We have two forceExpandWideMUL functions. One takes the low and high
half of 2 inputs and calculates the low and high half of their product.
This does not calculate the full 2x width product.
The other signature takes 2 inputs and calculates the low and high half
of their full 2x width product. Previously it did this by sign/zero
extending the inputs to create the high bits and then calling the other
function.
We can instead copy the algorithm from the other function and use the
Signed flag to determine whether we should do SRA or SRL. This avoids
the need to multiply the high part of the inputs and add them to the
high half of the result. This improves the generated code for signed
multiplication.
This should improve the performance of #123262. I don't know yet how
close we will get to gcc.
Based on feedback from the clastb codegen PR, I'm refactoring basic codegen for the vector.extract.last.active intrinsic to lower to an ISD node in SelectionDAGBuilder then expand in LegalizeVectorOps, instead of doing everything in the builder.
The new ISD node (vector_find_last_active) only covers finding the index of the last active element of the mask, and extracting the element + handling passthru is left to existing ISD nodes.
I think SDNodeIterator primarily exists because GraphTraits requires an
iterator that dereferences to SDNode*. op_iterator dereferences to
SDUse* which is implicitly convertible to SDValue.
This piece of code can use SDValue instead of SDNode* so we should
prefer to use the the more common op_iterator.
Add DAG legalization support for expanding i1 SETCC nodes using
appropriate logical operations to simulate integer comparisons. Use
these expansions to handle i1 SETCC in NVPTX.
fixes#58428 and #57405
I want to use this function for GISel too so Type * is a better common
interface. All of the callers already convert EVT to Type * as needed
by calling lowering anyway.
This was overlooked in 7d940432c46be83b8fcb5dbefee439585fa820cd - when
inline assembly has multiple outputs, they are returned as members of a
struct, and the `getAsmOperandType` needs to be called for each member
of struct. The difference between this and the single-output case is
that in the latter, there isn't a struct wrapping the outputs.
I noticed this when trying to use the same mechanism in the RISC-V
backend.
Committing two tests:
- One that shows a crash before this change, which is fixed by this
change.
- One (commented out) that shows a different crash with tied
inputs/outputs. This is commented as it is not fixed by this change and
needs more work in target-independent inline asm handling code.
The dpbusd_const.ll test change is due to us losing the expanded add reduction pattern as one of the elements is known to be zero (removing one of the adds from the reduction pyramid). I don't think its of concern.
Noticed while working on #107423
Check to see if we are only demanding (shifted) signbits from a SRL node that are also signbits in the source node.
We can't demand any upper zero bits that the SRL will shift in (up to max shift amount), and the lower demanded bits bound must already be all signbits.
Same fold as #114389 which added this for SimplifyMultipleUseDemandedBits
When trying to evaluate an expression in a narrower type, the
DAGCombine should propagate the disjoint flag, as it's equally
valid on the narrower expression.
This helps improve better use of addressing modes for some
Arm SME instructions, for example.
Check to see if we are only demanding (shifted) signbits from a SRL node that are also signbits in the source node.
We can't demand any upper zero bits that the SRL will shift in (up to max shift amount), and the lower demanded bits bound must already be all signbits.
This fixes all the places that hit the new assertion added in
https://github.com/llvm/llvm-project/pull/106524 in tests. That is,
cases where the value passed to the APInt constructor is not an N-bit
signed/unsigned integer, where N is the bit width and signedness is
determined by the isSigned flag.
The fixes either set the correct value for isSigned, set the
implicitTrunc flag, or perform more calculations inside APInt.
Note that the assertion is currently still disabled by default, so this
patch is mostly NFC.
If SETCC or VSELECT is not legal for vector, we should not expand it,
instead we can split the vectors.
So that, some simple scale instructions can be emitted instead of
some pairs of comparation+selection.
For the purpose of verifying proper arguments extensions per the target's ABI,
introduce the NoExt attribute that may be used by a target when neither sign-
or zeroextension is required (e.g. with a struct in register). The purpose of
doing so is to be able to verify that there is always one of these attributes
present and by this detecting cases where sign/zero extension is actually
missing.
As a first step, this patch has the verification step done for the SystemZ
backend only, but left off by default until all known issues have been
addressed.
Other targets/front-ends can now also add NoExt attribute where needed and do
this check in the backend.
Unlike scalar, where AArch64 prefers expanding scmp/ucmp with select,
under Neon we can use the arithmetic expansion to generate fewer
instructions. Notably it also prevents the scalarization of vselect
during vector-legalization.
This is a follow-up to #92289 that adds lowering of the new
`@llvm.experimental.vector.compress` intrinsic on x86 with AVX512
instructions. This intrinsic maps directly to `vpcompress`.
GCC supports code like "asm volatile ("" : "=r" (i) : "0" (f))" where i
is integer type and f is floating point type. Currently this code
produces an error with Clang. The change allows mixed scalar types
between input and output constraints.
Co-authored-by: Matt Arsenault <Matthew.Arsenault@amd.com>
PR #80309 proposes to have users of APInt's uint64_t
constructor opt-in to implicit truncation. Currently, that patch
requires SelectionDAG::getConstant to opt-in.
This patch adds getSignedConstant so we can start fixing some of the
cases that require implicit truncation.
C23 introduced new functions fminimum_num and fmaximum_num, and they
follow the minimumNumber and maximumNumber of IEEE754-2019. Let's
introduce new intrinsics to support them.
This patch introduces support only support for scalar values. The
support of
vector (vp, vp.reduce, vector.reduce),
experimental.constrained
will be added in future patches.
With this patch, MIPSr6 and LoongArch can work out of box with
fcanonical and fmax/fmin.
Aarch64/PowerPC64 can use the same login as MIPSr6 and LoongArch, while
they have no fcanonical support yet.
I will add it in future patches.
The FMIN/FMAX of RISC-V instructions follows the
minimumNumber/maximumNumber of IEEE754-2019. We can just add it in
future patch.
Background
https://discourse.llvm.org/t/rfc-fix-llvm-min-f-and-llvm-max-f-intrinsics/79735
Currently we have fminnum/fmaxnum, which have different behavior on
different platform for NUM vs sNaN:
1) Fallback to fmin(3)/fmax(3): return qNaN.
2) ARM64/ARM32+Neon: same as libc.
3) MIPSr6/LoongArch/RISC-V: return NUM.
And the fix of fminnum/fmaxnum to follow minNUM/maxNUM of IEEE754-2008
will submit as separated patches.
If the upper bits of the shr aren't demanded.
This helps with cases where the outer srl was originally an sra and was
converted to a srl by SimplifyDemandedBits before it had a chance to
combine with the inner sra. This can occur when the inner sra was part
of a sign_extend_inreg expansion.
There are some regressions in ARM and Thumb2.
In TargetLowering::expandFixedPointMul when expanding fixed point
multiplication, and when using a widened MUL as strategy for the
lowering, there was a bug resulting in assertion failures like this:
Assertion `VT.isVector() == N1.getValueType().isVector() &&
"SIGN_EXTEND result type type should be vector iff the operand "
"type is vector!"' failed.
Problem was that we did not consider that VT could be a vector type
when setting up the WideVT. This patch should fix that bug.
Always match ABD patterns pre-legalization, and use TargetLowering::expandABD to expand again during legalization.
abdu(lhs, rhs) -> sub(xor(sub(lhs, rhs), usub_overflow(lhs, rhs)), usub_overflow(lhs, rhs))
Alive2: https://alive2.llvm.org/ce/z/dVdMyv
REAPPLIED: Fix regression issue with "abs(ext(x) - ext(y)) -> zext(abd(x, y))" fold failing after type legalization
