Refactor extraction of tuning policy selectors#8975
Conversation
📝 WalkthroughSummary by CodeRabbit
WalkthroughThree device algorithms (adjacent difference, merge, merge-sort) refactor their dispatch mechanisms from explicit ChangesEnvironment-based dispatch refactoring for adjacent difference, merge, and merge-sort
Comment |
There was a problem hiding this comment.
Actionable comments posted: 1
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Enterprise
Run ID: dc6a4ef3-3b13-47e5-b4d3-eff83e9bb37b
📒 Files selected for processing (5)
cub/cub/device/device_adjacent_difference.cuhcub/cub/device/device_merge.cuhcub/cub/device/device_merge_sort.cuhcub/cub/device/dispatch/dispatch_adjacent_difference.cuhcub/cub/device/dispatch/dispatch_merge.cuh
😬 CI Workflow Results🟥 Finished in 2h 21m: Pass: 37%/283 | Total: 4d 08h | Max: 2h 20m | Hits: 37%/354677See results here. |
I realized that
dispatch_with_env_and_tuningis not a good default mechanism for handling the policy selector extraction and defaulting, since it will lead to a lot of redundancy for setting up the default policy selector.In this PR I am providing two alternatives: