Fix mul cuda for bool (#47031)
Summary:
Also, add tests for tensor by scalar multiplication / division
Fixes https://github.com/pytorch/pytorch/issues/47007
Pull Request resolved: https://github.com/pytorch/pytorch/pull/47031
Reviewed By: walterddr
Differential Revision: D24608874
Pulled By: malfet
fbshipit-source-id: 4e15179904814d6e67228276d3d11ff1b5d15d0d
