Device max memory watermark tracking #2392
Merged
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
Signed-off-by: Zach Puller <zpuller@nvidia.com>
Signed-off-by: Zach Puller <zpuller@nvidia.com>
Signed-off-by: Zach Puller <zpuller@nvidia.com>
Signed-off-by: Zach Puller <zpuller@nvidia.com>
Signed-off-by: Zach Puller <zpuller@nvidia.com>
Signed-off-by: Zach Puller <zpuller@nvidia.com>
|
build |
1 similar comment
|
build |
revans2
reviewed
Sep 19, 2024
| @@ -1360,6 +1375,11 @@ class spark_resource_adaptor final : public rmm::mr::device_memory_resource { | |||
| } | |||
| transition(thread->second, thread_state::THREAD_RUNNING); | |||
| thread->second.is_cpu_alloc = false; | |||
| // num_bytes is likely not padded, which could cause slight inaccuracies | |||
| // but for now it shouldn't matter for watermark purposes | |||
There was a problem hiding this comment.
Please make sure that we document this in the metrics page when it is fully ready
There was a problem hiding this comment.
Which page are you referring to exactly?
There was a problem hiding this comment.
Sorry it is an external page https://docs.nvidia.com/spark-rapids/user-guide/latest/tuning-guide.html#metrics
Signed-off-by: Zach Puller <zpuller@nvidia.com>
revans2
reviewed
Sep 23, 2024
Signed-off-by: Zach Puller <zpuller@nvidia.com>
revans2
previously approved these changes
Sep 23, 2024
|
build |
|
build |
|
build |
revans2
approved these changes
Sep 24, 2024
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Outdated suggestions cannot be applied.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.
Unblocks NVIDIA/spark-rapids#11457
Adds tracking of the max memory allocated on device over the lifespan of each task, and exposes this metric so that the Spark RAPIDS plugin can hook it up to a Spark accumulator.