use fshl/fshr to implement bit rotate #3519

Job	Run time
Ubuntu / Release / gcc	21m 32s
macOS / Release / homebrew-clang	9m 17s
Ubuntu / Debug / gcc	17m 29s
macOS / Debug / homebrew-clang	7m 16s
Ubuntu / Release / clang	20m 6s
macOS / Release / system-clang	9m 1s
Ubuntu / Debug / clang	13m 54s
macOS / Debug / system-clang	4m 48s
Windows / Release / cl	30m 21s
Windows / Release / clang	27m 21s
Windows / Release / clang-cl	24m 39s
Windows / Debug / cl	26m 37s
Windows / Debug / clang	22m 52s
Windows / Debug / clang-cl	21m 14s
	4h 16m 27s

Provide feedback