Skip to content

Handle not overriden for TensorIndex #218

@naoyam

Description

@naoyam

🐛 Bug

Cloning a fusion fails as no override is provided for TensorIndex.

unknown file: Failure
C++ exception with description "false INTERNAL ASSERT FAILED at "../torch/csrc/jit/codegen/cuda/dispatch.h":201, please report a bug to PyTorch. Handle not overriden for TensorIndex.
Exception raised from handle at ../torch/csrc/jit/codegen/cuda/dispatch.h:201 (most recent call first):
frame #0: <unknown function> + 0xe9f8d (0x7f486b5caf8d in /home/nmaruyama/pytorch/src/csarofeen.3/build/lib/libc10.so)
frame #1: std::function<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > ()>::operator()() const + 0x4c (0x55ca79706afc in ./build/bin/test_jit)
frame #2: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >) + 0x40 (0x7f486b5ca228 in /home/nmaruyama/pytorch/src/csarofeen.3/build/lib/libc10.so)

To Reproduce

Add fusion.printKernel() before calling compileFusion. Tested with testGPU_FusionCodegen.

Metadata

Metadata

Assignees

Labels

No labels
No labels

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions