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.

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

[Core] Adding option to avoid Plasma Fetch and Deserialisation + e2e benchmarks #6

Open
wants to merge 9 commits into
base: master
Choose a base branch
from

Conversation

alindkhare
Copy link
Owner

Why are these changes needed?

Related issue number

Checks

  • I've run scripts/format.sh to lint the changes in this PR.
  • I've included any doc changes needed for https://docs.ray.io/en/master/.
  • I've made sure the tests are passing. Note that there might be a few flaky tests, see the recent failures at https://flakey-tests.ray.io/
  • Testing Strategy
    • Unit tests
    • Release tests
    • This PR is not tested :(

Add GPU num_gpus=1 for replicas

Intermediate tensor is torch tensor

Only 1 iteration is enough for calculating throughput and closed loop latency

Image as raw bytes

Removed the need to pass args as list for enqueue remote

Added server code

Error Fix

Added server code

Fix bug for subprocess submission

Few more print fixes

Fire more queries to see
Final config

Final Config
Comment on lines 2499 to 2507
memory_store_->GetAsync(
object_id, [python_future, success_callback, fallback_callback, object_id,
fetch_plasma_data](std::shared_ptr<RayObject> ray_object) {
if (ray_object->IsInPlasmaError() && fetch_plasma_data) {
fallback_callback(ray_object, object_id, python_future, fetch_plasma_data);
} else {
success_callback(ray_object, object_id, python_future);
}
});
Copy link
Owner Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@atumanov This is the main change that I made

@alindkhare
Copy link
Owner Author

Performance Benchmark Results

router_comparison

Latency Table

Implementation [95th, 99th, 100th] ms latency percentile
Vanilla Ray Serve [ 58.30999687 91.6196993 135.20691544]
Reference [44.97625791 51.53332099 88.14558759]
Reference + Pipeline Orch. [46.11008726 55.16686797 97.10958973]
Ray Hack + Callbacks + Pipeline Orch. [42.77230315 45.28134219 50.52455887]

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant