Update torch.norm to torch.linalg.norm and torch.linalg.vector_norm

[loadams]

• Update PR since torch.norm and torch.linalg.norm have different function signatures.
• Check if there are any numeric differences between the functions.
• Determine why there appear to be performance differences from others here.
• Update to torch.linalg.vectornorm
  Follow up PR handles these in the comm folder: Explicitly use the linalg.vector_norm call in comm/ #6960