tqchen commented on PR #20350: URL: https://github.com/apache/tvm/pull/20350#issuecomment-5688863901
Validation for `be452a616f52f3ffcee84c5800bcc7fb7df1ab0e`: - LLVM/CUDA compiler build and all 138 existing C++ tests passed. - 139 existing tests passed across cache-index, buffer-access/LCA analysis, compaction, double buffering, allocation placement, storage rewrite, common-subexpression elimination, undefined variables, and tile-primitive dispatch/TRN transforms. - All 15 saved schedule diagnostics across detail/fast/none rendering still match the baseline byte-for-byte. - Full changed-file lint passed; the inapplicable TOML formatter remains skipped because no TOML files changed. Tile-primitive region operands now honor the typed BufferRegion hook, and DirectSubexpr safely narrows primitive expressions. Definition metadata traversal (including layout and allocated_addr uses) remains as requested. The visitor documentation explains the native-child requirement for preserving thrown error subclasses. GPU execution was not rerun in this focused pass; the previously reported driver limitation remains. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
