Enable torch.compile validation on Amazon 2023

[atalman]

Which is currently failing: https://github.com/pytorch/test-infra/actions/runs/13973253199/job/39120358041

Test:

python3 smoke_test.py --package torchonly
torch: 2.7.0+cu126
ATen/Parallel:
	at::get_num_threads() : 8
	at::get_num_interop_threads() : 16
OpenMP 201511 (a.k.a. OpenMP 4.5)
	omp_get_max_threads() : 8
Intel(R) oneAPI Math Kernel Library Version 2024.2-Product Build 20240605 for Intel(R) 64 architecture applications
	mkl_get_max_threads() : 8
Intel(R) MKL-DNN v3.7.1 (Git Hash 8d263e693366ef8db40acc569cc7d8edf644556d)
std::thread::hardware_concurrency() : 16
Environment variables:
	OMP_NUM_THREADS : [not set]
	MKL_NUM_THREADS : [not set]
ATen parallel backend: OpenMP

Skip version check for channel None as stable version is None
Testing smoke_test_conv2d
Testing smoke_test_linalg on cpu
Testing smoke_test_compile for cpu and torch.float16
Testing smoke_test_compile for cpu and torch.float32
Testing smoke_test_compile for cpu and torch.float64
Picked CPU ISA VecAVX2 bit width 256
Testing smoke_test_compile with mode 'max-autotune' for torch.float32

[vercel]

@atalman is attempting to deploy a commit to the Meta Open Source Team on Vercel.

A member of the Team first needs to authorize it.

[atalman]

Please note, because of this issue Amazon 2023 its still failing: pytorch/pytorch#149681

Validated with local Amazon 2023 (if numpy is installed test is working)

[seemethere]

Are these things we recommend people to install when using torch.compile?

Want to level set that we're testing out what a user might actually be installing / have expectations to install

[atalman]

Error 1 if Development Tools are not installed:

compiler_info = get_compiler_version_info(get_cpp_compiler())
  File "/usr/local/lib64/python3.9/site-packages/torch/_inductor/cpp_builder.py", line 156, in get_cpp_compiler
    compiler = cpp_compiler_search(search)
  File "/usr/local/lib64/python3.9/site-packages/torch/_inductor/cpp_builder.py", line 102, in cpp_compiler_search
    raise exc.InvalidCxxCompiler
torch._inductor.exc.InductorError: InvalidCxxCompiler: No working C++ compiler found in torch._inductor.config.cpp.cxx: (None, 'g++')

Error 2 if python-devel is not installed:

Testing smoke_test_compile for cpu and torch.float16
/usr/local/lib64/python3.9/site-packages/torch/_inductor/cpp_builder.py:797: UserWarning: Can't find Python.h in /usr/include/python3.9
  warnings.warn(f"Can't find Python.h in {str(include_dir)}")
/usr/local/lib64/python3.9/site-packages/torch/_inductor/cpp_builder.py:797: UserWarning: Can't find Python.h in /usr/include/python3.9
  warnings.warn(f"Can't find Python.h in {str(include_dir)}")
Traceback (most recent call last):
  File "/pytorch/.ci/pytorch/smoke_test/smoke_test.py", line 430, in <module>
    main()
  File "/pytorch/.ci/pytorch/smoke_test/smoke_test.py", line 424, in main
    smoke_test_cuda(
  File "/pytorch/.ci/pytorch/smoke_test/smoke_test.py", line 222, in smoke_test_cuda
    smoke_test_compile("cuda" if torch.cuda.is_available() else "cpu")

[atalman]

Same problem with 2.6 release:
Error1:

torch._dynamo.exc.BackendCompilerFailed: backend='inductor' raised:
InvalidCxxCompiler: No working C++ compiler found in torch._inductor.config.cpp.cxx: (None, 'g++')

Error2:

/usr/local/lib64/python3.9/site-packages/torch/_subclasses/functional_tensor.py:275: UserWarning: Failed to initialize NumPy: No module named 'numpy' (Triggered internally at /pytorch/torch/csrc/utils/tensor_numpy.cpp:81.)
  cpu = _conversion_method_template(device=torch.device("cpu"))
/usr/local/lib64/python3.9/site-packages/torch/_inductor/cpp_builder.py:799: UserWarning: Can't find Python.h in /usr/include/python3.9
  warnings.warn(f"Can't find Python.h in {str(include_dir)}")
/usr/local/lib64/python3.9/site-packages/torch/_inductor/cpp_builder.py:799: UserWarning: Can't find Python.h in /usr/include/python3.9

[malfet]

Please mention the followup issue in the comment, but otherwise LGTM

[atalman]

Documentation update is tracked in pytorch/pytorch#149740