[TORCH] Support aten.logaddexp & aten.logaddexp2 ops via torch decomposition

[zahidwx]

aten.logaddexp & aten.logaddexp2 ops are not supported in torch dialect and there is no decomposition for such ops to enable lowering to backend contract.

TORCH_VERSION_FOR_COMPARISON = 2.8.0.dev20250512
Running tests sequentially with progress status
*** RUNNING TEST: LogaddexpModule_basic ***
Compiling LogaddexpModule_basic...
Running LogaddexpModule_basic...

====================

TorchFX IR
module {
  func.func @LogaddexpModule(%arg0: !torch.vtensor<[3,4],f32>, %arg1: !torch.vtensor<[3,4],f32>) -> !torch.vtensor<[3,4],f32> {
    %0 = torch.operator "torch.aten.logaddexp"(%arg0, %arg1) : (!torch.vtensor<[3,4],f32>, !torch.vtensor<[3,4],f32>) -> !torch.vtensor<[3,4],f32> 
    return %0 : !torch.vtensor<[3,4],f32>
  }
}

loc("/home/zahidw/torch-mlir/build/tools/torch-mlir/python_packages/torch_mlir/torch_mlir_e2e_test/test_suite/basic.py":97:0): error: failed to legalize operation 'torch.operator' that was explicitly marked illegal
TORCH_VERSION_FOR_COMPARISON = 2.8.0.dev20250512
loc("/home/zahidw/torch-mlir/build/tools/torch-mlir/python_packages/torch_mlir/torch_mlir_e2e_test/test_suite/basic.py":97:0): error: failed to legalize operation 'torch.operator' that was explicitly marked illegal
FAIL - "LogaddexpModule_basic"

Unexpected outcome summary: (fx_importer)

****** Failed tests - 1 tests
    FAIL - "LogaddexpModule_basic"
        Runtime error: Traceback (most recent call last):
          File "/home/zahidw/torch-mlir/build/tools/torch-mlir/python_packages/torch_mlir/torch_mlir_e2e_test/framework.py", line 354, in compile_and_run_test
            trace = config.run(compiled, golden_trace)
                    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
          File "/home/zahidw/torch-mlir/build/tools/torch-mlir/python_packages/torch_mlir/torch_mlir_e2e_test/configs/fx_importer_backend.py", line 61, in run
            self._export_run(artifact, trace)
          File "/home/zahidw/torch-mlir/build/tools/torch-mlir/python_packages/torch_mlir/torch_mlir_e2e_test/configs/fx_importer_backend.py", line 140, in _export_run
            module = fx.export_and_import(
                     ^^^^^^^^^^^^^^^^^^^^^
          File "/home/zahidw/torch-mlir/build/tools/torch-mlir/python_packages/torch_mlir/torch_mlir/fx.py", line 124, in export_and_import
            return _module_lowering(
                   ^^^^^^^^^^^^^^^^^
          File "/home/zahidw/torch-mlir/build/tools/torch-mlir/python_packages/torch_mlir/torch_mlir/fx.py", line 61, in _module_lowering
            run_pipeline_with_repro_report(
          File "/home/zahidw/torch-mlir/build/tools/torch-mlir/python_packages/torch_mlir/torch_mlir/compiler_utils.py", line 127, in run_pipeline_with_repro_report
            raise TorchMlirCompilerError(trimmed_message) from None
        torch_mlir.compiler_utils.TorchMlirCompilerError: Lowering TorchFX IR -> Torch Backend IR failed with the following diagnostics:


        python exception: Failure while executing pass pipeline

        For Torch-MLIR developers, the error can be reproduced with:
        $ torch-mlir-opt -pass-pipeline='builtin.module(func.func(torch-match-quantized-custom-ops), torchdynamo-export-to-torch-backend-pipeline{backend-legal-ops=aten.adaptive_max_pool2d extra-library=})' /tmp/UnnammedModule.mlir
        Add '-mlir-print-ir-after-all -mlir-disable-threading' to get the IR dump for debugging purpose.



Summary:
    Failed: 1

Minimal repro steps:

import torch
import torch.nn as nn
from torch_mlir import fx

def run(f):
    print(f"{f.__name__}")
    print("-" * len(f.__name__))
    f()
    print()

@run
def test_logaddexp():
    class LogaddexpModule(nn.Module):
        def __init__(self):
            super().__init__()

        def forward(self, x, y):
            return torch.logaddexp(x,y)

    x = torch.randn(3, 4)
    y = torch.randn(3, 4)
    exported = fx.export_and_import(LogaddexpModule(), x, y, output_type="torch")
    print(exported)