Fold
binop(shift(ShiftedC1, ShAmt), shift(ShiftedC2, add(ShAmt, AddC)))
->
shift(binop(ShiftedC1, shift(ShiftedC2, AddC)), ShAmt)
where both shifts are the same and AddC is a valid shift amount.
Proofs: https://alive2.llvm.org/ce/z/PhVVeg
Differential Revision: https://reviews.llvm.org/D152927
If `Mask` and `Amt` are not constants and `binop1` and `binop2` are
the same we can transform to:
`(binop (lshift (binop X, Y), Amt), Mask)`
If `binop` is `add`, `lshift` must be `shl`.
If `Mask` and `Amt` are constants `C` and `C1` respectively.
We can transform to:
`(lshift1 (binop1 (binop2 X, (inv_lshift1 C, C1), Y)), C1)`
Saving an instruction IFF:
`lshift1` is same opcode as `lshift2`
Either `bitwise1` and/or `bitwise2` is `and`.
Proofs(1/2): https://alive2.llvm.org/ce/z/BjN-m_
Proofs(2/2): https://alive2.llvm.org/ce/z/bZn5QB
This is to help fix the regression caused in D151807
Reviewed By: nikic
Differential Revision: https://reviews.llvm.org/D152568
Unfortunately alive2 cannot prove the correctness due to fails by timeout even for
float type half.
However it should be correct. If a and b are not NaN, maximum and minimum will just
return different values (a and b) and take into account a + b == b + a this is the same.
If a or b is NaN, than maximum and minimum are equal to NaN and NaN + NaN is NaN.
a + b is also a NaN.
In terms of preserving fast flags, we cannot preserve ninf due to
minimum(NaN, Infinity) == maximum(NaN, Infinity) == NaN,
minimum(NaN, Infinity) +ninf maximum(NaN, Infinity) == NaN +ninf NaN = NaN
However transformation will change
minimum(NaN, Infinity) + maximum(NaN, Infinity) to NaN +ninf Infinity == poison.
But if fadd is marked as nnan, we can preserve because NaN +ninf/nnan NaN = poison as well.
The same optimization for
maximum(a,b) * minimum(a,b) => a * b
is added.
All said above for fadd is correct for fmul.
Reviewed By: mkazantsev
Differential Revision: https://reviews.llvm.org/D147299
Revert commit due to failure on buildbot:
error: 'match_combine_or' may not intend to support class template argument deduction
This reverts commit b86a06ef284f2637bef89bf5bb20157a8b195568.
(C - X) + Y --> (Y - X) + C
Moving the constant operand to an 'add' gives more
flexibility to subsequent reassociation patterns,
and it may be better for codegen on targets that
don't have subtract-from-immediate instructions.
This artifact can appear from the vectorizer. (add X, -1) is the
backedge taken count. It gets zero extended and then 1 is added to
it to get the trip count.
There is usually a dominating branch that rules out X being zero.
Alive: https://alive2.llvm.org/ce/z/NsRDwX
This transform was added with 68c197f07eeae71b9b7,
and the post-commit review noted the potential
for miscompiles at narrow bitwidths.
I'm not sure how to expose the i1 nuw bug because we
already simplify that, but other cases show that
there are missing transforms to add in follow-up
patches.
(X * X) - (Y * Y) --> (X + Y) * (X - Y)
https://alive2.llvm.org/ce/z/BAuRCf
The no-wrap propagation could be relaxed in some cases,
but there does not seem to be an obvious rule for that.
(A s>> (BW - 1)) + (zext (A s> 0)) --> (A s>> (BW - 1)) | (zext (A != 0))
https://alive2.llvm.org/ce/z/V-nM8N
This is not the form that we currently match as m_Signum(),
but I'm not sure if one is better than the other, so there's
a follow-up patch needed either way.
For this patch, it should be better for analysis to use a
not-null test and bitwise logic rather than >0 with add.
Codegen doesn't seem significantly different on any targets
that I looked at.
Also note that none of these variants is shown in issue #60012 -
those generally include at least one 'select', so that's likely
where these patterns will end up.
It's a bigger match than usual, but I have not found any
sub-patterns that reduce:
(X / DivC) + sext ((X & (SMin | (DivC - 1)) >u SMin) --> X >>s log2(DivC)
https://alive2.llvm.org/ce/z/MJzlhl
Fixes issue #55741
We were conservatively intersecting flags, but we can take the union here
because forbidden special values (nsz/nnan/ninf) are not altered by fneg.
So if they were guaranteed not present on the select or fneg, then they
are guaranteed not present on the new select. Alive2 appears to agree on
the test diffs (reduced to not include irrelevant flags like reassoc):
https://alive2.llvm.org/ce/z/ViqqrO
This prevents a potential regression if we tighten up the FMF behavior
for fabs with NAN as suggested in issue #59279.
`(A * -2**C) + B --> B - (A << C)`
https://alive2.llvm.org/ce/z/A6BWkf
This inverts what Negator was doing before:
D134310 / 0f32a5dea0e9
Analysis and codegen are generally better without multiply,
so we should favor this form even if we trade add for sub
(because those are generally equivalent cost operations).
Fixes#57531
This transformation may be particularly useful on x86-64,
because x & (x - 1) can be performed by a single blsr instruction.
Differential Revision: https://reviews.llvm.org/D133362
LLVM contains a helpful function for getting the size of a C-style
array: `llvm::array_lengthof`. This is useful prior to C++17, but not as
helpful for C++17 or later: `std::size` already has support for C-style
arrays.
Change call sites to use `std::size` instead.
Differential Revision: https://reviews.llvm.org/D133429
This pattern is handled more generally in SimplifySelectsFeedingBinaryOp().
Tests to confirm that added to the add.ll test file in the previous commit.
https://alive2.llvm.org/ce/z/j_8Wz9
The arithmetic shift was converted to logical shift with:
246078604c871d43
That does not seem to uncover any other missing/conflicting folds,
so convert directly to signbit test + cast.
We still need to fold the pattern with logical shift to test + cast.
This allows reducing patterns where the output type is not
the same as the input value:
https://alive2.llvm.org/ce/z/nydwFVFixes#57394
~(A * C1) + A --> (A * (1 - C1)) - 1
This is a non-obvious mix of bitwise logic and math:
https://alive2.llvm.org/ce/z/U7ACVT
The pattern may be produced by Negator from the more typical
code seen in issue #57255.
Canonicalize ((x + C1) & C2) --> ((x & C2) + C1) for suitable constants
C1 and C2, instead of the other way round. This should allow more
constant ADDs to be matched as part of addressing modes for loads and
stores.
Differential Revision: https://reviews.llvm.org/D130080
We manage to iteratively achieve this result with no extra
uses, and the reassociate pass can also do this, but this
pattern falls through the cracks in the example from
issue #57053.
Replace ConstantExpr:getFAdd etc with call to
ConstantFoldBinaryOpOperands(). I'm using the constant folding API
rather than IRBuilder here to ensure that this does actually
constant fold. These transforms don't use m_ImmConstant(), so this
would not otherwise be guaranteed (and apparently, they can't use
m_ImmConstant because they want to handle scalable vector splats).
There is an opportunity here to further migrate these to the
ConstantFoldFPInstOperands() API, which would respect the denormal
mode. I've held off on doing so here, because some of this code
explicitly checks for denormal results, and I don't want to touch
it in a mostly NFC change.
