[NDTensors][Bug] Fix CPU performance issue

[kmp5VT]

Description

Joey found that contractions slowed down by roughly a factor of 5-10. It turned out that a permutedim was being evaluated lazily resulting in a complicated type that wasn't being parsed properly and as a result generic mul! was begin called. Here I create a function to unwrap a strided array to properly show its parenttype and I force evaluate of a permutation to ensure the strided version of mul is called.

Checklist:

• Fix Joeys CPU mul! slowdown

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (871e59d) 83.63% compared to head (aecad59) 67.47%.
Report is 1 commits behind head on main.

❗ Current head aecad59 differs from pull request most recent head 0e52496. Consider uploading reports for the commit 0e52496 to get more accurate results

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@             Coverage Diff             @@
##             main    #1218       +/-   ##
===========================================
- Coverage   83.63%   67.47%   -16.16%     
===========================================
  Files          89       88        -1     
  Lines        8445     8397       -48     
===========================================
- Hits         7063     5666     -1397     
- Misses       1382     2731     +1349     

see 55 files with indirect coverage changes

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.