Don't terminate the whole process when Compile fails

[will-cromar]

See #6700

The error won't kill your whole notebook anymore:

>>> import torch_xla as xla >>> import torch >>> torch.dot(torch.tensor([2, 3], device=xla.device()), torch.tensor([2, 1], device=xla.device())) WARNING:root:PJRT is now the default runtime. For more information, see https://github.com/pytorch/xla/blob/master/docs/pjrt.md WARNING:root:libtpu.so and TPU device found. Setting PJRT_DEVICE=TPU. WARNING: All log messages before absl::InitializeLog() is called are written to STDERR I0000 00:00:1709942790.379654 501354 pjrt_api.cc:100] GetPjrtApi was found for tpu at /usr/local/lib/python3.8/site-packages/libtpu/libtpu.so I0000 00:00:1709942790.379741 501354 pjrt_api.cc:79] PJRT_Api is set for device type tpu I0000 00:00:1709942790.379751 501354 pjrt_api.cc:146] The PJRT plugin has PJRT API version 0.40. The framework PJRT API version is 0.40. Traceback (most recent call last): File "<stdin>", line 1, in <module> File "/usr/local/lib/python3.8/site-packages/torch/_tensor.py", line 464, in __repr__ return torch._tensor_str._str(self, tensor_contents=tensor_contents) File "/usr/local/lib/python3.8/site-packages/torch/_tensor_str.py", line 697, in _str return _str_intern(self, tensor_contents=tensor_contents) File "/usr/local/lib/python3.8/site-packages/torch/_tensor_str.py", line 432, in _str_intern self = self.to("cpu") RuntimeError: Bad StatusOr access: UNIMPLEMENTED: While rewriting computation to not contain X64 element types, XLA encountered an HLO for which this rewriting is not implemented: %dot.3 = s64[] dot(s64[2]{0} %p1.2, s64[2]{0} %p0.1), lhs_contracting_dims={0}, rhs_contracting_dims={0}