Remove singular check for LU decomposition and add const qualifier to the CUDA kernels

[sjsprecious]

This PR:

• removes the singular check for LU decomposition (previously the GPU code is not checking the singularity correctly).
• adds the const qualifier to the CUDA kernels (it reduces the time of 64 tests (including all CUDA tests) to 245 seconds now!)

fix #664
fix #637

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 93.35%. Comparing base (c756a3d) to head (cc70d62).

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #671      +/-   ##
==========================================
+ Coverage   93.33%   93.35%   +0.02%     
==========================================
  Files          53       53              
  Lines        3603     3569      -34     
==========================================
- Hits         3363     3332      -31     
+ Misses        240      237       -3     

Flag	Coverage Δ
	93.35% <100.00%> (+0.02%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

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