[PATCH] D154496: InstCombine: Fold ldexp(ldexp(x, a), b) -> ldexp(x, a + b)

[Matt Arsenault via Phabricator via llvm-commits]

arsenm created this revision.
arsenm added reviewers: jcranmer-intel, foad, kpn, sepavloff, andrew.w.kaylor.
Herald added subscribers: StephenFan, hiraditya.
Herald added a project: All.
arsenm requested review of this revision.
Herald added a subscriber: wdng.
Herald added a project: LLVM.

The problem here is overflow or underflow which would have occurred in
the inner operation, which the exponent offsetting avoids. We can do
this if we know the two exponents are in the same direction, or
reassoc flags allow unsafe reassociates.


https://reviews.llvm.org/D154496

Files:
  llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp
  llvm/test/Transforms/InstCombine/ldexp.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D154496.537296.patch
Type: text/x-patch
Size: 9275 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20230705/97e088ba/attachment.bin>