pytorch

206da97b - [aot_inductor][easy] enable test_triton_kernel_multi_output_arg (#122052)

[aot_inductor][easy] enable test_triton_kernel_multi_output_arg (#122052) looks like we already support aoti_torch_cuda_sort in C shim. Pull Request resolved: https://github.com/pytorch/pytorch/pull/122052 Approved by: https://github.com/oulgen