tree afc65d8cf268122bd8012d6b73ee0ed4f089f8c8
parent caacfffc71931797c85d8176ba7eb78195d4072f
author Marek Sedláček <mar3k.sedlacek@gmail.com> 1759935992 +0200
committer GitHub <noreply@github.com> 1759935992 +0000
gpgsig -----BEGIN PGP SIGNATURE-----
 
 wsFcBAABCAAQBQJo5n34CRC1aQ7uu5UhlAAA2L8QAA7yWAW+sYjLbEYIo8/7DCK3
 xIfjPAJhn2M3dpx5Opyo6BzunmfI9+DYuZCtsxvWZhOlvV8oUsmyzMnCBX2k/HU5
 5ThRDcGHVp7O8IKkMCyaAlR4BGNLV4k93MC4GPKvpf6vHdqKlEQ+1mVbhzi/6Gjt
 WXY1VLmsEPdQBD5dwBV6oMVE5AbohBzizvOtuHf/2sRDQHuP0H6hyTTDC1STMdxj
 ocuBFg3TyEBU86g8QhZWOZyBW/hqgj6dcUQs6nx/oJXgfUsx3fQw5HGkfwy36aiL
 w+MC9D0V+jY8JPw/QR3wq9e6EIp1hm+XeV82JJuQcZTohpzVCe6CWb6fF+G1Zn6k
 1nJJPhEJv6FP9ze9DPd3ZMC0tlYSCh8r7SKQCt9bmIlFntrxhB1UBLB7qpUeq4Qc
 5xlrqtoycvQ8gvf0lLoso8uAQ8GdXX03Qp4IkC9zrrBxtu3Jf1EVrpFJTK1n8xtk
 V03eLKY6Ql/eKRXA4lP3zQF8Sh8cdyDkQy3sTKNhYNNCV871L8jtCTDgO9xvupq2
 pRT4tPQdmhAstYldPY0K1vmzjh11He2XM0xuTHQI3PX5NRPlreyXbdaNadDhGnCp
 jRSumdASgdtd7huBhZVdiK3jd5fxy07OabTQTuyfu/x69TXzaTyhxCaYGLKoKkfD
 Ki4Wjs14h0kphkWvd8RK
 =aG2d
 -----END PGP SIGNATURE-----
 

[LoopRotate] Remove canRotateDeoptimizingLatchExit and multi-rotate (#162482)

This patch removes canRotateDeoptimizingLatchExit check from loop roate
and connected multi rotation option for loop roate (-loop-rotate-multi
option).

The heuristic in canRotateDeoptimizingLatchExit returns true if any of
the loop exits are non-deoptimizing. This means if the loop has multiple
deopt exits, then without multi-rotate, we may still end up having deopt
exit at the latch. This multi-rotate option is introduced but it is off
by default. 

We have not seen any improvements downstream as well, where
we have frequent number of deoptimizations in IR. Since the original
heuristic without the multi-rotate effectively ends up being "rotate
under some conditions", but the loop may still not be in the form we
want, we should remove the heuristic and multi-rotate framework entirely

Note that the diff is big, but it's just removal of while loop and
indentation change.

After this patch I would like to continue here and propose adding a
computability check for exit count, but that will be in a separate PR.

Requests for review: @annamthomas @fhahn @davemgreen

Co-authored-by: Marek Sedlacek <msedlacek@azul.com>