in Vector.hashCode #5697

Changes from 1 commit
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
1ebad60
only scan the first 16 nonzeros in Vector.hashCode
mengxr Apr 25, 2015
8fb7d74
update impl
mengxr Apr 25, 2015
2abc86d
typo
mengxr Apr 26, 2015
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 7 additions & 2 deletions mllib/src/main/scala/org/apache/spark/mllib/linalg/Vectors.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -52,7 +52,7 @@ sealed trait Vector extends Serializable {

override def equals(other: Any): Boolean = {
other match {
case v2: Vector => {
case v2: Vector =>
if (this.size != v2.size) return false
(this, v2) match {
case (s1: SparseVector, s2: SparseVector) =>
Expand All @@ -63,20 +63,25 @@ sealed trait Vector extends Serializable {
Vectors.equals(0 until d1.size, d1.values, s1.indices, s1.values)
case (_, _) => util.Arrays.equals(this.toArray, v2.toArray)
}
}
case _ => false
}
}

override def hashCode(): Int = {
Copy link
Member

@jkbradley jkbradley Apr 25, 2015

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why is Vector.hashCode overridden, even though it will never be called (since Sparse & Dense Vector both override it)?

Copy link
Contributor Author

@mengxr mengxr Apr 26, 2015

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is just a reference implementation for subclasses to optimize.

var result: Int = size + 31
var i = 0
this.foreachActive { case (index, value) =>
Copy link
Contributor

@rxin rxin Apr 25, 2015

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

you don't mean to have a case here do you? just a closure not partial function

// ignore explict 0 for comparison between sparse and dense
if (value != 0) {
result = 31 * result + index
// refer to {@link java.util.Arrays.equals} for hash algorithm
val bits = java.lang.Double.doubleToLongBits(value)
result = 31 * result + (bits ^ (bits >>> 32)).toInt
i += 1
// only scan the first 16 nonzeros
if (i > 16) {
return result
Copy link
Contributor

@rxin rxin Apr 25, 2015

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

note that return in a closure is pretty expensive because it triggers an exception (which requires traversing the current stack). this might be much slower than before. you probably want to pass the max number of things to run into foreachActive.

Copy link
Contributor Author

@mengxr mengxr Apr 25, 2015

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

foreachActive is usually on the critical path. This hashCode implementation is only for the cases when the vector is used as a map key, as in the Pyrolite SerDe. I added specialized version for DenseVector and SparseVector.

}
}
}
result
Expand Down