-
Notifications
You must be signed in to change notification settings - Fork 232
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
[BUG]The metrics "total time“ of GpuColumnarExchange is strange #952
Comments
some question about metics in scan node
for scan node
Thanks |
@GaryShen2008 I looked into this and I agree that the value we are currently reporting for total time for shuffles is confusing (it is actually measuring the time to create an internal iterator which is misleading). Given that Spark doesn't report this metric for shuffles I propose that we remove this metric. I will create a PR. |
@JustPlay apologies for the delay, here's some answers for your querstions.
The latter is a standard Spark metric for scans that is measuring time at a very high level for the node, i.e.: the |
…IDIA#952) Signed-off-by: spark-rapids automation <70000568+nvauto@users.noreply.github.com>
Describe the bug
When using the Legacy Shuffle, the SQL plan showed as below.
The "total time total" is less than the "shuffle write time" + "fetch wait time".
It may confuse the user.
Steps/Code to reproduce bug
Run any ETL app with the legacy shuffle.
Expected behavior
Make the metrics name correct or find a way to show the correct value of total time.
Environment details (please complete the following information)
Additional context
None
The text was updated successfully, but these errors were encountered: