[jvm-packages] ExtenralMemory: Overlap the caching time#11474
Open
wbo4958 wants to merge 3 commits intodmlc:masterfrom
Open
[jvm-packages] ExtenralMemory: Overlap the caching time#11474wbo4958 wants to merge 3 commits intodmlc:masterfrom
wbo4958 wants to merge 3 commits intodmlc:masterfrom
Conversation
trivialfis
reviewed
May 21, 2025
| count += 1 | ||
| Thread.sleep(50) | ||
| // If timeout, it's going to throw exception | ||
| val success = Await.result(futureOpt.get, 6.seconds) |
Member
There was a problem hiding this comment.
6 seconds is not a lot when data is large and system is busy, why do we need timeout here? Can writing to disk hang?
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This PR tries to overlap the caching time (caching cudf table to the disk when enabling external memory).
The previous design is
So, they are running in sequence which is time-consuming.
This PR moves the caching process to run in a separate thread, which can be overlapped when building QuantileDMatrix.
I did some test locally with total 13.1G dataset,
With this PR, it's going to eliminate the 90%+ caching time.