vortex-captain opened a new pull request, #691: URL: https://github.com/apache/tvm-ffi/pull/691
## Summary - add a native Windows ARM64 wheel job and enable ARM64 in cibuildwheel - select the MSVC developer environment from the running Python architecture so JIT extensions are built for ARM64 rather than x64 - omit the test-only PyTorch dependency on Windows ARM64 because official PyTorch wheels are not currently available for that platform PyTorch is not a runtime dependency of tvm-ffi. Its integration tests continue to run on CI platforms where official wheels are available; Windows ARM64 runs the remaining Python and C++ coverage. ## Validation - Windows ARM64 publish job: 465/465 C++ tests passed - CPython 3.12 ARM64: 2414 passed, 44 skipped, 3 xfailed - CPython 3.13 ARM64: 2416 passed, 42 skipped, 3 xfailed - CPython 3.14t ARM64: 2418 passed, 40 skipped, 3 xfailed - Windows AMD64, Linux x86_64/ARM64, and macOS ARM64 wheel jobs passed - regular lint, docs, Windows AMD64, Linux x86_64/ARM64, and macOS ARM64 CI passed A local private PyTorch 2.14 Windows ARM64 wheel was also tested with the released tvm-ffi 0.1.12 ARM64 wheel. The suite reported 2300 passed, 17 skipped, and 4 xfailed. Only these two optional addon-build tests failed: - `test_build_torch_c_dlpack_extension` - `test_parallel_build` Those failures are due to tvm-ffi 0.1.12 compiling the addon as C++17 while the PyTorch 2.14 headers require C++20. Current main already selects C++20 for PyTorch 2.13 and newer. -- 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]
