Skip to content
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.

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

Optimize Record._elements to not duplicate VectorMap if possible #4254

Merged
merged 1 commit into from
Jul 10, 2024

Conversation

jackkoenig
Copy link
Contributor

@jackkoenig jackkoenig commented Jul 9, 2024

The code is a little janky, but it is a massive reduction in memory use.

For a typical Bundle with 7 UInt fields, this reduces the memory footprint from 1992 bytes to 1552 bytes (the duplicate VectorMap of only 7 elements is 440 bytes!), a 22% reduction. Combined with my other recent performance improvement PRs (#4251, #4252, #4253 but excluding #4242), this 1992 bytes is reduced to 1424 for a total reduction of 28.5% memory use for this typical Bundle with 7 UInt fields.

I suspect we can do better by getting rid of any Map here at all and instead just use an Array[Data] (where the String names of the fields are stored in the children Data themselves as they are already stored there). Whereas the VectorMap of 7 elements is 440 bytes, an Array of 7 elements is 48 bytes.

That being said, this is a much simpler change that imparts a huge benefit so is worth it in the meantime.

Contributor Checklist

  • Did you add Scaladoc to every public function/method?
  • Did you add at least one test demonstrating the PR?
  • Did you delete any extraneous printlns/debugging code?
  • Did you specify the type of improvement?
  • Did you add appropriate documentation in docs/src?
  • Did you request a desired merge strategy?
  • Did you add text to be included in the Release Notes for this change?

Type of Improvement

  • Performance improvement

Desired Merge Strategy

  • Squash

Release Notes

This reduces memory use of a typical bundle by 20%.

Reviewer Checklist (only modified by reviewer)

  • Did you add the appropriate labels? (Select the most appropriate one based on the "Type of Improvement")
  • Did you mark the proper milestone (Bug fix: 3.6.x, 5.x, or 6.x depending on impact, API modification or big change: 7.0)?
  • Did you review?
  • Did you check whether all relevant Contributor checkboxes have been checked?
  • Did you do one of the following when ready to merge:
    • Squash: You/ the contributor Enable auto-merge (squash), clean up the commit message, and label with Please Merge.
    • Merge: Ensure that contributor has cleaned up their commit history, then merge with Create a merge commit.

@jackkoenig jackkoenig added the Performance Improves performance, will be included in release notes label Jul 9, 2024
@jackkoenig jackkoenig added this to the 6.x milestone Jul 9, 2024
val originalElements = elements
// Don't create a new map unless necessary, this is much more memory efficient in common case
// This is true if elements is not a VectorMap or if any names need sanitization
var needNewMap = !originalElements.isInstanceOf[VectorMap[_, _]]
Copy link
Contributor

Choose a reason for hiding this comment

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

is there a reason why this has to be done with a var and also why we aren't just doing it something like:

def makeNewMap(newNames: Seq[String) =       newNames.view
        .zip(originalElements)
        .map { case (name, (_, data)) => name -> data }
        .to(VectorMap) // VectorMap has O(1) lookup whereas ListMap is O(n)

val newElements = originalElements match {
  case original: VectorMap[_, _] => {
    val newNames = ...
    val allNamesUnchanged = newNames == origNames
    if (allNamesUnchanged) original else makeNewMap(newNames)
  case _ => makeNewMap(newNames)
}

...? I guess you are trying to only iterate over it once ...?

Copy link
Contributor Author

Choose a reason for hiding this comment

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

I guess you are trying to only iterate over it once ...?

Correct, I'm trying to minimize the total number of traversals

Copy link
Contributor

@adkian-sifive adkian-sifive left a comment

Choose a reason for hiding this comment

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

LGTM, the changes make sense

@jackkoenig jackkoenig merged commit 0df3ec9 into main Jul 10, 2024
18 checks passed
@jackkoenig jackkoenig deleted the jackkoenig/microoptimize-record-elements branch July 10, 2024 17:05
@mergify mergify bot added the Backported This PR has been backported label Jul 10, 2024
mergify bot pushed a commit that referenced this pull request Jul 10, 2024
chiselbot pushed a commit that referenced this pull request Jul 10, 2024
…) (#4265)

(cherry picked from commit 0df3ec9)

Co-authored-by: Jack Koenig <[email protected]>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Backported This PR has been backported Performance Improves performance, will be included in release notes
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

3 participants