Add a new NVTX range for task GPU ownership #11596
Open
Conversation
This file contains 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
jihoonson
changed the title
Signed-off-by: Jihoon Son <[email protected]>
jihoonson
force-pushed
the
semaphore-range
branch
from
89b6f6c
to
5400dfb
Compare
abellina
reviewed
Oct 17, 2024
sql-plugin/src/main/scala/com/nvidia/spark/rapids/RapidsConf.scala
Outdated
Show resolved
Hide resolved
abellina
reviewed
Oct 17, 2024
sql-plugin/src/main/scala/com/nvidia/spark/rapids/GpuSemaphore.scala
Outdated
Show resolved
Hide resolved
abellina
reviewed
Oct 17, 2024
sql-plugin/src/main/scala/com/nvidia/spark/rapids/GpuSemaphore.scala
Outdated
Show resolved
Hide resolved
abellina
reviewed
Oct 17, 2024
| @@ -258,6 +262,10 @@ private final class SemaphoreTaskInfo(val taskAttemptId: Long) extends Logging { | |||
| // We now own the semaphore so we need to wake up all of the other tasks that are | |||
| // waiting. | |||
| hasSemaphore = true | |||
| if (trackSemaphore) { | |||
| nvtxRange = | |||
| Some(new NvtxUniqueRange(s"Sem-${taskAttemptId}", NvtxColor.ORANGE)) | |||
There was a problem hiding this comment.
NvtxColor is a hex number in argb (alpha, red, green, blue). I think it would make this a lot more useable if task X hashed to a specific color, consistently. That way you can tell them apart while the task is alive. Bonus points to make it color match with the current Task ranges? That might make this even more useful.
There was a problem hiding this comment.
This seems a really good idea. I will try as you suggested. Thanks!
There was a problem hiding this comment.
Actually changed my mind, and want to try it later. It is a nice idea and I definitely want to try, but don't want to keep this PR open for unnecessarily long. Created #11645.
mythrocks
reviewed
Oct 17, 2024
sql-plugin/src/main/scala/com/nvidia/spark/rapids/GpuSemaphore.scala
Outdated
Show resolved
Hide resolved
mythrocks
previously approved these changes
Oct 17, 2024
sameerz
added
the
task
|
build |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
If you have ever been curious about the semaphore-based GPU concurrency control mechanism, you may have wondered at some point exactly what tasks are holding the semaphore at a given point of time during query processing. This PR can help you in that case. This PR adds a new NVTX range that shows what task owns GPU in the nsys profile result. This feature is off by default (since I don't think it's always useful), and can be enabled by setting
spark.rapids.sql.traceTaskGpuOwnership = true. The screenshot below shows an example nsys result with the new semaphore ranges. The orange boxes withSem-${taskAttemptId}represent the ranges in which each spark task was holding the semaphore.