benchmark

043b07d4 - Add Inductor torch.compile benchmark for jagged_layer_norm operator

Add Inductor torch.compile benchmark for jagged_layer_norm operator Summary: Add `torch.compile` benchmark to `jagged_layer_norm` operator in TritonBench to compare the performance of Inductor-generated Triton kernels against existing TritonBench benchmarks, both in PyTorch and Triton. Reviewed By: davidberard98 Differential Revision: D60072951 fbshipit-source-id: b0803269b16e3aff136a989b8489f969164e51de