[fused_rmsnorm] Avoid conditional on dynamic stride

[wconstab]

Stack from ghstack (oldest at bottom):

• 301

• -> #300

• 161

The conditional expression forces evaluation of a symbolic quantity, which is not a problem in eager but prevents tracing/export and blocks using Pipeline Parallelism via tracing frontend.

Tensor.contiguous operator already handles returning self if the striding of self is already contiguous.

Fixes https://github.com/pytorch/PiPPy/issues/1108

[wconstab]

i think i abandoned this becuase even with this fix the PP tracer would not work due to other issues with the fused-rmsnorm custom op registration.

did we figure out a solution there? or are we totally giving up on fused_rmsnorm and switching to compile?

cc @lessw2020

[wconstab]

abandoned