tqchen commented on PR #20350: URL: https://github.com/apache/tvm/pull/20350#issuecomment-5688571555
Validation for `ad2dbf31b2d34f3eb038445fb455f836b9f5820e`, rebased onto main `319b760d27c61cf69b682f6d5975c98b1085094a`: - LLVM/CUDA compiler build and all 138 existing C++ tests passed. - Existing IR/TE/Relax tests: 500 passed, 10 expected failures, 1 baseline-reproduced unexpected pass. - S-TIR analysis/schedule/transform/meta-schedule tests: 1,012 passed, 17 skipped, 2 expected failures. - Backend codegen/layout tests: 422 passed, 378 skipped. - TIRx CUDA/IKET tests: 70 passed, 164 skipped, 18 failed. All 18 failure messages reproduce identically on the untouched baseline: one existing conflicting PTX launch-bounds failure and 17 architecture-detection failures caused by unavailable GPU driver access. GPU execution remains unverified on this head. - All 15 scheduling diagnostics across detail, fast, and none render levels match the baseline byte-for-byte. - Full changed-file lint passed. The TOML-only formatter was skipped because no TOML files changed; all applicable hooks ran. The Tensor structural leaf hooks already exist unchanged in the baseline. Five redundant TensorLoad overrides were removed; ExprSideEffect retains its required read-effect bookkeeping and delegates traversal to the base. -- 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]
