In some codes flushing the registers was costly: we processed each
register whereas all the registers alone in their equivalence class need
not to be processed. We now overapproximate easily which classes are of
size 2 so as to save many iterations in the Flush() loop in some cases.

Bug: v8:6432
Change-Id: I945e151736e8a515263ac76312127d930fd20d74
Reviewed-on: https://chromium-review.googlesource.com/525795
Commit-Queue: Alexandre Talon <alexandret@google.com>
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/master@{#45805}