-
Notifications
You must be signed in to change notification settings - Fork 706
perf: try out new cliping numba kernel #3952
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
base: main
Are you sure you want to change the base?
Conversation
ilan-gold
commented
Jan 23, 2026
- Trying out kernel from [RFC] Refactoring Scanpy's Numba Backend: C-Style Kernel Layer & Biology-Optimized Sparse Matrices #3944 (comment)
- Tests included or not required because:
- Release notes not necessary because:
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #3952 +/- ##
==========================================
+ Coverage 78.26% 78.35% +0.09%
==========================================
Files 117 117
Lines 12633 12637 +4
==========================================
+ Hits 9887 9902 +15
+ Misses 2746 2735 -11
Flags with carried forward coverage won't be shown. Click here to find out more.
|
Benchmark changes
Comparison: https://github.com/scverse/scanpy/compare/4c7c5c0e8bacc958370fcb0988495a6bf8410add..8b3f328748a0f2cc350509b14adb85fc9b2aa7b5 More details: https://github.com/scverse/scanpy/pull/3952/checks?check_run_id=61986886627 |
|
@ilan-gold I tried my old kernel that failed two years ago. It works now as well and might even be a bit faster than the proposed kernels, likely due to better memory coalescing per thread. Numba must have done some work on the parallel execution engine. |