[llvm] [InstCombine] Fold umax(nuw_mul(x, C0), x + 1) into (x == 0 ? 1 : nuw_mul(x, C0)) (PR #123468)
via llvm-commits
llvm-commits at lists.llvm.org
Sat Jan 18 16:36:47 PST 2025
================
@@ -1847,6 +1847,29 @@ Instruction *InstCombinerImpl::visitCallInst(CallInst &CI) {
return CastInst::Create(Instruction::ZExt, NarrowMaxMin, II->getType());
}
}
+ // If C0 is not 0:
+ // umax(nuw_shl(x, C0), x + 1) -> x == 0 ? 1 : nuw_shl(x, C0)
+ // If C0 is not 0 or 1:
+ // umax(nuw_mul(x, C0), x + 1) -> x == 0 ? 1 : nuw_mul(x, C0)
+ const APInt *C0;
+ auto matchShiftOrMul = [&](Value *I) {
+ return ((match(I, m_OneUse(m_NUWShl(m_Value(X), m_APInt(C0))))) ||
+ (match(I, m_OneUse(m_NUWMul(m_Value(X), m_APInt(C0)))) &&
----------------
Ruhung wrote:
Removed the restriction.
https://github.com/llvm/llvm-project/pull/123468
More information about the llvm-commits
mailing list