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.

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

Jump to bottom

doc: Add memory tuning section to user guide #684

Merged
merged 3 commits into from
Jul 20, 2024
Merged

doc: Add memory tuning section to user guide #684

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
1fe9cc3
doc: Add memory tuning section to user guide
viirya Jul 18, 2024
a217ff7
Update docs/source/user-guide/tuning.md
viirya Jul 19, 2024
6388c03
More
viirya Jul 19, 2024
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
15 changes: 15 additions & 0 deletions docs/source/user-guide/tuning.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,6 +21,21 @@ under the License.

Comet provides some tuning options to help you get the best performance from your queries.

## Memory Tuning

Comet currently doesn't share the memory allocation from Spark but owns its own memory allocation.
That's said, Comet requires additional memory to be allocated. Comet provides some memory related configurations to help you tune the memory usage.

By default, the amount of memory is `spark.comet.memory.overhead.factor` * `spark.executor.memory`.
viirya marked this conversation as resolved.
Show resolved Hide resolved
The default value for `spark.comet.memory.overhead.factor` is 0.2. You can increase the factor to require more
memory for Comet to use, if you see OOM error. Generally, increasing memory overhead will improve the performance of your queries.
For example, some operators like `SortMergeJoin` and `HashAggregate` may require more memory to run.
Once the memory is not enough, the operator will spill to disk, which will slow down the query.

Besides, you can also set the memory explicitly by setting `spark.comet.memoryOverhead` to the desired value.
Comet will allocate at least `spark.comet.memory.overhead.min` memory.

If both `spark.comet.memoryOverhead` and `spark.comet.memory.overhead.factor` are set, the former will be used.

## Shuffle

Expand Down