leeexyz opened a new pull request #10325: URL: https://github.com/apache/tvm/pull/10325
Currently, `cache_write` does not support the split reduce axis. This PR tries to support it. ```bash E 2: tvm::te::Schedule::cache_write(tvm::te::Tensor const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&) E at ../src/te/schedule/schedule_dataflow_rewrite.cc:457 E 1: tvm::te::CacheWriteWithReLayout(tvm::te::Schedule, tvm::runtime::Array<tvm::te::Tensor, void> const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&) E at ../src/te/schedule/schedule_dataflow_rewrite.cc:299 E 0: void tvm::te::PrepareAxisMapping<tvm::te::ComputeOpNode const>(tvm::te::Stage, tvm::te::ComputeOpNode const*, std::unordered_set<tvm::tir::IterVar, std::hash<tvm::tir::IterVar>, std::equal_to<tvm::tir::IterVar>, std::allocator<tvm::tir::IterVar> >*, tvm::runtime::Array<tvm::tir::IterVar, void>*, std::unordered_map<tvm::tir::IterVar, tvm::Range, std::hash<tvm::tir::IterVar>, std::equal_to<tvm::tir::IterVar>, std::allocator<std::pair<tvm::tir::IterVar const, tvm::Range> > >*, std::unordered_map<tvm::tir::VarNode const*, tvm::PrimExpr, std::hash<tvm::tir::VarNode const*>, std::equal_to<tvm::tir::VarNode const*>, std::allocator<std::pair<tvm::tir::VarNode const* const, tvm::PrimExpr> > >*, std::unordered_map<tvm::tir::VarNode const*, tvm::PrimExpr, std::hash<tvm::tir::VarNode const*>, std::equal_to<tvm::tir::VarNode const*>, std::allocator<std::pair<tvm::tir::VarNode const* const, tvm::PrimExpr> > >*, std::vector<tvm::PrimExpr, std::allocator<tvm::PrimExpr> >*) E at ../src/te/schedule/schedule_dataflow_rewrite.cc:217 E File "../src/te/schedule/schedule_dataflow_rewrite.cc", line 217 E TVMError: E --------------------------------------------------------------- E An error occurred during the execution of TVM. E For more information, please see: https://tvm.apache.org/docs/errors.html E --------------------------------------------------------------- E E Check failed: iv->iter_type == kDataPar (2 vs. 0) : Can only relayout with in data parallel dimensions ``` -- 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]
