We no longer require dst == src (output = input[0]) in all cases, only
when AVX is not supported. This can help remove an extra move when AVX
is supported. Also in many cases (when input[0] is an immediate), we
require less temporary registers.

Bug: v8:11589
Change-Id: I0d272df12de54f55b4c7a0a330c38ccaca82e927
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3092553Reviewed-by: Clemens Backes <clemensb@chromium.org>
Commit-Queue: Zhi An Ng <zhin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#76286}