Traceback (most recent call last): File "/builds/software-machine-learning-infra-frameworks-workspaces-robhar02/pytorch/test/inductor/test_aot_inductor.py", line 1744, in test_quanatized_int8_linear self.check_model(Model(self.device), example_inputs) File "/builds/software-machine-learning-infra-frameworks-workspaces-robhar02/pytorch/test/inductor/test_aot_inductor_utils.py", line 249, in check_model expected = ref_model(*ref_inputs) File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1778, in _wrapped_call_impl return self._call_impl(*args, **kwargs) File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1789, in _call_impl return forward_call(*args, **kwargs) File "/builds/software-machine-learning-infra-frameworks-workspaces-robhar02/pytorch/test/inductor/test_aot_inductor.py", line 1729, in forward return torch.ops._quantized.wrapped_quantized_linear( File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/_ops.py", line 1275, in call return self._op(*args, **kwargs) RuntimeError: unknown architecure

To execute this test, run the following from the base repo dir: python test/inductor/test_aot_inductor_arrayref.py AOTInductorTestABICompatibleCpuWithStackAllocation.test_quanatized_int8_linear_cpu_with_stack_allocation

Fix Plan

The fix involves updating the torch.ops._quantized.wrapped_quantized_linear function to support AArch64 architecture.

• Update the torch.ops module to include AArch64 architecture support.
• Modify the wrapped_quantized_linear function to handle the new architecture.

Example code:

import torch

# Update torch.ops to support AArch64
torch.ops._quantized.register_module("wrapped_quantized_linear", 
                                     lambda *args, **kwargs: 
                                     torch.ops._quantized._wrapped_quantized_linear_aa(*args, **kwargs))

# Define the new _wrapped_quantized_linear_aa function
def _wrapped_quantized_linear_aa(input, weight, bias, scale, zero_point, dtype):
    # Implementation for AArch64 architecture
    # ...
    pass

Verification

To verify the fix, run the test case again:

python test/inductor/test_aot_inductor_arrayref.py AOTInductorTestABICompatibleCpuWithStackAllocation.test_quanatized_int8_linear_cpu_with_stack_allocation

If the test passes, the fix is successful.

Extra Tips

• Ensure that the torch.ops module is updated correctly to support the new architecture.
• Test the fix on different platforms to ensure compatibility.
• Consider adding additional logging or error handling to the wrapped_quantized_linear function to handle any potential issues.