Version 14: Add shared-memory tiles - #10

Merged

pbelevich merged 1 commit into

mainfrom

version-14-shared-memory-tiles

Aug 23, 2026

pbelevich commented Aug 23, 2026

Owner

Summary

recognize canonical masked matrix loads feeding tl.dot
cooperatively stage both dot operands in shared memory and synchronize the block
preserve externally used SSA dependencies and validate layout and shared-memory limits
cover direct, AST frontend, and runtime K-loop staging paths

Testing

make (283 passed)


          Version 14: Add shared-memory tiles

c7e5f1c

pbelevich merged commit af00e9c into main

4 checks passed

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet