[GLUTEN-9163][VL] Separate compression buffer and disk write buffer configuration#9356
Merged
marin-ma merged 3 commits intoapache:mainfrom Apr 23, 2025
Merged
Conversation
|
Run Gluten Clickhouse CI on x86 |
marin-ma
force-pushed
the
shuffle-compression-config
branch
from
22ad109 to
b514a4d
Compare
|
Run Gluten Clickhouse CI on x86 |
|
Run Gluten Clickhouse CI on x86 |
Contributor
Author
|
@zhouyuan Could you help to review? Thanks! |
|
Run Gluten Clickhouse CI on x86 |
Contributor
Author
|
@zhouyuan Could you help to review? Thanks! |
zhouyuan
approved these changes
Apr 23, 2025
| GlutenShuffleUtils.getSortEvictBufferSize(sparkConf, compressionCodec); | ||
| GlutenShuffleUtils.getCompressionBufferSize(sparkConf, compressionCodec); | ||
| diskWriteBufferSize = | ||
| (int) (long) sparkConf.get(package$.MODULE$.SHUFFLE_DISK_WRITE_BUFFER_SIZE()); |
Member
There was a problem hiding this comment.
the code is little diffcult to understand, is it necessary to cast to long and then cast to int
Contributor
Author
There was a problem hiding this comment.
In Spark's source code, the configurations are converted in this way. Here's an explanation apache/spark#24187 (comment)
If we don't convert to long first , it will encounter exception like this:
Caused by: java.lang.ClassCastException: java.lang.Long cannot be cast to java.lang.Integer
marin-ma
added a commit
to marin-ma/gluten
that referenced
this pull request
Jul 16, 2025
warrenzhu25
pushed a commit
to warrenzhu25/gluten
that referenced
this pull request
Jan 10, 2026
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
A follow-up to #9278
spark.shuffle.spill.diskWriteBufferSizeis used for setting the buffer size before spill to store the sorted rows. Spiller will write the data in this buffer to the output stream.spark.io.compression.lz4.blockSize,spark.io.compression.zstd.bufferSizeare used to set the compression buffer size in the compressed output stream, depending on which compression codec is set.The memory allocation of two buffers in spark are counted into overhead memory, so we use arrow::default_memory_pool to allocate them.
Add
spark.gluten.sql.columnar.shuffle.sort.deserializerBufferSize: Buffer size in bytes for sort-based shuffle reader deserializing raw input to columnar batch.