This reverts commit e8b3ffa532b8ebac5dcdf17bb91b47817382c14d.
The AMDGPU/mad_64_32.ll seems to fail on some of the build bots but
passes locally. I'm really confused.
(sra X, BW-1) is either 0 or -1. So the multiply is a conditional
negate of Y.
This pattern shows up when type legalizing wide multiplies involving
a sign extended value.
Fixes PR57549.
Reviewed By: RKSimon
Differential Revision: https://reviews.llvm.org/D133399
MachineInstr's copy constructor works by calling the addOperand method
to add each operand of the old MachineInstr to the new one, one by
one. But addOperand deliberately avoids trying to replicate ties
between operands, on the grounds that the tie refers to operands by
index, and the indices aren't necessarily finalized yet.
This led to a code generation fault when the machine pipeliner cloned
an Arm conditional instruction, and lost the tie between the output
register and the input value to be used when the condition failed to
execute.
Reviewed By: dmgreen
Differential Revision: https://reviews.llvm.org/D135434
This reverts commit 0148df8157f05ecf3b1064508e6f012aefb87dad.
Getting a lit test failures on AMDGPU but I can't reproduce it so far.
Reverting to investigate.
(sra X, BW-1) is either 0 or -1. So the multiply is a conditional
negate of Y.
This pattern shows up when type legalizing wide multiplies involving
a sign extended value.
Fixes PR57549.
Reviewed By: RKSimon
Differential Revision: https://reviews.llvm.org/D133399
pipeliner-preserve-ties.mir demonstrates a current bug in which the
output of the Modulo Software Pipelining pass has left off a tie
between operands in the conditional `t2ADDri` instruction. It should
look like this:
%19:rgpr = t2ADDri %1, 1, 1 /* CC::ne */, $cpsr, $noreg, implicit %1(tied-def 0)
in which the final input operand is tied to the output, because that's
the input that will become the output value if the conditionalized add
instruction does not execute, and hence, must necessarily be whatever
was in the output register beforehand.
In the input to the pipeliner, those `tied-def` specifications are
present and correct. But when the pipeliner clones MachineInstrs, it
loses them.
pipeliner-inlineasm.mir does not demonstrate any bug: the output is
already correct, because of compensation code in the machine pipeliner
that applies only to INLINEASM instructions. But no test previously
exercised that code, so I add one now before making changes in that
area.
The `CodeGenPrepare` pass can sink bitwise `and` used by compare to
zero into the basic blocks where the users are. This operation is
guarded by lowering hook, which is disabled for ARM. In the ARM
architecture versions from v7-M up these two operations can be folded
into `tst rN, #imm` instruction. Sinking of `and` can also enable
the cmov-to-bfi DAG combiner.
This patch fixes some benchmark regressions caused
by https://reviews.llvm.org/D129370 as well scoring slightly better overall.
Reviewed By: dmgreen
Differential Revision: https://reviews.llvm.org/D134360
The method of counting resource consumption is modified to be based on
"Cycles" value when DFA is not used.
The calculation of ResMII is modified to total "Cycles" and divide it
by the number of units for each resource. Previously, ResMII was
excessive because it was assumed that resources were consumed for
the cycles of "Latency" value.
The method of resource reservation is modified similarly. When a
value of "Cycles" is larger than 1, the resource is considered to be
consumed by 1 for cycles of its length from the scheduled cycle.
To realize this, ResourceManager maintains a resource table for all
slots. Previously, resource consumption was always 1 for 1 cycle
regardless of the value of "Cycles" or "Latency".
In addition, the number of micro operations per cycle is modified to
be constrained by "IssueWidth". To disable the constraint,
--pipeliner-force-issue-width=100 can be used.
For the case of using DFA, the scheduling results are unchanged.
Reviewed By: dpenry
Differential Revision: https://reviews.llvm.org/D133572
This was only trying this to relax register class constraints, but
this can also help if there are subranges involved.
This solves a compilation failure for AMDGPU when there is high
pressure created by large register tuples. If one virtual register is
using most of the available budget, we need to be able to evict
subranges.
This solves the immediate failure, but this solution leaves a lot to
be desired. In the relevant testcases, we have 32-element tuples but
most of the uses are operations on 1 element subranges of it. What
we're now getting is a spill and restore of the full 1024 bits and an
extract of the used 32-bits. It would be far better if we introduced a
copy to a new virtual register with a smaller register class and used
narrower spills.
Furthermore, we could probably do a better job if the allocator were
to introduce new subranges where none previously existed in the
highest pressure scenarios. The block and region splits should also
try to split specific subranges out.
The mve-vst3.ll test changes looks like noise to me, but instruction
count increased by one. mve-vst4.ll looks like a solid improvement
with several 16-byte spills eliminated. splitkit-copy-live-lanes.mir
also shows a solid reduction in total spill count.
This could use more tests but it's pretty tiring to come up with cases
that fail on this.
This interface allows a target to reject a proposed
SMS schedule. For Hexagon/PowerPC, all schedules
are accepted, leaving behavior unchanged. For ARM,
schedules which exceed register pressure limits are
rejected.
Also, two RegisterPressureTracker methods now need to be public so
that register pressure can be computed by more callers.
Reapplication of D128941/(reversion:D132037) with small fix.
Differential Revision: https://reviews.llvm.org/D132170
This reverts commit 8c4aea438c310816bb4e4f9a32d783381ef3182e.
Needed because buildbot failures (warnings) gave a clue that there was
a functional bug in the ARM rejection logic.
Reviewed By: dmgreen
Differential Revision: https://reviews.llvm.org/D132037
This interface allows a target to reject a proposed
SMS schedule. For Hexagon/PowerPC, all schedules
are accepted, leaving behavior unchanged. For ARM,
schedules which exceed register pressure limits are
rejected.
Also, two RegisterPressureTracker methods now need to be public so
that register pressure can be computed by more callers.
Reviewed By: dmgreen
Differential Revision: https://reviews.llvm.org/D128941
FoldConstantArithmetic can fold constant vectors hidden behind bitcasts (e.g. vXi64 -> v2Xi32 on 32-bit platforms), but currently bails if either vector contains undef elements. These undefs can often occur due to SimplifyDemandedBits/VectorElts calls recognising that the upper bits are often unnecessary (e.g. funnel-shift/rotate implicit-modulo and AND masks).
This patch adds a basic 'FoldValueWithUndef' handler that will attempt to constant fold if one or both of the ops are undef - so far this just handles the AND and MUL cases where we always fold to zero.
The RISCV codegen increase is interesting - it looks like the BUILD_VECTOR lowering was loading a constant pool entry but now (with all elements defined constant) it can materialize the constant instead?
Differential Revision: https://reviews.llvm.org/D130839
issue #56775
I rearranged the Thumb2 codegen test to avoid simplifying the chain
of rounding instructions. I'm assuming the intent of the test is
to verify lowering of each of those intrinsics.
SimplifyDemandedBits currently early-outs for multi-use values beyond the root node (just returning the knownbits), which is missing a number of optimizations as there are plenty of cases where we can still simplify when initially demanding all elements/bits.
@lenary has confirmed that the test cases in aea-erratum-fix.ll need refactoring and the current increase codegen is not a major concern.
Differential Revision: https://reviews.llvm.org/D129765
Given a patch like D129506, using instructions not valid for the current
target feature set becomes an error. This fixes an issue in
ARMExpandPseudo::ExpandCMP_SWAP where Thumb2 compares were used in
Thumb1Only code, such as thumbv8m.baseline targets.
Differential Revision: https://reviews.llvm.org/D129695
If none of the bits of a VBICimm are demanded, we can remove the node
entirely using the input operand instead.
Differential Revision: https://reviews.llvm.org/D129966
These instructions are only available when fp is available, so cannot be
used with just +mve. Add predicates to ensure we fall-back under the
right circumstances.
If we have a variable shift amount and the demanded mask has leading
zeros, we can propagate those leading zeros to not demand those bits
from operand 0. This can allow zero_extend/sign_extend to become
any_extend. This pattern can occur due to C integer promotion rules.
This transform is already done by InstCombineSimplifyDemanded.cpp where
sign_extend can be turned into zero_extend for example.
Reviewed By: spatel, foad
Differential Revision: https://reviews.llvm.org/D121833
If the add has more than one use then applying the transformation
won't cause it to be removed, so we can end up applying it again
causing an infinite loop.
Differential Revision: https://reviews.llvm.org/D129361
D89489 added some logic to the interleaved access pass to attempt to
undo the folding of shuffles into binops, that instcombine performs. If
early-cse is run too, the binops may be commoned into a single operation
with multiple shuffle uses. It is still profitable reverse the transform
though, so long as all the uses are shuffles.
Differential Revision: https://reviews.llvm.org/D129419
This fixes the combining of constant vector GEP operands in the
optimization of MVE gather/scatter addresses, when opaque pointers are
enabled. As opaque pointers reduce the number of bitcasts between geps,
more can be folded than before. This can cause problems if the index
types are now different between the two geps.
This fixes that by making sure each constant is scaled appropriately,
which has the effect of transforming the geps to have a scale of 1,
changing [r0, q0, uxtw #1] gathers to [r0, q0] with a larger q0. This
helps use a simpler instruction that doesn't need the extra uxtw.
Differential Revision: https://reviews.llvm.org/D127733
Although this doesn't usually come up, we can have uses of the
BaseAccess of a distributed postinc being a PHI. This doesn't need the
usual dominance check as we will dominate along the phi edge, allowing
us to still create a postinc load/store.
Differential Revision: https://reviews.llvm.org/D127676
The included test hits a verifier problems as one of the instructions:
```
%113:tgpreven, %114:tgprodd = MVE_VMLSLDAVas16 %12:tgpreven(tied-def 0), %11:tgprodd(tied-def 1), %7:mqpr, %8:mqpr, 0, $noreg, $noreg
```
Has two inputs that come from different PHIs with the same base reg, but
conflicting regclasses:
```
%11:tgprodd = PHI %103:gpr, %bb.1, %16:gpr, %bb.2
%12:tgpreven = PHI %103:gpr, %bb.1, %17:gpr, %bb.2
```
The MachinePipeliner would attempt to use %103 for both the %11 and %12
operands in the prolog, constraining the register class to the common
subset of both. Unfortunately there are no registers that are both odd
and even, so the second constrainRegClass fails. Fix this situation by
inserting a COPY for the second if the call to constrainRegClass fails.
The register allocation can then fold that extra copy away. The register
allocation of Q regs changed with this test, but the R regs were the
same and no new instructions are needed in the final assembly.
Differential Revision: https://reviews.llvm.org/D127971
This reverts commit d9ef307e9bb3b636a18c4051a236f1aafd7600e6 as it is
causeing expensive check verification errors. Remove the test again
until we can fix them.
We already have patterns for matching fadd(select(..., -0.0)),
but an upcoming patch will lead to patterns using +0.0 as the
identity instead of -0.0. I'm adding support for these patterns
now to avoid any regressions for MVE.
Differential Revision: https://reviews.llvm.org/D127275
This matches what we do in IR. For the RISC-V test case, this allows
us to use -8 for the AND mask instead of materializing a constant in a register.
Reviewed By: spatel
Differential Revision: https://reviews.llvm.org/D127335
Adds MVT::v128i2, MVT::v64i4, and implied MVT::i2, MVT::i4.
Keeps MVT::i2, MVT::i4 lowering actions as expand, which should be
removed once targets set this explicitly.
Adjusts 11 lit tests to reflect slightly different behavior during
DAG combine.
Differential Revision: https://reviews.llvm.org/D125247
Adds MVT::v128i2, MVT::v64i4, and implied MVT::i2, MVT::i4.
Keeps MVT::i2, MVT::i4 lowering actions as `expand`, which should be
removed once targets set this explicitly.
Adjusts 11 lit tests to reflect slightly different behavior during
DAG combine.
Differential Revision: https://reviews.llvm.org/D125247
As discussed on D77804, instcombine will have already performed a similar SimplifyMultipleUseDemandedBits call which will break the UXTB16 pattern that was being match in these DAG tests
I've updated the existing tests so that it match the instcombine IR (with a suitable FIXME) and added an equivalent test pattern suggested by @dmgreen
- Add t2LoopEnd to TargetInstrInfo::analyzeBranch and
related functions. As there are many side effects of
analyzing a branch, only do so if software pipelining
is enabled to maintain previous behavior when pipelining
is not desired.
- Make sure that t2LoopEndDec is immediately followed by
a t2B when it is synthesized from a t2LoopEnd. This is
done because the t2LoopEnd might have acquired a
fall-through path, but IfConversion assumes that
fall-through are only possible on analyzable branches.
Reviewed By: dmgreen
Differential Revision: https://reviews.llvm.org/D126322
The `vcvtb.f16.f32 Sd, Sn` (and vcvtt.f16.f32) instruction convert a f32
into a f16, writing either the top or bottom halves of the register.
That means that half of the input register Sd is used in the output.
This wasn't being modelled in the instructions, leading later analyses
to believe that the registers were dead where they were not, generating
invalid scheduling
Fix that be specifying the input Sda register for the instructions too,
allowing them to be set for cases like vector inserts. Most of the
changes are plumbing through the constraint string, cstr.
Differential Revision: https://reviews.llvm.org/D126118
Pulled out of D77804 as its going to be easier to address the regressions individually.
This patch allows SimplifyDemandedBits to call SimplifyMultipleUseDemandedBits in cases where the source operand has other uses, enabling us to peek through the shifted value if we don't demand all the bits/elts.
The lost RISCV gorc2 fold shouldn't be a problem - instcombine would have already destroyed that pattern - see https://github.com/llvm/llvm-project/issues/50553
Differential Revision: https://reviews.llvm.org/D124839
If the mask is made up of elements that form a mask in the higher type
we can convert shuffle(bitcast into the bitcast type, simplifying the
instruction sequence. A v4i32 2,3,0,1 for example can be treated as a
1,0 v2i64 shuffle. This helps clean up some of the AArch64 concat load
combines, along with helping simplify a number of other tests.
The PowerPC combine for v16i8 splat vector loads needed some fixes to
keep it working for v16i8 vectors. This improves the handling of v2i64
shuffles to match too, hopefully improving them in general.
Differential Revision: https://reviews.llvm.org/D123801
This adds a big endian run line for the AArch64 TRN tests and
regenerated the check lines, along with adding an extra MVE VMOVN case
and regenerating vector-DAGCombine.ll for easier updating.
The reasoning for creating VSHLIMM/VSHRsIMM/VSHRuIMM nodes in a combine
- because matching i64 constants is difficult - does not apply for MVE,
as there are not v2i64 shifts. Delaying the creation of the nodes can
allow extra transforms on target independant shl/shr.
