tvm-0.8.0-bp154.2.20

- Add patch to fix boo#1197347:
  * tvm-do-not-force-synr-version.patch

- Skip test_meta_schedule_local_runner_time_out and
  test_meta_schedule_local_runner_exception tests on Leap/SLE
  since they require python 3.7+

- Add deps on python3-tvm and python3-setuptools for tvmc - boo#1196646

- Add 'Requires: python-cloudpickle' - boo#1195952#c1

- Update to v0.8
  * PaddlePaddle frontend
  * TVMScript: round-trippable python-based syntax for TIR
  * TorchScript integration
  * TensorIR scheduling language
  * TensorRT and CUTLASS integration via BYOC
  * Int4 TensorCore support in AutoTVM
  * MicroTVM Project API and Zephyr, Arduino support
  * AOT executor
  * Robust Windows support
  * Affine analysis infra: iter-affine-map
  * Improved Vulkan backend
  * CUDA graph support in TVM runtime
- Hopefully fixes boo#1195952
- Drop Patches merged upstream:
  * tvm-fix-catch.patch -- gh#apache/tvm#7319
  * tvm-fix-llvm12.patch -- gh#apache/tvm#6717, gh#apache/tvm#6738
- Add patches
  * tvm-fix-relay-test.patch -- gh#apache/tvm#10402
  * tvm-disable-vulkan-test-check.patch
- Skip Python 3.9+ -- gh#apache/tvm#8577

- Trim future plans/marketing from description.

- Build fails with LLVM 13, so still use LLVM 12 on Tumbleweed

- Backport patches to fix build on Tumbleweed with llvm12:
  * tvm-fix-llvm12.patch

- NEP 29: Last minorversion bump deprecated Python 3.6
  https://numpy.org/neps/nep-0029-deprecation_policy.html

- Exclude %arm %{ix86} as tests fail

- Skip tests requiring opencl
- Skip test requiring a specific arch

- Recommends onnx and PIL for tvmc

- Fix upstream tarball

- Disable LTO as it breaks runtime, see: https://github.com/apache/tvm/issues/6832
- Disable python tests as it does not find Vulkan/openCL

- Add patch to fix a warning with gcc10:
  * tvm-fix-catch.patch

- Add Arm Compute Library support

- Update to 0.7.0: https://github.com/apache/incubator-tvm/blob/main/NEWS.md
  v0.7 brings many major features. The community works together
  to refactor the internal code base to bring an unified IR code
  structure with a unified IRModule, type system and pass
  infrastructure. We have also bought many exciting new features,
  some highlights include:
  * Initial automatic scheduling support
  * Initial command line driver interface
  * WebGPU and webassembly support
  * Better first class rust support in the codebase
  * Intial Hexagon support
  * Bring your own codegen (BYOC) support
  The community also continues to bring high quality improvements
  to the existing modules including, but not limited to:
  better frontend coverage, performance, quantization, uTVM
  and dynamic shape support.
- Rebase pacth:
  * tvm-fix-openblas.patch
- Drop patch:
  * lib-finder-python-cmake.patch

- Enable onednn (formerly mkl-dnn) for aarch64 and ppc64le as well
- Add support to:
  * Antlr4
  * openBLAS
- Add patch to fix build with openblas enabled:
  * tvm-fix-openblas.patch

- Add _constraints with 4GB min disk space
- Use limit_build macro for all arches
- Exclude more tests for ppc64 & ppc64le

- Fix typo

- Exclude %arm %{ix86} as tests fail

- Skip tests requiring opencl
- Skip test requiring a specific arch

- Split cmake folder to not override it with python builds
- Fix detection of this folder in the python modules:
  * lib-finder-python-cmake.patch

- Add provides/obsoletes on TVM

- Build fails on Tumbleweed with llvm10, so use llvm9 - boo#1176220

- Format with spec-cleaner
- Expand properly the fdupes call
- Fix headers install
- Fix python pkg names
- Fix python requires

- Package nnvm and topi python packages

- mkl-dnn is not available on %{ix86}, so do not enable it there

- Make use of python Macros: %python_subpackages and %python_files

- Enable Python binding

- Enable UnitTests (only when LLVM is enabled as it is a requirement)