Teach combineVectorSizedSetCCEquality() to handle arbitrary memcmp
expansions but do not change any default policy for now.
This also fixes a bug in the memcmp expansion itself when large
displacements are needed.
https://reviews.llvm.org/D69507
Enable the new SelectionDAG representation for unordered loads and stores introduced in r371441 by default. As a reminder, the new lowering changes the representation of an unordered atomic load from an AtomicSDNode - which is essentially a black box which gets passed through without combines messing with it - to a LoadSDNode w/a atomic marker on the MMO. The later parallels the way we handle volatiles, and I've audited the code to ensure that every location which checks one checks the other.
This has been fairly heavily fuzzed, and I examined diffs in a reasonable large corpus of assembly by hand, so I'm reasonable sure this is correct for the common case. Late in the review for this, it was discovered that I hadn't correctly handled cases which could be legalized into CAS operations. This points out that there's a strong bias in the IR of the frontend I'm working with towards only legal atomics. If there are problems with this patch, the most likely area will be legalization.
Differential Revision: https://reviews.llvm.org/D69219
This catches some cases. There are probably ways to improve this.
I tried doing it as a combine on the setcc, but that broke
some cases involving flag reuse in place of test.
I renamed the isX86CCUnsigned to isX86CCSigned and flipped its
polarity to make it consistent with the similar functions for
ISD::SETCC. This avoids calling EQ/NE as being signed or unsigned.
Fixes PR43823.
Differential Revision: https://reviews.llvm.org/D69499
The legalization of v2i1->i2 or v4i1->i4 bitcasts followed by a setcc can create an and after the bitcast. If we're lucky enough that the input to the bitcast is a concat_vectors where the first operand is a setcc that can natively 0 all the upper bits of ak-register, then we should replace the other operands of the concat_vectors with zero in order to remove the AND.
With the AND removed we might be able to use a kortest on the result.
Differential Revision: https://reviews.llvm.org/D69205
PTEST and especially the MOVMSK instructions are slow on Knights Landing
or later. As a bonus, this patch increases instruction parallelism by
emitting:
KORTEST(PCMPNEQ(a, b), PCMPNEQ(c, d)) == 0
Instead of:
KORTEST(AND(PCMPEQ(a, b), PCMPEQ(c, d))) == ~0
https://reviews.llvm.org/D69157
tryToWidenViaDuplication lowers using the shuffle_v8i16(unpack_v16i8(shuffle_v8i16(x),shuffle_v8i16(x))) pattern, but the unpack only needs the even/odd 16i8 args if the original v16i8 shuffle mask references the even/odd elements - which isn't true for many extension style shuffles.
llvm-svn: 375342
We were always generating a single source HADDPD, but really we should only do this if shouldUseHorizontalOp says its a good idea.
Differential Revision: https://reviews.llvm.org/D69175
llvm-svn: 375341
Add generic DAG combine for extending masked loads.
Allow us to generate sext/zext masked loads which can access v4i8,
v8i8 and v4i16 memory to produce v4i32, v8i16 and v4i32 respectively.
Differential Revision: https://reviews.llvm.org/D68337
llvm-svn: 375085
Exposes an issue in getFauxShuffleMask where the OR(SHUFFLE,SHUFFLE) decode should always resolve zero/undef elements.
Part of the fix for PR43024 where ideally we shouldn't call resolveTargetShuffleAndZeroables for Depth == 0
llvm-svn: 374928
The only things VBROADCAST_LOAD uses is an address and a chain
node. It has no vector inputs.
So if its a user of the source of another broadcast that could
only mean one of two things. The other broadcast is broadcasting
the address of the broadcast_load. Or the source is a load and
the use we're seeing is the chain result from that load. Neither
of these cases make sense to combine here.
This issue was reported post-commit r373871. Test case has not
been reduced yet.
llvm-svn: 374862
This prevents isel from emitting a TEST instruction that
optimizeCompareInstr will need to remove later.
In some of the modified tests, the SUB gets duplicated due to
the flags being needed in two places and being clobbered in
between. optimizeCompareInstr was able to optimize away the TEST
that was using the result of one of them, but optimizeCompareInstr
doesn't know to turn SUB into CMP after removing the TEST. It
only knows how to turn SUB into CMP if the result was already
dead.
With this change the TEST never exists, so optimizeCompareInstr
doesn't have to remove it. Then it can just turn the SUB into
CMP immediately.
Fixes PR43649.
llvm-svn: 374755
We were already controlling whether the KnownZero elements were being written to the target mask, this extends it to the KnownUndef elements as well so we can prevent the target shuffle mask being manipulated at all.
llvm-svn: 374732
This enables use of the saturating truncate instructions when the
result type is less than 128 bits. It also enables the use of
saturating truncate instructions on KNL when the input is less
than 512 bits. We can do this by widening the input and then
extracting the result.
llvm-svn: 374731
This seems to improve std::midpoint code where we have a min and
a max with the same condition. If we split the setcc we can end
up with two compares if the one of the operands is a constant.
Since we aggressively canonicalize compares with constants.
For non-constants it can interfere with our ability to share
control flow if we need to expand cmovs into control flow.
I'm also not sure I understand this min/max canonicalization code.
The motivating case talks about comparing with 0. But we don't
check for 0 explicitly.
Removes one instruction from the codegen for PR43658.
llvm-svn: 374706
Since the input type is larger than 256-bits we'll need to some
concatenating to reassemble the results. The pack instructions
ability to concatenate while packing make this a shorter/faster
sequence.
llvm-svn: 374643
We already did this for VTRUNCUS with a specific combination of
types. This extends this to VTRUNCS and handles any types where
a truncating store is legal.
llvm-svn: 374615
If we don't have VLX we won't end up selecting a saturating
truncate for 256-bit or smaller vectors so we should just use
the pack lowering.
llvm-svn: 374487
When handling the packus pattern for i32->i8 we do a two step
process using a packss to i16 followed by a packus to i8. If the
final i8 step is a type with less than 64-bits the packus step
will return SDValue(), but the i32->i16 step might have succeeded.
This leaves the nodes from the middle step dangling.
Guard against this by pre-checking that the number of elements is
at least 8 before doing the middle step.
With that check in place this should mean the only other
case the middle step itself can fail is when SSE2 is disabled. So
add an early SSE2 check then just assert that neither the middle
or final step ever fail.
llvm-svn: 374460
If we've disable zmm registers, the v16i32 will need to be split. This split will propagate through min/max the truncate. This creates two sequences that need to be concatenated back to v16i8. We can instead use packusdw to do part of the clamping, truncating, and concatenating all at once. Then we can use a vpmovuswb to finish off the clamp.
Differential Revision: https://reviews.llvm.org/D68763
llvm-svn: 374431
As background, starting in D66309, I'm working on support unordered atomics analogous to volatile flags on normal LoadSDNode/StoreSDNodes for X86.
As part of that, I spent some time going through usages of LoadSDNode and StoreSDNode looking for cases where we might have missed a volatility check or need an atomic check. I couldn't find any cases that clearly miscompile - i.e. no test cases - but a couple of pieces in code loop suspicious though I can't figure out how to exercise them.
This patch adds defensive checks and asserts in the places my manual audit found. If anyone has any ideas on how to either a) disprove any of the checks, or b) hit the bug they might be fixing, I welcome suggestions.
Differential Revision: https://reviews.llvm.org/D68419
llvm-svn: 374261
Gather instructions can use i32 or i64 elements for indices. If
the index is zero extended from a type smaller than i32 to i64, we
can shrink the extend to just extend to i32.
llvm-svn: 373982
When the target option GuaranteedTailCallOpt is specified, calls with
the fastcc calling convention will be transformed into tail calls if
they are in tail position. This diff adds a new calling convention,
tailcc, currently supported only on X86, which behaves the same way as
fastcc, except that the GuaranteedTailCallOpt flag does not need to
enabled in order to enable tail call optimization.
Patch by Dwight Guth <dwight.guth@runtimeverification.com>!
Reviewed By: lebedev.ri, paquette, rnk
Differential Revision: https://reviews.llvm.org/D67855
llvm-svn: 373976
If a fp scalar is loaded and then used as both a scalar and a vector broadcast, perform the load as a broadcast and then extract the scalar for 'free' from the 0th element.
This involved switching the order of the X86ISD::BROADCAST combines so we only convert to X86ISD::BROADCAST_LOAD once all other canonicalizations have been attempted.
Adds a DAGCombinerInfo::recursivelyDeleteUnusedNodes wrapper.
Fixes PR43217
Differential Revision: https://reviews.llvm.org/D68544
llvm-svn: 373871
Replaces setTargetShuffleZeroElements with getTargetShuffleAndZeroables which reports the Zeroable elements but doesn't merge them into the decoded target shuffle mask (the merging has been moved up into getTargetShuffleInputs until we can get rid of it entirely).
This is part of the work to fix PR43024 and allow us to use SimplifyDemandedElts to simplify shuffle chains - we need to get to a point where the target shuffle mask isn't adjusted by its source inputs but instead we cache them in a parallel Zeroable mask.
llvm-svn: 373867
Summary:
The default legalization for v16i64->v16i8 tries to create a multiple stage truncate concatenating after each stage and truncating again. But avx512 implements truncates with multiple uops. So it should be better to truncate all the way to the desired element size and then concatenate the pieces using unpckl instructions. This minimizes the number of 2 uop truncates. The unpcks are all single uop instructions.
I tried to handle this by just custom splitting the v16i64->v16i8 shuffle. And hoped that the DAG combiner would leave the two halves in the state needed to make D68374 do the job for each half. This worked for the first half, but the second half got messed up. So I've implemented custom handling for v8i64->v8i8 when v8i64 needs to be split to produce the VTRUNCs directly.
Reviewers: RKSimon, spatel
Reviewed By: RKSimon
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D68428
llvm-svn: 373864
We can make use of the Zeroable mask to indicate which elements we can safely set to zero instead of creating a target shuffle mask on the fly.
This allows us to remove createTargetShuffleMask.
This is part of the work to fix PR43024 and allow us to use SimplifyDemandedElts to simplify shuffle chains - we need to get to a point where the target shuffle masks isn't adjusted by its source inputs in setTargetShuffleZeroElements but instead we cache them in a parallel Zeroable mask.
llvm-svn: 373846
As discussed on PR42025, with more complex boolean math we can end up with many truncations/extensions of the comparison results through each bitop.
This patch handles the cases introduced in combineBitcastvxi1 by pushing the sign extension through the AND/OR/XOR ops so its just the original SETCC ops that gets extended.
Differential Revision: https://reviews.llvm.org/D68226
llvm-svn: 373834
Rename some variables to match lowerShuffleAsRepeatedMaskAndLanePermute - prep work toward adding some equivalent sublane functionality.
llvm-svn: 373832
We already do this for ISD::TRUNCATE, but we can do the same for X86ISD::VTRUNC
Differential Revision: https://reviews.llvm.org/D68432
llvm-svn: 373765
This patch recognizes the shuffle pattern we get from a
v8i64->v8i8 truncate when v8i64 isn't a legal type.
With VLX we can use two VTRUNCs, unpckldq, and a insert_subvector.
Diffrential Revision: https://reviews.llvm.org/D68374
llvm-svn: 373645
We can make use of the Zeroable mask to indicate which elements we can safely set to zero instead of creating a target shuffle mask on the fly.
This only leaves one user of createTargetShuffleMask which we can hopefully get rid of in a similar manner.
This is part of the work to fix PR43024 and allow us to use SimplifyDemandedElts to simplify shuffle chains - we need to get to a point where the target shuffle masks isn't adjusted by its source inputs in setTargetShuffleZeroElements but instead we cache them in a parallel Zeroable mask.
llvm-svn: 373641
This improves broadcast load folding of i64 elements on 32-bit
targets where i64 isn't legal.
Previously we had to represent these as vXf64 vbroadcast_loads and
a bitcast to vXi64. But we didn't have any isel patterns
looking for that.
This also allows us to remove or simplify some isel patterns that
were looking for bitcasted vbroadcast_loads.
llvm-svn: 373566
The previous code tried to do a trick where we would extract the subvector from the location we were inserting. Then xor that with the new value. Take the xored value and clear out the bits above the subvector size. Then shift that xored subvector to the insert location. And finally xor that with the original vector. Since the old subvector was used in both xors, this would leave just the new subvector at the inserted location. Since the surrounding bits had been zeroed no other bits of the original vector would be modified.
Unfortunately, if the old subvector came from undef we might aggressively propagate the undef. Then we end up with the XORs not cancelling because they aren't using the same value for the two uses of the old subvector. @bkramer gave me a case that demonstrated this, but we haven't reduced it enough to make it easily readable to see what's happening.
This patch uses a safer, but more costly approach. It isolate the bits above the insertion and bits below the insert point and ORs those together leaving 0 for the insertion location. Then widens the subvector with 0s in the upper bits, shifts it into position with 0s in the lower bits. Then we do another OR.
Differential Revision: https://reviews.llvm.org/D68311
llvm-svn: 373495
The gather/scatter instructions can implicitly sign extend the indices. If we're operating on 32-bit data, an v16i64 index can force a v16i32 gather to be split in two since the index needs 2 registers. If we can shrink the index to the i32 we can avoid the split. It should always be safe to shrink the index regardless of the number of elements. We have gather/scatter instructions that can use v2i32 index stored in a v4i32 register with v2i64 data size.
I've limited this to before legalize types to avoid creating a v2i32 after type legalization. We could check for it, but we'd also need testing. I'm also only handling build_vectors with no bitcasts to be sure the truncate will constant fold.
Differential Revision: https://reviews.llvm.org/D68247
llvm-svn: 373408
Summary:
This adds the ISD opcode and a DAG combine to create it. There are
probably some places where we can directly create it, but I'll
leave that for future work.
This updates all of the isel patterns to look for this new node.
I had to add a few additional isel patterns for aligned extloads
which we should probably fix with a DAG combine or something. This
does mean that the broadcast load folding for avx512 can no
longer match a broadcasted aligned extload.
There's still some work to do here for combining a broadcast of
a broadcast_load. We also need to improve extractelement or
demanded vector elements of a broadcast_load. I'll try to get
those done before I submit this patch.
Reviewers: RKSimon, spatel
Reviewed By: RKSimon
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D68198
llvm-svn: 373349
Replace with the MachineFunction. X86 is the only user, and only uses
it for the function. This removes one obstacle from using this in
GlobalISel. The other is the more tolerable EVT argument.
The X86 use of the function seems questionable to me. It checks hasFP,
before frame lowering.
llvm-svn: 373292
The i1 scalar would have been type legalized to i8, but that
doesn't guarantee anything about the upper bits. If we're going
to use it as condition we need to make sure the upper bits are 0.
I've special cased ISD::SETCC conditions since that should
guarantee zero upper bits. We could go further and use
computeKnownBits, but we have no tests that would need that.
Fixes PR43507.
llvm-svn: 373246
ANY_EXTEND of v8i8 is marked Custom on AVX512 for handling extends
from v8i8. But the type legalization infrastructure will call
ReplaceNodeResults for v8i8 results. We should just defer it the
default handling instead of asserting in the default of the switch.
Fixes PR43509.
llvm-svn: 373234
This was added back to allow some performance regressions to be
investigated. The main perf issue was fixed shortly after adding
this back and no other major issues have been reported. So I
think its safe to remove this again.
llvm-svn: 373174
Creating new nodes is what we usually do. Have to explicitly
check that we don't update to an existing node and having
to manually manage the worklist is unusual.
We can probably add a helper function to reduce the duplication
of having to check if we should create a gather or scatter, but
I wanted to just get the simple thing done.
llvm-svn: 373137
The majority of the code doesn't run on the X86 nodes today since
its gated by isBeforeLegalizeOps and we don't formm X86 nodes
until after that. Except for a couple special case in type
legalization. But I think we would probably break those if
some of the transforms fire on them.
I want to remove the hardcoded operand numbers and the unusual
use of UpdateNodeOperands. Being able to know which ISD opcodes
are present should help with that.
llvm-svn: 373136
This removes the need for ConvertToTarget opcodes in the isel table.
It's also consistent with the recent changes to use TargetConstant
for intrinsic nodes that always take immediates.
Differential Revision: https://reviews.llvm.org/D67902
llvm-svn: 372645
The attached test case would previous infinite loop after
r365711.
I'm going to move this to X86ISelDAGToDAG.cpp to get the setcc
to match VPTEST in 32-bit mode in a follow up commit.
llvm-svn: 372543
This allows us to use timm in the isel table which is more
consistent with other intrinsics that take an immediate now.
We can't declare the intrinsic as taking an ImmArg because we
need to match non-constants to the shift by MMX register
instruction which we do by mutating the intrinsic id during
lowering.
llvm-svn: 372537
This intrinsics should be shift by immediate, but gcc allows any
i32 scalar and clang needs to match that. So we try to detect the
non-constant case and move the data from an integer register to an
MMX register.
Previously this was done by creating a v2i32 build_vector and
bitcast in SelectionDAGBuilder. This had to be done early since
v2i32 isn't a legal type. The bitcast+build_vector would be DAG
combined to X86ISD::MMX_MOVW2D which isel will turn into a
GPR->MMX MOVD.
This commit just moves the whole thing to lowering and emits
the X86ISD::MMX_MOVW2D directly to avoid the illegal type. The
test changes just seem to be due to nodes being linearized in a
different order.
llvm-svn: 372535
This reverts commit 52621307bc.
Tests have been failing all night with
[0/2] ACTION //llvm/test:check-llvm(//llvm/utils/gn/build/toolchain:unix)
-- Testing: 33647 tests, 64 threads --
Testing: 0 .. 10..
UNRESOLVED: LLVM :: CodeGen/AMDGPU/GlobalISel/isel-blendi-gettargetconstant.ll (6943 of 33647)
******************** TEST 'LLVM :: CodeGen/AMDGPU/GlobalISel/isel-blendi-gettargetconstant.ll' FAILED ********************
Test has no run line!
********************
Since there were other concerns on https://reviews.llvm.org/D67785,
I'm just reverting for now.
llvm-svn: 372383
We reuse an ISD opcode here that can be reached from BMI that
doesn't require it to be an immediate. Our isel patterns to match
the TBM immediate form require a Constant and not a TargetConstant.
We were accidentally getting the Constant due to a quirk of
combineBEXTR calling SimplifyDemandedBits. The call to
SimplifyDemandedBits ended up constant folding the TargetConstant
to a regular Constant. But we should probably instead be asserting
if SimplifyDemandedBits on a TargetConstant so we shouldn't rely
on this behavior.
llvm-svn: 372373
Summary: This fixes a crasher introduced by r372338.
Reviewers: echristo, arsenm
Subscribers: jvesely, wdng, nhaehnle, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D67785
Tighten up the test case.
llvm-svn: 372366
The later code that generates a constant when there are
some non-const elements works basically the same and doesn't
require there to be any non-const elements.
llvm-svn: 372365
This reverts r372314, reapplying r372285 and the commits which depend
on it (r372286-r372293, and r372296-r372297)
This was missing one switch to getTargetConstant in an untested case.
llvm-svn: 372338
This patch converts the DAGCombine isNegatibleForFree/GetNegatedExpression into overridable TLI hooks and includes a demonstration X86 implementation.
The intention is to let us extend existing FNEG combines to work more generally with negatible float ops, allowing it work with target specific combines and opcodes (e.g. X86's FMA variants).
Unlike the SimplifyDemandedBits, we can't just handle target nodes through a Target callback, we need to do this as an override to allow targets to handle generic opcodes as well. This does mean that the target implementations has to duplicate some checks (recursion depth etc.).
I've only begun to replace X86's FNEG handling here, handling FMADDSUB/FMSUBADD negation and some low impact codegen changes (some FMA negatation propagation). We can build on this in future patches.
Differential Revision: https://reviews.llvm.org/D67557
llvm-svn: 372333
This broke the Chromium build, causing it to fail with e.g.
fatal error: error in backend: Cannot select: t362: v4i32 = X86ISD::VSHLI t392, Constant:i8<15>
See llvm-commits thread of r372285 for details.
This also reverts r372286, r372287, r372288, r372289, r372290, r372291,
r372292, r372293, r372296, and r372297, which seemed to depend on the
main commit.
> Encode them directly as an imm argument to G_INTRINSIC*.
>
> Since now intrinsics can now define what parameters are required to be
> immediates, avoid using registers for them. Intrinsics could
> potentially want a constant that isn't a legal register type. Also,
> since G_CONSTANT is subject to CSE and legalization, transforms could
> potentially obscure the value (and create extra work for the
> selector). The register bank of a G_CONSTANT is also meaningful, so
> this could throw off future folding and legalization logic for AMDGPU.
>
> This will be much more convenient to work with than needing to call
> getConstantVRegVal and checking if it may have failed for every
> constant intrinsic parameter. AMDGPU has quite a lot of intrinsics wth
> immarg operands, many of which need inspection during lowering. Having
> to find the value in a register is going to add a lot of boilerplate
> and waste compile time.
>
> SelectionDAG has always provided TargetConstant for constants which
> should not be legalized or materialized in a register. The distinction
> between Constant and TargetConstant was somewhat fuzzy, and there was
> no automatic way to force usage of TargetConstant for certain
> intrinsic parameters. They were both ultimately ConstantSDNode, and it
> was inconsistently used. It was quite easy to mis-select an
> instruction requiring an immediate. For SelectionDAG, start emitting
> TargetConstant for these arguments, and using timm to match them.
>
> Most of the work here is to cleanup target handling of constants. Some
> targets process intrinsics through intermediate custom nodes, which
> need to preserve TargetConstant usage to match the intrinsic
> expectation. Pattern inputs now need to distinguish whether a constant
> is merely compatible with an operand or whether it is mandatory.
>
> The GlobalISelEmitter needs to treat timm as a special case of a leaf
> node, simlar to MachineBasicBlock operands. This should also enable
> handling of patterns for some G_* instructions with immediates, like
> G_FENCE or G_EXTRACT.
>
> This does include a workaround for a crash in GlobalISelEmitter when
> ARM tries to uses "imm" in an output with a "timm" pattern source.
llvm-svn: 372314
Encode them directly as an imm argument to G_INTRINSIC*.
Since now intrinsics can now define what parameters are required to be
immediates, avoid using registers for them. Intrinsics could
potentially want a constant that isn't a legal register type. Also,
since G_CONSTANT is subject to CSE and legalization, transforms could
potentially obscure the value (and create extra work for the
selector). The register bank of a G_CONSTANT is also meaningful, so
this could throw off future folding and legalization logic for AMDGPU.
This will be much more convenient to work with than needing to call
getConstantVRegVal and checking if it may have failed for every
constant intrinsic parameter. AMDGPU has quite a lot of intrinsics wth
immarg operands, many of which need inspection during lowering. Having
to find the value in a register is going to add a lot of boilerplate
and waste compile time.
SelectionDAG has always provided TargetConstant for constants which
should not be legalized or materialized in a register. The distinction
between Constant and TargetConstant was somewhat fuzzy, and there was
no automatic way to force usage of TargetConstant for certain
intrinsic parameters. They were both ultimately ConstantSDNode, and it
was inconsistently used. It was quite easy to mis-select an
instruction requiring an immediate. For SelectionDAG, start emitting
TargetConstant for these arguments, and using timm to match them.
Most of the work here is to cleanup target handling of constants. Some
targets process intrinsics through intermediate custom nodes, which
need to preserve TargetConstant usage to match the intrinsic
expectation. Pattern inputs now need to distinguish whether a constant
is merely compatible with an operand or whether it is mandatory.
The GlobalISelEmitter needs to treat timm as a special case of a leaf
node, simlar to MachineBasicBlock operands. This should also enable
handling of patterns for some G_* instructions with immediates, like
G_FENCE or G_EXTRACT.
This does include a workaround for a crash in GlobalISelEmitter when
ARM tries to uses "imm" in an output with a "timm" pattern source.
llvm-svn: 372285
This generates worse code, but matches what is done for avx2 and
prevents crashes when more arguments are passed than we have
registers for.
llvm-svn: 372200
The case were Immediate is 0 and HasConstElts is true should never
happen since that would mean the constant elts were all zero. But
we check for all zero build vector earlier. So just use HasConstElts
and blindly take Immediate without checking if its 0.
Move the code that bitcasts and extract the immediate into the
the HasConstElts case since the other code just creates an undef
with the right type. No casting needed.
llvm-svn: 372153
* Reordered MVT simple types to group scalable vector types
together.
* New range functions in MachineValueType.h to only iterate over
the fixed-length int/fp vector types.
* Stopped backends which don't support scalable vector types from
iterating over scalable types.
Reviewers: sdesmalen, greened
Reviewed By: greened
Differential Revision: https://reviews.llvm.org/D66339
llvm-svn: 372099
Previously we tried to split them into narrower v64i1 or v16i1
pieces that each got promoted to vXi8 and then passed in a zmm
or xmm register. But this crashes when you need to pass more
pieces than available registers reserved for argument passing.
The scalarizing done here generates much longer and slower code,
but is consistent with the behavior of avx2 and earlier targets
for these types.
Fixes PR43323.
llvm-svn: 372069
Some prep work for PR42863, this change allows us to move all the FMA opcode mappings into the negateFMAOpcode helper.
For the FMADDSUB/FMSUBADD cases, we can only negate the accumulator - any other negations will result in an error.
llvm-svn: 371840
The X86 decision assumes the compare will produce a result in an XMM
register, but that can't happen for an fp128 compare since those
go to a libcall the returns an i32. Pass the VT so X86 can check
the type.
llvm-svn: 371775
I found three issues:
1. the loop over E[ABCD]X copies run over BB start
2. the direct address of cmpxchg8b could be a frame index
3. the displacement of cmpxchg8b could be a global instead of an
immediate
These were all introduced together in r287875, and should be fixed with
this change.
Issue reported by Zachary Turner.
llvm-svn: 371678
fp128 is considered a legal type for a register, but has almost no legal operations so everything needs to be converted to a libcall. Previously this was implemented by tricking type legalization into softening the operations with various checks for "is legal in hardware register" to change the behavior to still use f128 as the resulting type instead of converting to i128.
This patch abandons this approach and instead moves the libcall conversions to LegalizeDAG. This is the approach taken by AArch64 where they also have a legal fp128 type, but no legal operations. I think this is more in spirit with how SelectionDAG's phases are supposed to work.
I had to make some hacks for STRICT_FP_ROUND because some of the strict FP handling checks if ISD::FP_ROUND is Legal for a given result type, but I had to make ISD::FP_ROUND Custom to allow making a libcall when the input is f128. For all other types the Custom handler just returns the original node. These hacks are incomplete and don't work for a strict truncate from f128, but I don't think it worked before either since LegalizeFloatTypes doesn't know about strict ops yet. I've also raised PR43209 against AArch64 which currently crashes on a strict ftrunc from f64->f32 because of FP_ROUND being marked Custom for the same reason there.
Differential Revision: https://reviews.llvm.org/D67128
llvm-svn: 371672
See D66309 for context.
This is the first sweep of x86 target specific code to add isAtomic bailouts where appropriate. The intention here is to have the switch from AtomicSDNode to LoadSDNode/StoreSDNode be close to NFC; that is, I'm not looking to allow additional optimizations at this time.
Sorry for the lack of tests. As discussed in the review, most of these are vector tests (for which atomicity is not well defined) and I couldn't figure out to exercise the anyextend cases which aren't vector specific.
Differential Revision: https://reviews.llvm.org/D66322
llvm-svn: 371547
This is the first patch in a large sequence. The eventual goal is to have unordered atomic loads and stores - and possibly ordered atomics as well - handled through the normal ISEL codepaths for loads and stores. Today, there handled w/instances of AtomicSDNodes. The result of which is that all transforms need to be duplicated to work for unordered atomics. The benefit of the current design is that it's harder to introduce a silent miscompile by adding an transform which forgets about atomicity. See the thread on llvm-dev titled "FYI: proposed changes to atomic load/store in SelectionDAG" for further context.
Note that this patch is NFC unless the experimental flag is set.
The basic strategy I plan on taking is:
introduce infrastructure and a flag for testing (this patch)
Audit uses of isVolatile, and apply isAtomic conservatively*
piecemeal conservative* update generic code and x86 backedge code in individual reviews w/tests for cases which didn't check volatile, but can be found with inspection
flip the flag at the end (with minimal diffs)
Work through todo list identified in (2) and (3) exposing performance ops
(*) The "conservative" bit here is aimed at minimizing the number of diffs involved in (4). Ideally, there'd be none. In practice, getting it down to something reviewable by a human is the actual goal. Note that there are (currently) no paths which produce LoadSDNode or StoreSDNode with atomic MMOs, so we don't need to worry about preserving any behaviour there.
We've taken a very similar strategy twice before with success - once at IR level, and once at the MI level (post ISEL).
Differential Revision: https://reviews.llvm.org/D66309
llvm-svn: 371441
Current for SAE instructions we only allow _MM_FROUND_CUR_DIRECTION(bit 2) or _MM_FROUND_NO_EXC(bit 3) to be used as the immediate passed to the inrinsics. But these instructions don't perform rounding so _MM_FROUND_CUR_DIRECTION is just sort of a default placeholder when you don't want to suppress exceptions. Using _MM_FROUND_NO_EXC by itself is really bit equivalent to (_MM_FROUND_NO_EXC | _MM_FROUND_TO_NEAREST_INT) since _MM_FROUND_TO_NEAREST_INT is 0. Since we aren't rounding on these instructions we should also accept (_MM_FROUND_CUR_DIRECTION | _MM_FROUND_NO_EXC) as equivalent to (_MM_FROUND_NO_EXC). icc allows this, but gcc does not.
Differential Revision: https://reviews.llvm.org/D67289
llvm-svn: 371430
This patch decodes target and faux shuffles with getTargetShuffleInputs - a reduced version of resolveTargetShuffleInputs that doesn't resolve SM_SentinelZero cases, so we can correctly remove zero vectors if they aren't demanded.
llvm-svn: 371353
If the two zero vectors have undefs in different places they
won't get combined by simplifySelect.
This fixes a regression from an earlier commit.
llvm-svn: 371351
The change to avx512-vec-cmp.ll is a regression, but should be
easy to fix. It occurs because the getZeroVector call was
canonicalizing both sides to the same node, then SimplifySelect
was able to simplify it. But since only called getZeroVector
on some VTs this isn't a robust way to combine this.
The change to vector-shuffle-combining-ssse3.ll is more
instructions, but removes a constant pool load so its unclear
if its a regression or not.
llvm-svn: 371350
This generalizes the existing <32 x i1> pre-AVX2 split code to support reductions from <64 x i1> as well, we can probably generalize to any larger pow2 case in the future if the (unlikely) need ever arises.
We still need to tweak combineBitcastvxi1 to improve AVX512F codegen as its assumes vXi1 types should be handled on the mask registers even when they aren't legal.
Differential Revision: https://reviews.llvm.org/D67070
llvm-svn: 371328
isel used to require zero vectors to be canonicalized to a single
type to minimize the number of patterns needed to match. This is
no longer required.
I plan to do this to integers too, but floating point was simpler
to start with. Integer has a complication where v32i16/v64i8 aren't
legal when the other 512-bit integer types are.
llvm-svn: 371325
Use getAPIntValue() directly - this is mainly a best practice style issue to help prevent fuzz tests blowing up when a i12345 (or whatever) is generated.
Use getConstantOperandVal/getConstantOperandAPInt wrappers where possible.
llvm-svn: 371315
Fix for https://bugs.llvm.org/show_bug.cgi?id=43230.
When creating PSHUFLW from a repeated shuffle mask, we have to apply
the checks to the repeated mask, not the original one. For the test
case from PR43230 the inspected part of the original mask is all undef.
Differential Revision: https://reviews.llvm.org/D67314
llvm-svn: 371307
Summary:
This is patch is part of a series to introduce an Alignment type.
See this thread for context: http://lists.llvm.org/pipermail/llvm-dev/2019-July/133851.html
See this patch for the introduction of the type: https://reviews.llvm.org/D64790
Reviewers: courbet
Subscribers: nemanjai, javed.absar, hiraditya, kbarton, asb, rbar, johnrusso, simoncook, apazos, sabuasal, niosHD, jrtc27, MaskRay, zzheng, edward-jones, rogfer01, MartinMosbeck, brucehoult, the_o, PkmX, jocewei, jsji, s.egerton, pzheng, ychen, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D67267
llvm-svn: 371212
As noted in PR43197, we can use test+add+cmov+sra to implement
signed division by a power of 2.
This is based off the similar version in AArch64, but I've
adjusted it to use target independent nodes where AArch64 uses
target specific CMP and CSEL nodes. I've also blocked INT_MIN
as the transform isn't valid for that.
I've limited this to i32 and i64 on 64-bit targets for now and only
when CMOV is supported. i8 and i16 need further investigation to be
sure they get promoted to i32 well.
I adjusted a few tests to enable cmov to demonstrate the new
codegen. I also changed twoaddr-coalesce-3.ll to 32-bit mode
without cmov to avoid perturbing the scenario that is being
set up there.
Differential Revision: https://reviews.llvm.org/D67087
llvm-svn: 371104
As discussed on D64551 and PR43227, we don't correctly handle cases where the base load has a non-zero byte offset.
Until we can properly handle this, we must bail from EltsFromConsecutiveLoads.
llvm-svn: 371078
Summary:
This patch renames functions that takes or returns alignment as log2, this patch will help with the transition to llvm::Align.
The renaming makes it explicit that we deal with log(alignment) instead of a power of two alignment.
A few renames uncovered dubious assignments:
- `MirParser`/`MirPrinter` was expecting powers of two but `MachineFunction` and `MachineBasicBlock` were using deal with log2(align). This patch fixes it and updates the documentation.
- `MachineBlockPlacement` exposes two flags (`align-all-blocks` and `align-all-nofallthru-blocks`) supposedly interpreted as power of two alignments, internally these values are interpreted as log2(align). This patch updates the documentation,
- `MachineFunctionexposes` exposes `align-all-functions` also interpreted as power of two alignment, internally this value is interpreted as log2(align). This patch updates the documentation,
Reviewers: lattner, thegameg, courbet
Subscribers: dschuff, arsenm, jyknight, dylanmckay, sdardis, nemanjai, jvesely, nhaehnle, javed.absar, hiraditya, kbarton, fedor.sergeev, asb, rbar, johnrusso, simoncook, apazos, sabuasal, niosHD, jrtc27, MaskRay, zzheng, edward-jones, atanasyan, rogfer01, MartinMosbeck, brucehoult, the_o, dexonsmith, PkmX, jocewei, jsji, Jim, s.egerton, llvm-commits, courbet
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D65945
llvm-svn: 371045
This reverts r370525 (git commit 0bb1630685)
Also reverts r370543 (git commit 185ddc08ee)
The approach I took only works for functions marked `noreturn`. In
general, a call that is not known to be noreturn may be followed by
unreachable for other reasons. For example, there could be multiple call
sites to a function that throws sometimes, and at some call sites, it is
known to always throw, so it is followed by unreachable. We need to
insert an `int3` in these cases to pacify the Windows unwinder.
I think this probably deserves its own standalone, Win64-only fixup pass
that runs after block placement. Implementing that will take some time,
so let's revert to TrapUnreachable in the mean time.
llvm-svn: 370829
This merges the 32-bit and 64-bit mode code to just use Custom
for both i32 and i64. We already had most of the handling in
the custom handling due to the AVX512 having legal fp_to_uint.
Just needed to add the i32->i64 promotion handling. Refactor
the fp_to_uint code in the custom handler to simplify the
number of times we check things.
Tweak cost model tables to match the default handling we were
getting due to Expand before.
llvm-svn: 370700
Use Custom lowering instead. Fall back to default expansion only
when the scalar FP type belongs in an XMM register. This improves
lowering for i32 to fp80, and also i32 to double on SSE1 only.
llvm-svn: 370699
FP128 values are passed in xmm registers so should be asssociated
with an SSE feature rather than MMX which uses a different set
of registers.
llc enables sse1 and sse2 by default with x86_64. But does not
enable mmx. Clang enables all 3 features by default.
I've tried to add command lines to test with -sse
where possible, but any test that returns a value in an xmm
register fails with a fatal error with -sse since we have no
defined ABI for that scenario.
llvm-svn: 370682
Rename to lowerShuffleAsLanePermuteAndShuffle to make it clear that not just blends are performed.
Cleanup the in-lane shuffle mask generation to make it more obvious what's going on.
Some prep work noticed while investigating the poor shuffle code mentioned in D66004.
llvm-svn: 370613
EltsFromConsecutiveLoads was assuming that the number of input elts was the same as the number of elements in the output vector type when creating a zeroing shuffle, causing an assert when subvectors were being combined instead of just scalars.
llvm-svn: 370592
Users have complained llvm.trap produce two ud2 instructions on Win64,
one for the trap, and one for unreachable. This change fixes that.
TrapUnreachable was added and enabled for Win64 in r206684 (April 2014)
to avoid poorly understood issues with the Windows unwinder.
There seem to be two major things in play:
- the unwinder
- C++ EH, _CxxFrameHandler3 & co
The unwinder disassembles forward from the return address to scan for
epilogues. Inserting a ud2 had the effect of stopping the unwinder, and
ensuring that it ran the EH personality function for the current frame.
However, it's not clear what the unwinder does when the return address
happens to be the last address of one function and the first address of
the next function.
The Visual C++ EH personality, _CxxFrameHandler3, needs to figure out
what the current EH state number is. It does this by consulting the
ip2state table, which maps from PC to state number. This seems to go
wrong when the return address is the last PC of the function or catch
funclet.
I'm not sure precisely which system is involved here, but in order to
address these real or hypothetical problems, I believe it is enough to
insert int3 after a call site if it would otherwise be the last
instruction in a function or funclet. I was able to reproduce some
similar problems locally by arranging for a noreturn call to appear at
the end of a catch block immediately before an unrelated function, and I
confirmed that the problems go away when an extra trailing int3
instruction is added.
MSVC inserts int3 after every noreturn function call, but I believe it's
only necessary to do it if the call would be the last instruction. This
change inserts a pseudo instruction that expands to int3 if it is in the
last basic block of a function or funclet. I did what I could to run the
Microsoft compiler EH tests, and the ones I was able to run showed no
behavior difference before or after this change.
Differential Revision: https://reviews.llvm.org/D66980
llvm-svn: 370525
gcc and icc pass these types in zmm registers in zmm registers.
This patch implements a quick hack to override the register
type before calling convention handling to one that is legal.
Longer term we might want to do something similar to 256-bit
integer registers on AVX1 where we just split all the operations.
Fixes PR42957
Differential Revision: https://reviews.llvm.org/D66708
llvm-svn: 370495
ISD::isBuildVectorAllZeros permits undef elements to be present, which means we can't return it as a zero vector. PMULDQ/PMULUDQ is an extending multiply so a multiply by zero of the lower 32-bits should result in a zero 64-bit element.
llvm-svn: 370404
Summary:
We were previously doing it in DAGCombine.
But we also want to do `sub %x, C` -> `add %x, (sub 0, C)` for vectors in DAGCombine.
So if we had `sub %x, -1`, we'll transform it to `add %x, 1`,
which `combineIncDecVector()` will immediately transform back into `sub %x, -1`,
and here we go again...
I've marked this as NFC since not a single test changes,
but since that 'changes' DAGCombine, probably this isn't fully NFC.
Reviewers: RKSimon, craig.topper, spatel
Reviewed By: craig.topper
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D62327
llvm-svn: 370327
We had an isel pattern to perform this, but its better to
do it in DAG combine as a simplification. This also fixes the lack
of patterns for AVX512 targets.
llvm-svn: 370294
Including a type legalizer fix to make bitcast operand promotion
work correctly when getSoftenedFloat returns f128 instead of i128.
Fixes PR43157
llvm-svn: 370293
Neither libgcc or compiler-rt are usually used on Windows, so these
functions can't be called.
Differential revision: https://reviews.llvm.org/D66880
llvm-svn: 370204
Probably better to keep add over sub in early DAG combines.
It might make sense to push this to lowering or delay it all
the way to isel. But this was the simplest change.
llvm-svn: 369981
ANY_EXTEND_VECTOR_INREG isn't currently marked Legal which prevents SimplifyDemandedBits from turning SIGN/ZERO_EXTEND_VECTOR_INREG into it after op legalization. And even if we did make it Legal, combineExtInVec doesn't do shuffle combining on the VECTOR_INREG nodes until AVX1.
This patch adds a quick hack to combinePMULDQ to directly emit a vector shuffle corresponding to an ANY_EXTEND_VECTOR_INREG operation. This avoids both of those issues without creating any other regressions on our tests. The xop-ifma.ll change here also showed up when I tried to resurrect D56306 and seemed to be the only improvement that patch creates now. This is a more direct way to get the benefit.
Differential Revision: https://reviews.llvm.org/D66436
llvm-svn: 369942
Summary:
Concat_vectors is more canonical during early DAG combine. For example, its what's used by SelectionDAGBuilder when converting IR shuffles into SelectionDAG shuffles when element counts between inputs and mask don't match. We also have combines in DAGCombiner than can pull concat_vectors through a shuffle. See partitionShuffleOfConcats. So it seems like concat_vectors is a better operation to use here. I had to teach DAGCombiner's SimplifyVBinOp to also handle concat_vectors with undef. I haven't checked yet if we can remove the INSERT_SUBVECTOR version in there or not.
I didn't want to mess with the other caller of getShuffleHalfVectors that's used during shuffle lowering where insert_subvector probably is what we want to produce so I've enabled this via a boolean passed to the function.
Reviewers: spatel, RKSimon
Reviewed By: RKSimon
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D66504
llvm-svn: 369872
CONCAT_VECTORS and INSERT_SUBVECTORS can both call combineConcatVectorOps,
but we shouldn't produce INSERT_SUBVECTORS from there. We should
keep CONCAT_VECTORS until vector legalization.
Noticed while looking at the madd_quad_reduction test from madd.ll
llvm-svn: 369802
Patch showing the effect of enabling bool vector oversimplification.
Non-VLX builds can simplify a kshift shuffle, but VLX builds simplify:
insert_subvector v8i zeroinitializer, v2i --> insert_subvector v8i undef, v2i
Preventing the removal of the AND to clear the upper bits of result
Differential Revision: https://reviews.llvm.org/D53022
llvm-svn: 369780
For v2i32 we only feed 2 i8 elements into the psadbw instructions
with 0s in the other 14 bytes. The resulting psadbw instruction
will produce zeros in bits [127:16] of the output. We need to take
the result and feed it to a v2i32 add where the first element
includes bits [15:0] of the sad result. The other element should
be zero.
Prior to this patch we were using a truncate to take 0 from
bits 95:64 of the psadbw. This results in a pshufd to move those
bits to 63:32. But since we also have zeroes in bits 63:32 of
the psadbw output, we should just take those bits.
The previous code probably worked better with promoting legalization,
but now we use widening legalization. I've preserved the old
behavior if -x86-experimental-vector-widening-legalization=false
until we get that option removed.
llvm-svn: 369733
I also had to add a new combine to X86's combineExtractSubvector to prevent a regression.
This helps our vXi1 code see the full concat operation and allow it optimize undef to a zero if there is already a zero in the concat. This helped us use a movzx instead of an AND in some of the tests. In those tests, one concat comes from SelectionDAGBuilder and the second comes from type legalization of v4i1->i4 bitcasts which uses an additional concat. Though these changes weren't my original motivation.
I'm looking at making X86ISelLowering's narrowShuffle emit a concat_vectors instead of an insert_subvector since concat_vectors is more canonical during early DAG combine. This patch helps prevent a regression from my experiments with that.
Differential Revision: https://reviews.llvm.org/D66456
llvm-svn: 369459
Without AVX512DQ we don't have KMOVB so we can't really copy 8-bits of a k-register to a GPR. We have to copy 16 bits instead. We do this even if the DAG copy is from v8i1->v16i1. If we detect the (i8 (bitcast (v8i1 (extract_subvector (v16i1 X), 0)))) we should rewrite the types to match the copy we do support. By doing this, we can help known bits to propagate without losing the upper 8 bits of the input to the extract_subvector. This allows some zero extends to be removed since we have an isel pattern to use kmovw for (zero_extend (i16 (bitcast (v16i1 X))).
Differential Revision: https://reviews.llvm.org/D66489
llvm-svn: 369434
Google is reporting performance issues with the new default behavior
and have asked for a way to switch back to the old behavior while we
investigate and make fixes.
I've restored all of the code that had since been removed and added
additional checks of the command flag onto code paths that are
not otherwise guarded by a check of getTypeAction.
I've also modified the cost model tables to hopefully get us back
to the previous costs.
Hopefully we won't need to support this for very long since we
have no test coverage of the old behavior so we can very easily
break it.
llvm-svn: 369332
The motivating case are the changes in vector-reduce-add.ll where
we were doing extra work in the scalar domain instead of shuffling.
There may be some one use check that needs to be looked into there,
but this patch sidesteps the issue by avoiding broadcasts that
aren't really broadcasting.
Differential Revision: https://reviews.llvm.org/D66071
llvm-svn: 369287
We can support these by widening to a supported type,
then shifting all the way to the left and then
back to the right to ensure that we shift in zeroes.
llvm-svn: 369232
Not sure how to test this as we have tests that exercise this code,
but nothing failed for the types not matching. Since all the k-registers
use equivalent register classes everything just ends up working.
llvm-svn: 369228
This allows us to widen the type when the KSHIFTR instruction
doesn't exist for the type. If we need to shift in zeroes into
the upper elements we would need more work to guarantee zeroes
when widening.
llvm-svn: 369227
For such a case we should only need a KSHIFTL, but we were
previously generating a KSHIFTL followed by a KSHIFTR because
we mistakenly believed we need to zero the undef elements.
llvm-svn: 369224
Add similar functionality to isShuffleEquivalent - if the mask elements don't match, try matching the BUILD_VECTOR scalars instead.
As target shuffles need to handle SM_Sentinel values, this can get a bit tricky, so commit just adds actual mask element index handling - full SM_SentinelZero support will be added when the need arises.
Also, enables support in matchVectorShuffleWithPACK
llvm-svn: 369212
Simplifies shuffle mask comparisons by just bailing out if the shuffle mask has any out of range values - will make an upcoming patch much simpler.
llvm-svn: 369211
Eventually we need to generalize combineExtractWithShuffle to handle all faux shuffles and handle truncate (and X86ISD::VTRUNC etc.) there, but we're not ready yet (still creates nodes on the fly, incomplete DemandedElts support, bad use of recursive Depth limit).
llvm-svn: 369134
Summary:
This clang-tidy check is looking for unsigned integer variables whose initializer
starts with an implicit cast from llvm::Register and changes the type of the
variable to llvm::Register (dropping the llvm:: where possible).
Partial reverts in:
X86FrameLowering.cpp - Some functions return unsigned and arguably should be MCRegister
X86FixupLEAs.cpp - Some functions return unsigned and arguably should be MCRegister
X86FrameLowering.cpp - Some functions return unsigned and arguably should be MCRegister
HexagonBitSimplify.cpp - Function takes BitTracker::RegisterRef which appears to be unsigned&
MachineVerifier.cpp - Ambiguous operator==() given MCRegister and const Register
PPCFastISel.cpp - No Register::operator-=()
PeepholeOptimizer.cpp - TargetInstrInfo::optimizeLoadInstr() takes an unsigned&
MachineTraceMetrics.cpp - MachineTraceMetrics lacks a suitable constructor
Manual fixups in:
ARMFastISel.cpp - ARMEmitLoad() now takes a Register& instead of unsigned&
HexagonSplitDouble.cpp - Ternary operator was ambiguous between unsigned/Register
HexagonConstExtenders.cpp - Has a local class named Register, used llvm::Register instead of Register.
PPCFastISel.cpp - PPCEmitLoad() now takes a Register& instead of unsigned&
Depends on D65919
Reviewers: arsenm, bogner, craig.topper, RKSimon
Reviewed By: arsenm
Subscribers: RKSimon, craig.topper, lenary, aemerson, wuzish, jholewinski, MatzeB, qcolombet, dschuff, jyknight, dylanmckay, sdardis, nemanjai, jvesely, wdng, nhaehnle, sbc100, jgravelle-google, kristof.beyls, hiraditya, aheejin, kbarton, fedor.sergeev, javed.absar, asb, rbar, johnrusso, simoncook, apazos, sabuasal, niosHD, jrtc27, MaskRay, zzheng, edward-jones, atanasyan, rogfer01, MartinMosbeck, brucehoult, the_o, tpr, PkmX, jocewei, jsji, Petar.Avramovic, asbirlea, Jim, s.egerton, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D65962
llvm-svn: 369041
If the last step in an FP add reduction allows reassociation and doesn't care
about -0.0, then we are free to recognize that computation as a reduction
that may reorder the intermediate steps.
This is requested directly by PR42705:
https://bugs.llvm.org/show_bug.cgi?id=42705
and solves PR42947 (if horizontal math instructions are actually faster than
the alternative):
https://bugs.llvm.org/show_bug.cgi?id=42947
Differential Revision: https://reviews.llvm.org/D66236
llvm-svn: 368995
If the width is 256 bits, then we must have AVX so the else here
was unnecessary. Once that's removed then the >= 256 bit code is
identical to the 128 bit code with a different VT so combine them.
llvm-svn: 368956
If the target shuffle mask is from a wider type, attempt to scale the mask so that the extraction can attempt to peek through.
Fixes the regression mentioned in rL368662
Reapplying this as rL368308 had to be reverted as part of rL368660 to revert rL368276
llvm-svn: 368663
If we don't demand all elements, then attempt to combine to a simpler shuffle.
At the moment we can only do this if Depth == 0 as combineX86ShufflesRecursively uses Depth to track whether the shuffle has really changed or not - we'll need to change this before we can properly start merging combineX86ShufflesRecursively into SimplifyDemandedVectorElts.
The insertps-combine.ll regression is because XFormVExtractWithShuffleIntoLoad can't see through shuffles of different widths - this will be fixed in a follow-up commit.
Reapplying this as rL368307 had to be reverted as part of rL368660 to revert rL368276
llvm-svn: 368662
This introduced a false positive MemorySanitizer warning about use of
uninitialized memory in a vectorized crc function in Chromium. That suggests
maybe something is not right with this transformation. See
https://crbug.com/992853#c7 for a reproducer.
This also reverts the follow-up commits r368307 and r368308 which
depended on this.
> This patch attempts to peek through vectors based on the demanded bits/elt of a particular ISD::EXTRACT_VECTOR_ELT node, allowing us to avoid dependencies on ops that have no impact on the extract.
>
> In particular this helps remove some unnecessary scalar->vector->scalar patterns.
>
> The wasm shift patterns are annoying - @tlively has indicated that the wasm vector shift codegen are to be refactored in the near-term and isn't considered a major issue.
>
> Differential Revision: https://reviews.llvm.org/D65887
llvm-svn: 368660
We need AVX512BW to be able to truncate an i16 vector. If we don't
have that we have to extend i16->i32, then trunc, i32->i8. But we
won't be able to remove the min/max if we do that. At least not
without more special handling.
llvm-svn: 368623
If we're after type legalize, we should make sure we won't create
a store with an illegal type when we separate the AVG pattern
from the truncating store.
I don't know of a way to fail for this today. Just noticed while
I was in the vicinity.
llvm-svn: 368608
The test case that changed is probably better served through
allowing combineTruncatedArithmetic to create narrow vectors. It
also appears InstCombine would have simplified this test case
to remove the zext and trunc anyway.
llvm-svn: 368522
On SSE41+ targets we always lower vector shuffles to ZERO_EXTEND_VECTOR_INREG, even if we don't need the extended bits.
This patch relaxes this so that we lower to ANY_EXTEND_VECTOR_INREG if we can, meaning that shuffle combines have a better idea of what elements need to be kept zero. This helps the multiple reduction code as we can now combine away a lot more of the pack+extend codes.
Differential Revision: https://reviews.llvm.org/D65741
llvm-svn: 368515
Summary:
This patch adds a special DAG combine for SSE1 to recognize the IR pattern InstCombine gives us for movmsk. This only does the recognition for a few cases where its obvious the input won't be scalarized resulting in building a vector just do to the movmsk. I've made it separate from our existing matching for movmsk since that's called in multiple places and I didn't spend time to see if the other callers would make sense here. Plus the restrictions and additional checks would complicate that.
This fixes the case from PR42870. Buts its probably still broken the presence of logic ops feeding the movmsk pattern which would further hide the v4f32 type.
Reviewers: spatel, RKSimon, xbolva00
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D65689
llvm-svn: 368506
Summary:
On windows if the frame size exceed 4096 bytes, compiler need to
generate a call to _alloca_probe. X86CallFrameOptimization pass
changes the reserved stack size and cause of stack probe function
not be inserted. This patch fix the issue by detecting the call
frame size, if the size exceed 4096 bytes, drop X86CallFrameOptimization.
Reviewers: craig.topper, wxiao3, annita.zhang, rnk, RKSimon
Reviewed By: rnk
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D65923
llvm-svn: 368503
As discussed on PR42825, if we are inverting the selection mask we can just swap the inputs and avoid the inversion.
Differential Revision: https://reviews.llvm.org/D65522
llvm-svn: 368438
The only way to generate these was through promoting legalization
of narrow vectors, but we widen those types now. So we shouldn't
produce these nodes.
llvm-svn: 368396
Under this configuration we'll want to split the v8i64 or v16i32 into two vectors. The default legalization will try to truncate each of those 256-bit pieces one step to 128-bit, concatenate those, then truncate one more time from the new 256 to 128 bits.
With this patch we now truncate the two splits to 64-bits then concatenate those. We have to do this two different ways depending on whether have widening legalization enabled. Without widening legalization we have to manually construct X86ISD::VTRUNC to prevent the ISD::TRUNCATE with a narrow result being promoted to 128 bits with a larger element type than what we want followed by something like a pshufb to grab the lower half of each element to finish the job. With widening legalization we just get the right thing. When we switch to widening by default we can just delete the other code path.
Differential Revision: https://reviews.llvm.org/D65626
llvm-svn: 368349
If the target shuffle mask is from a wider type, attempt to scale the mask so that the extraction can attempt to peek through.
Fixes the regression mentioned in rL368307
llvm-svn: 368308
If we don't demand all elements, then attempt to combine to a simpler shuffle.
At the moment we can only do this if Depth == 0 as combineX86ShufflesRecursively uses Depth to track whether the shuffle has really changed or not - we'll need to change this before we can properly start merging combineX86ShufflesRecursively into SimplifyDemandedVectorElts.
The insertps-combine.ll regression is because XFormVExtractWithShuffleIntoLoad can't see through shuffles of different widths - this will be fixed in a follow-up commit.
llvm-svn: 368307
If we're splitting the 512-bit vector anyway and we have zero/sign bits, then we might as well use pack instructions to concat and truncate at once.
Differential Revision: https://reviews.llvm.org/D65904
llvm-svn: 368210
The assert that caused this to be reverted should be fixed now.
Original commit message:
This patch changes our defualt legalization behavior for 16, 32, and
64 bit vectors with i8/i16/i32/i64 scalar types from promotion to
widening. For example, v8i8 will now be widened to v16i8 instead of
promoted to v8i16. This keeps the elements widths the same and pads
with undef elements. We believe this is a better legalization strategy.
But it carries some issues due to the fragmented vector ISA. For
example, i8 shifts and multiplies get widened and then later have
to be promoted/split into vXi16 vectors.
This has the potential to cause regressions so we wanted to get
it in early in the 10.0 cycle so we have plenty of time to
address them.
Next steps will be to merge tests that explicitly test the command
line option. And then we can remove the option and its associated
code.
llvm-svn: 368183
If we're after type legalization we should only be trying to turn
v2i64 into v2i32. So bitcast to v4i32, shuffle the even elements
together. Then use X86ISD::CVTSI2P. The alternative is to leave
the v2i64 type alone and let it scalarized. Hopefully keeping
it packed is better.
Fixes PR42905.
llvm-svn: 368091
If we don't demand any non-undef shuffle elements then the assert will fail as all shuffle inputs would still be flagged as 'identity' safe.
Exposed by an incoming patch.
llvm-svn: 368022
Summary:
Before this patch MGATHER/MSCATTER is capable of representing all
common addressing modes, but only when illegal types are used.
This patch adds an IndexType property so more representations
are available when using legal types only.
Original modes:
vector of bases
base + vector of signed scaled offsets
New modes:
base + vector of signed unscaled offsets
base + vector of unsigned scaled offsets
base + vector of unsigned unscaled offsets
The current behaviour of addressing modes for gather/scatter remains
unchanged.
Patch by Paul Walker.
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D65636
llvm-svn: 368008
This patch changes our defualt legalization behavior for 16, 32, and
64 bit vectors with i8/i16/i32/i64 scalar types from promotion to
widening. For example, v8i8 will now be widened to v16i8 instead of
promoted to v8i16. This keeps the elements widths the same and pads
with undef elements. We believe this is a better legalization strategy.
But it carries some issues due to the fragmented vector ISA. For
example, i8 shifts and multiplies get widened and then later have
to be promoted/split into vXi16 vectors.
This has the potential to cause regressions so we wanted to get
it in early in the 10.0 cycle so we have plenty of time to
address them.
Next steps will be to merge tests that explicitly test the command
line option. And then we can remove the option and its associated
code.
llvm-svn: 367901
The test case is based on the example from the post-commit thread for:
https://reviews.llvm.org/rGc9171bd0a955
This replaces the x86-specific simple-type check from:
rL367766
with a check in the DAGCombiner. Adding the check isn't
strictly necessary after the fix from:
rL367768
...but it seems likely that we're heading for trouble if
we are creating weird types in this transform.
I combined the earlier legality check into the initial
clause to simplify the code.
So we should only try the trunc/sext transform at the
earliest combine stage, but we limit the transform to
simple types anyway because the TLI hook is probably
too lax about what it considers a free truncate.
llvm-svn: 367834
Summary:
This is patch is part of a serie to introduce an Alignment type.
See this thread for context: http://lists.llvm.org/pipermail/llvm-dev/2019-July/133851.html
See this patch for the introduction of the type: https://reviews.llvm.org/D64790
Reviewers: courbet, jfb, jakehehrlich
Reviewed By: jfb
Subscribers: wuzish, jholewinski, arsenm, dschuff, nemanjai, jvesely, nhaehnle, javed.absar, sbc100, jgravelle-google, hiraditya, aheejin, kbarton, asb, rbar, johnrusso, simoncook, apazos, sabuasal, niosHD, jrtc27, MaskRay, zzheng, edward-jones, rogfer01, MartinMosbeck, brucehoult, the_o, dexonsmith, PkmX, jocewei, jsji, s.egerton, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D65514
llvm-svn: 367828
Summary:
The SimplifyDemandedVectorElts function can replace with undef
when no elements are demanded, but due to how it interacts with
TargetLoweringOpts, it can only do this when the node has
no other users.
Remove a now unneeded DAG combine from the X86 backend.
Reviewers: RKSimon, spatel
Reviewed By: RKSimon
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D65713
llvm-svn: 367788
This avoids the crash from:
https://bugs.llvm.org/show_bug.cgi?id=42880
...and I think it's a proper constraint for the TLI hook.
But that example raises questions about what happens to get us
into this situation (created i29 types) and what happens later
(why does legalization die on those types), so I'm not sure if
we will resolve the bug based on this change.
llvm-svn: 367766
This is consistent with the target independent intrinsic handling.
Not sure this really matters since we just pull the constant out
using getZExtValue later.
llvm-svn: 367736
If a type is larger than a legal type and needs to be split, we would previously allow the multiply to be decomposed even if the split multiply is legal. Since the shift + add/sub code would also need to be split, its not any better to decompose it.
This patch figures out what type the mul will eventually be legalized to and then uses that type for the query. I tried just returning false illegal types and letting them get handled after type legalization, but then we can't recognize and i64 constant splat on 32-bit targets since will be destroyed by type legalization. We could special case vectors of i64 to avoid that...
Differential Revision: https://reviews.llvm.org/D65533
llvm-svn: 367601
This adds SimplifyMultipleUseDemandedBitsForTargetNode X86 support and uses it to allow us to peek through vector insertions to avoid dependencies on entire insertion chains.
llvm-svn: 367570
We have custom code that ignores the normal promoting type legalization on less than 128-bit vector types like v4i8 to emit pavgb, paddusb, psubusb since we don't have the equivalent instruction on a larger element type like v4i32. If this operation appears before a store, we can be left with an any_extend_vector_inreg followed by a truncstore after type legalization. When truncstore isn't legal, this will normally be decomposed into shuffles and a non-truncating store. This will then combine away the any_extend_vector_inreg and shuffle leaving just the store. On avx512, truncstore is legal so we don't decompose it and we had no combines to fix it.
This patch adds a new DAG combine to detect this case and emit either an extract_store for 64-bit stoers or a extractelement+store for 32 and 16 bit stores. This makes the avx512 codegen match the avx2 codegen for these situations. I'm restricting to only when -x86-experimental-vector-widening-legalization is false. When we're widening we're not likely to create this any_extend_inreg+truncstore combination. This means we should be able to remove this code when we flip the default. I would like to flip the default soon, but I need to investigate some performance regressions its causing in our branch that I wasn't seeing on trunk.
Differential Revision: https://reviews.llvm.org/D65538
llvm-svn: 367488
Before combining insert_subvector(insert_subvector(vec, sub0, c0), sub1, c1) patterns, ensure that the subvectors are all the same type. On AVX512 targets especially we might have a mixture of 128/256 subvector insertions.
llvm-svn: 367429
PR42819 showed an issue that we couldn't handle the case where we demanded a 'sub-sub-vector' of the SUBV_BROADCAST 'sub-vector' source.
This patch recognizes these cases and extracts the sub-sub-vector instead of trying to broadcast to a type smaller than the 'sub-vector' source.
llvm-svn: 367306
As discussed on rL367171, we have a problem where the depth recursion used in combineX86ShufflesRecursively was subtly different to computeKnownBits etc. - it starts at Depth=1 instead of Depth=0 like the others and has a different maximum recursion depth.
This NFC patch fixes the recursion depth to start at 0, so we can more easily reuse depth values in calls from combineX86ShufflesRecursively and its helper functions in computeKnownBits etc.
llvm-svn: 367232
The pmaddwd inserts a truncate, if that truncate would end up
creating additional instructions instead of making a zext
narrower, then we shouldn't do it.
I've restricted this to only sse4.1 targets since on prior
targets the zext will be done in stages. So the truncate will
probably not create additional instructions. Might need some
more investigation of mul shrinking and the other pmaddwd
transform to be sure this is the right decision.
There might be a slight regression on AVX1 targets due to add
splitting. Hard to say for sure. Maybe we need to look into
using the vector reduction flag to use 2 narrow loads and a
blend instead of extracting and inserting.
llvm-svn: 367198
This reverts r340478 and r340631 and replaces them with a simpler
method of just letting DAG combine revisit the nodes to handle
the other operand.
llvm-svn: 367195
Recommit rL367100 which was reverted at rL367141. Until PR42777 is fixed, we no longer get the benefits of peeking through bitcasts but it does still remove a GetDemandedBits user and gives us the equivalent combines.
llvm-svn: 367172
Summary:
This was originally reported in D62818.
https://rise4fun.com/Alive/oPH
InstCombine does the opposite fold, in hope that `C l>>/<< Y` expression
will be hoisted out of a loop if `Y` is invariant and `X` is not.
But as it is seen from the diffs here, if it didn't get hoisted,
the produced assembly is almost universally worse.
Much like with my recent "hoist add/sub by/from const" patches,
we should get almost universal win if we hoist constant,
there is almost always an "and/test by imm" instruction,
but "shift of imm" not so much, so we may avoid having to
materialize the immediate, and thus need one less register.
And since we now shift not by constant, but by something else,
the live-range of that something else may reduce.
Special care needs to be applied not to disturb x86 `BT` / hexagon `tstbit`
instruction pattern. And to not get into endless combine loop.
Reviewers: RKSimon, efriedma, t.p.northover, craig.topper, spatel, arsenm
Reviewed By: spatel
Subscribers: hiraditya, MaskRay, wuzish, xbolva00, nikic, nemanjai, jvesely, wdng, nhaehnle, javed.absar, tpr, kristof.beyls, jsji, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D62871
llvm-svn: 366955
This patch adds support for recognizing cases where a larger vector type is being used to reduce just the elements in the lower subvector:
e.g. <8 x i32> reduction pattern in a <16 x i32> vector:
<4,5,6,7,u,u,u,u,u,u,u,u,u,u,u,u>
<2,3,u,u,u,u,u,u,u,u,u,u,u,u,u,u>
<1,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u>
matchBinOpReduction returns the lower extracted subvector in such cases, assuming isExtractSubvectorCheap accepts the extraction.
I've only enabled it for X86 reduction sums so far. I intend to enable it for the bitop/minmax cases in future patches, and eventually I think its worth turning it on all the time. This is mainly just a case of ensuring calls to matchBinOpReduction don't make assumptions on the vector width based on the original vector extraction.
Fixes the x86 partial reduction sum cases in PR33758 and PR42023.
Differential Revision: https://reviews.llvm.org/D65047
llvm-svn: 366933
LegalizeDAG tries to legal the DAG by legalizing nodes before
their operands.
If we create a new node, we end up legalizing it after its operands.
This prevents some of the optimizations that can be done when the
operand is a build_vector since the build_vector will have been
legalized to something else.
Differential Revision: https://reviews.llvm.org/D65132
llvm-svn: 366835
The build_vector will become a constant pool load. By using the
desired type initially, it ensures we don't generate a bitcast
of the constant pool load which will need to be folded with
the load.
While experimenting with another patch, I noticed that when the
load type and the constant pool type don't match, then
SimplifyDemandedBits can't handle it. While we should probably
fix that, this was a simple way to fix the issue I saw.
llvm-svn: 366732
This patch enables us to find the source loads for each element, splitting them into a Load and ByteOffset, and attempts to recognise consecutive loads that are in fact from the same source load.
A helper function, findEltLoadSrc, recurses to find a LoadSDNode and determines the element's byte offset within it. When attempting to match consecutive loads, byte offsetted loads then attempt to matched against a previous load that has already been confirmed to be a consecutive match.
Next step towards PR16739 - after this we just need to account for shuffling/repeated elements to create a vector load + shuffle.
Fixed out of bounds load assert identified in rL366501
Differential Revision: https://reviews.llvm.org/D64551
llvm-svn: 366681
As detailed on PR42674, we can reduce a vXi8 down until we have the final <8 x i8>, and then use PSADBW with zero, to sum those values. We then extract the bottom i8, discarding any overflow from the upper bits of the i16 result.
llvm-svn: 366636
This patch enables us to find the source loads for each element, splitting them into a Load and ByteOffset, and attempts to recognise consecutive loads that are in fact from the same source load.
A helper function, findEltLoadSrc, recurses to find a LoadSDNode and determines the element's byte offset within it. When attempting to match consecutive loads, byte offsetted loads then attempt to matched against a previous load that has already been confirmed to be a consecutive match.
Next step towards PR16739 - after this we just need to account for shuffling/repeated elements to create a vector load + shuffle.
Differential Revision: https://reviews.llvm.org/D64551
llvm-svn: 366441
I'm not convinced the code this calls is properly vetted for
vXi1 vectors. Experimental vector widening legalization testing
for D55251 is now hitting an assertion failure inside
EltsFromConsecutiveLoads. This is occurring from a v2i1 load
having a store size different than its VT size. Hopefully
this commit will keep such issues from happening.
llvm-svn: 366405
This is part of what is requested by PR42023:
https://bugs.llvm.org/show_bug.cgi?id=42023
There's an extension needed for FP add, but exactly how we would specify
that using flags is not clear to me, so I left that as a TODO.
We're still missing patterns for partial reductions when the input vector
is 256-bit or 512-bit, but I think that's a failure of vector narrowing.
If we can reduce the widths, then this matching should work on those tests.
Differential Revision: https://reviews.llvm.org/D64760
llvm-svn: 366268
inttofp (trunc (extelt X, 0)) --> inttofp (extelt (bitcast X), 0)
We have pseudo-vectorization of scalar int to FP casts, so this tries to
make that more likely by replacing a truncate with a bitcast. I didn't see
any test diffs starting from 'uitofp', so I left that as a TODO. We can't
only match the shorter trunc+extract pattern because there's an opposing
transform somewhere, so we infinite loop. Waiting to try this during
lowering is another possibility.
A motivating case is shown in PR39975 and included in the test diffs here:
https://bugs.llvm.org/show_bug.cgi?id=39975
Differential Revision: https://reviews.llvm.org/D64710
llvm-svn: 366098
I think we only turn out of range shiftss to undef when
all elements are out of range or the shift amount is a splat out
of range. I'm not sure which, I didn't check.
During lowering we can split a shift where some elements
are out of range into multiple shifts. This can create a
new shift with a splat shift amount that is out of range.
This patch returns undef for this case.
Fixes PR42615.
Differential Revision: https://reviews.llvm.org/D64699
llvm-svn: 366096
Summary:
SSE1 only supports v4f32. But does have instructions like movlps/movhps that load/store 64-bits of memory.
This patch breaks the connection between the node VT of the vzext_load/vextract_store patterns and the memory VT. Enabling a v4f32 node with a 64-bit memory VT. I've used i64 as the memory VT here. I've written the PatFrag predicate to just check the store size not the specific VT. I think the VT will only matter for CSE purposes. We could use v2f32, but if we want to start using these operations in more places a simple integer type might make the most sense.
I'd like to maybe use this same thing for SSE2 and later as well, but that will need more work to be supported by EltsFromConsecutiveLoads to avoid regressing lit tests. I'd maybe also like to combine bitcasts with these load/stores nodes now that the types are disconnected. And I'd also like to consider canonicalizing (scalar_to_vector + load) to vzext_load.
If you want I can split the mechanical tablegen stuff where I added the 32/64 off from the sse1 change.
Reviewers: spatel, RKSimon
Reviewed By: RKSimon
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D64528
llvm-svn: 366034
Follow up to D58597, where it was noted that the commuted ISD::SUB variant
was having problems with lack of combines.
See also D63958 where we untangled setcc/sub pairs.
Differential Revision: https://reviews.llvm.org/D58875
llvm-svn: 365791
This renames the type so it doesn't sound like its based off the load size - as we're moving towards supporting combining loads of different sizes.
llvm-svn: 365655
Cache the LoadSDNode nodes so we can easily map to/from the element index instead of packing them together - this will be useful for future patches for PR16739 etc.
llvm-svn: 365620
This patch checks to see if the vector element loads are based off a dereferenceable pointer that covers the entire vector width, in which case we don't need to have element loads at both extremes of the vector width - just the start (base pointer) of it.
Another step towards partial vector loads......
Differential Revision: https://reviews.llvm.org/D64205
llvm-svn: 365614
This should prevent doing this on pre-sse4.1 targets or for 256
bit vectors without avx2.
I don't know of a failure from this. Op legalization will probably
take care of, but seemed better to be safe.
llvm-svn: 365577
Basically the problem is that X86 doesn't set the Fast flag from
allowsMemoryAccess on certain CPUs due to slow unaligned memory
subtarget features. This prevents bitcasts from being folded into
loads and stores. But all vector loads and stores of the same width
are the same cost on X86.
This patch merges the allowsMemoryAccess call into isLoadBitCastBeneficial to allow X86 to skip it.
Differential Revision: https://reviews.llvm.org/D64295
llvm-svn: 365549
Summary:
This makes it so that IR files using triples without an environment work
out of the box, without normalizing them.
Typically, the MSVC behavior is more desirable. For example, it tends to
enable things like constant merging, use of associative comdats, etc.
Addresses PR42491
Reviewers: compnerd
Subscribers: hiraditya, dexonsmith, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D64109
llvm-svn: 365387
Bitcast v4i32 to v8f32 and back again - it might be worth adding isel patterns for X86PShufd v8i32 on AVX1 targets like we did for X86Blendi to avoid the bitcasts?
llvm-svn: 365125
If we have more then 2 shuffle ops to combine, try to use combineX86ShuffleChainWithExtract to see if some are from the same super vector.
llvm-svn: 365050
iff the number of elements doesn't change.
This gets around an issue with combineX86ShuffleChain not being able to hint which domain is preferred for shuffles that can be done with either.
Fixes regression introduced in rL365041
llvm-svn: 365044
This better accounts for the cost/benefit of removing extract_subvectors from the shuffle and will be more useful in future patches.
The vpermq predicate regression will be fixed shortly.
llvm-svn: 365041
Assert that the shift amount is in range and create vXi8 shift masks in a way that doesn't cause MSVC/cppcheck shift result is truncated then extended warnings.
llvm-svn: 365024
Don't use APInt::getZExtValue() if you can avoid it - eventually someone will call it with i128 or something that doesn't fit into 64-bits.
In this case it was completely superfluous as we'd moved the rest of the code to always use APInt.
Fixes the <1 x i128> addition bug in PR42486
llvm-svn: 364953
Pull out CombineShuffleWithExtract lambda to new combineX86ShuffleChainWithExtract wrapper and refactored it to handle more than 2 shuffle inputs - this will allow combineX86ShufflesRecursively to call this in a future patch.
llvm-svn: 364924
We were relying on combineX86ShufflesRecursively to handle this - this patch gets it done earlier which should make it easier for other code to use resolveTargetShuffleInputsAndMask.
llvm-svn: 364906
These instructions only read 64-bits of memory so we shouldn't
allow a full vector width load to be pattern matched in case it
is marked volatile.
Instead allow vzload or scalar_to_vector+load.
Also add a DAG combine to turn full vector loads into vzload when
used by one of these instructions if the load isn't volatile.
This fixes another case for PR42079
llvm-svn: 364838
We can already widenSubVector to a specific type (of the same scalar type) - this variant just specifies the target vector size.
This will be useful when CombineShuffleWithExtract relaxes the need to have the same scalar type for all shuffle operand subvector sources.
llvm-svn: 364803
But only when the load isn't volatile.
This improves load folding during isel where we only have vzload
and scalar_to_vector+load patterns. We can't have full vector load
isel patterns for the same volatile load issue.
Also add some missing masked cvtsi2fp/cvtui2fp with vzload patterns.
llvm-svn: 364728
AVX masked loads only support 0 as the value for masked off elements.
So we need an extra blend to support other values. Previously we
expanded the masked load to two instructions with isel patterns.
With this patch we now insert the vselect during lowering and it
will be separately selected as a blend.
llvm-svn: 364718
We were requiring that both shuffle operands were EXTRACT_SUBVECTORs, but we can relax this to only require one of them to be.
Also, we shouldn't bother attempting this if both operands are from the lowest subvector (or not EXTRACT_SUBVECTOR at all).
llvm-svn: 364644
We already had the infrastructure for this, but were waiting for the fix for a number of regressions which were handled by the recent shuffle(extract_subvector(),extract_subvector()) -> extract_subvector(shuffle()) shuffle combines
llvm-svn: 364569
While lowering calls, collect info about registers that forward arguments
into following function frame. We store such info into the MachineFunction
of the call. This is used very late when dumping DWARF info about
call site parameters.
([9/13] Introduce the debug entry values.)
Co-authored-by: Ananth Sowda <asowda@cisco.com>
Co-authored-by: Nikola Prica <nikola.prica@rt-rk.com>
Co-authored-by: Ivan Baev <ibaev@cisco.com>
Differential Revision: https://reviews.llvm.org/D60715
llvm-svn: 364516
Without the fix gcc 7.4.0 complains with
../lib/Target/X86/X86ISelLowering.cpp: In function 'bool getFauxShuffleMask(llvm::SDValue, llvm::SmallVectorImpl<int>&, llvm::SmallVectorImpl<llvm::SDValue>&, llvm::SelectionDAG&)':
../lib/Target/X86/X86ISelLowering.cpp:6690:36: error: enumeral and non-enumeral type in conditional expression [-Werror=extra]
int Idx = (ZeroMask[j] ? SM_SentinelZero : (i + j + Ofs));
~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
cc1plus: all warnings being treated as errors
llvm-svn: 364507
This patch rewrites the loop iteration to only visit every other element starting with element 0. And we work on the "even" element and "next" element at the same time. The "First" logic has been moved to the bottom of the loop and doesn't run on every element. I believe it could create dangling nodes previously since we didn't check if we were going to use SCALAR_TO_VECTOR for the first insertion. I got rid of the "First" variable and just do a null check on V which should be equivalent. We also no longer use undef as the starting V for vectors with no zeroes to avoid false dependencies. This matches v8i16.
I've changed all the extends and OR operations to use MVT::i32 since that's what they'll be promoted to anyway. I've tried to use zero_extend only when necessary and use any_extend otherwise. This resulted in some improvements in tests where we are now able to promote aligned (i32 (extload i8)) to a 32-bit load.
Differential Revision: https://reviews.llvm.org/D63702
llvm-svn: 364469
This was trying to optimize concat_vectors with zero of setcc or
kand instructions. But I think it produced the same code we
produce for a concat_vectors with 0 even it it doesn't come from
one of those operations.
llvm-svn: 364463
Create a per-byte shuffle mask based on the computeKnownBits from each operand - if for each byte we have a known zero (or both) then it can be safely blended.
Fixes PR41545
llvm-svn: 364458
We have (almost) no target opcodes that have scalar/vector equivalents - for now assume we can't scalarize them (we can add exceptions if we need to).
llvm-svn: 364429
Ideally this needs to be a generic combine in DAGCombiner::visitEXTRACT_SUBVECTOR but there's some nasty regressions in aarch64 due to neon shuffles not handling bitcasts at all.....
llvm-svn: 364407
truncateVectorWithPACK is often used in conjunction with ComputeNumSignBits which struggles when peeking through bitcasts.
This fix tries to avoid bitcast(shuffle(bitcast())) patterns in the 256-bit 64-bit sublane shuffles so we can still see through at least until lowering when the shuffles will need to be bitcasted to widen the shuffle type.
llvm-svn: 364401
We currently have some isel patterns for treating vzmovl+load the same as vzload, but that shrinks the load which we shouldn't do if the load is volatile.
Rather than adding isel checks for volatile. This patch removes the patterns and teachs DAG combine to merge them into vzload when its legal to do so.
Differential Revision: https://reviews.llvm.org/D63665
llvm-svn: 364333
In LowerBuildVectorv16i8 we took care to use an any_extend if the first pair is in the lower 16-bits of the vector and no elements are 0. So bits [31:16] will be undefined. But we still emitted a vzext_movl to ensure that bits [127:32] are 0. If we don't need any zeroes we should be consistent and make all of 127:16 undefined.
In LowerBuildVectorv8i16 we can just delete the vzext_movl code because we only use the scalar_to_vector when there are no zeroes. So the vzext_movl is always unnecessary.
Found while investigating whether (vzext_movl (scalar_to_vector (loadi32)) patterns are necessary. At least one of the cases where they were necessary was where the loadi32 matched 32-bit aligned 16-bit extload. Seemed weird that we required vzext_movl for that case.
Differential Revision: https://reviews.llvm.org/D63700
llvm-svn: 364207
This patch does a few things to start cleaning up the isFNEG function.
-Remove the Op0/Op1 peekThroughBitcast calls that seem unnecessary. getTargetConstantBitsFromNode has its own peekThroughBitcast inside. And we have a separate peekThroughBitcast on the return value.
-Add a check of the scalar size after the first peekThroughBitcast to ensure we haven't changed the element size and just did something like f32->i32 or f64->i64.
-Remove an unnecessary check that Op1's type is floating point after the peekThroughBitcast. We're just going to look for a bit pattern from a constant. We don't care about its type.
-Add VT checks on several places that consume the return value of isFNEG. Due to the peekThroughBitcasts inside, the type of the return value isn't guaranteed. So its not safe to use it to build other nodes without ensuring the type matches the type being used to build the node. We might be able to replace these checks with bitcasts instead, but I don't have a test case so a bail out check seemed better for now.
Differential Revision: https://reviews.llvm.org/D63683
llvm-svn: 364206
Ideally we'd be able to represent this truncate as a any_extend to
v16i32 and a truncate, but SelectionDAG doens't know how to not
fold those together.
We have isel patterns to use a vpmovzxwd+vpdmovdb for the truncate,
but we aren't able to simultaneously fold the load and the store
from the isel pattern. By pulling the truncate into the store we
can successfully hide it from the DAG combiner. Then we can isel
pattern match the truncstore and load+any_extend separately.
llvm-svn: 364163
128/256 bit scalar_to_vectors are canonicalized to (insert_subvector undef, (scalar_to_vector), 0). We have isel patterns that try to match this pattern being used by a vzmovl to use a 128-bit instruction and a subreg_to_reg.
This patch detects the insert_subvector undef portion of this and pulls it through the vzmovl, creating a narrower vzmovl and an insert_subvector allzeroes. We can then match the insertsubvector into a subreg_to_reg operation by itself. Then we can fall back on existing (vzmovl (scalar_to_vector)) patterns.
Note, while the scalar_to_vector case is the motivating case I didn't restrict to just that case. I'm also wondering about shrinking any 256/512 vzmovl to an extract_subvector+vzmovl+insert_subvector(allzeros) but I fear that would have bad implications to shuffle combining.
I also think there is more canonicalization we can do with vzmovl with loads or scalar_to_vector with loads to create vzload.
Differential Revision: https://reviews.llvm.org/D63512
llvm-svn: 364095
This should be unreachable, but bugs can make it reachable. This
adds a debug print so we can see the bad node in the output when
the llvm_unreachable triggers.
llvm-svn: 364091
The sat add/sub tests still have unnecessary extract_subvector((vandnps ymm, ymm), 0) uses that should be split to (vandnps (extract_subvector(ymm, 0), extract_subvector(ymm, 0)), but its getting better.
llvm-svn: 364038
This is an exception to the rule that we should prefer xmm ops to ymm ops.
As shown in PR42305:
https://bugs.llvm.org/show_bug.cgi?id=42305
...the store folding opportunity with vextractf128 may result in better
perf by reducing the instruction count.
Differential Revision: https://reviews.llvm.org/D63517
llvm-svn: 363853
We already do this for ZERO_EXTEND/ZERO_EXTEND_VECTOR_INREG - this just extends the pattern matcher to recognize cases where we don't need the zeros in the extension.
llvm-svn: 363841
This allows targets to make more decisions about reserved registers
after isel. For example, now it should be certain there are calls or
stack objects in the frame or not, which could have been introduced by
legalization.
Patch by Matthias Braun
llvm-svn: 363757
FP_ROUND defaults to Legal for all MVT types and nothing changes
the v4f32 entry way from this default. If we needed this line
we'd also need one for v8f32 with AVX512 which we don't have.
llvm-svn: 363719
Part of fixing the X86 regression noted in D63281 - I've split this into X86 and generic parts - the generic commit will be coming shortly and will fix the vector-reduce-mul-widen.ll regression introduced here.
llvm-svn: 363693
If a XMM non-temporal store has less than natural alignment, scalarize the vector - with SSE4A we can stay on the vector and use MOVNTSD(f64), else we must move to GPRs and use MOVNTI(i32/i64).
llvm-svn: 363592
If a YMM/ZMM non-temporal store has less than natural alignment, split the vector - either they will be satisfactorily aligned or will continue to be split until they are XMMs - at which point the legalizer will scalarize it.
llvm-svn: 363582
This is currently only used for ymm->xmm splitting but we shouldn't hardcode the offsets/alignment.
This is necessary for an upcoming patch to split under-aligned non-temporal vector loads.
llvm-svn: 363570
For loads, pre-SSE41 we can't perform NT loads at all, and after that we can only perform vector aligned loads, so if the alignment is less than for a xmm we'll just end up using the regular unaligned vector loads anyway.
First step towards fixing PR42026 - the next step for stores will be to use SSE4A movntsd where possible and to avoid the stack spill on SSE2 targets.
Differential Revision: https://reviews.llvm.org/D63246
llvm-svn: 363564
This is similar logic/motivation to the select splitting in D62969.
In D63233, the pattern changes so that we no longer have an extract_subvector of vselect,
but the operands of the select are still being concatenated.
The closest case is represented in either the first or last test diffs here - we have an
extra instruction, but we converted 3-4 ymm instructions into 4-5 xmm instructions.
I think that's the right trade-off for most AVX1 targets.
In the example based on PR37428:
https://bugs.llvm.org/show_bug.cgi?id=37428
...this makes the loop about 30% faster (tested on Haswell by compiling with -mavx).
Differential Revision: https://reviews.llvm.org/D63364
llvm-svn: 363508
Previously it copied over MachineMemOperands verbatim which caused MOV32rm to have store flags set, and MOV32mr to have load flags set. This fixes some assertions being thrown with EXPENSIVE_CHECKS on.
Committed on behalf of @luke (Luke Lau)
Differential Revision: https://reviews.llvm.org/D62726
llvm-svn: 363268
As discussed on D62910, we need to check whether particular types of memory access are allowed, not just their alignment/address-space.
This NFC patch adds a MachineMemOperand::Flags argument to allowsMemoryAccess and allowsMisalignedMemoryAccesses, and wires up calls to pass the relevant flags to them.
If people are happy with this approach I can then update X86TargetLowering::allowsMisalignedMemoryAccesses to handle misaligned NT load/stores.
Differential Revision: https://reviews.llvm.org/D63075
llvm-svn: 363179
As suggested by @arsenm on D63075 - this adds a TargetLowering::allowsMemoryAccess wrapper that takes a Load/Store node's MachineMemOperand to handle the AddressSpace/Alignment arguments and will also implicitly handle the MachineMemOperand::Flags change in D63075.
llvm-svn: 363048
Summary:
Our default behavior is to use sign_extend for signed comparisons and zero_extend for everything else. But for equality we have the freedom to use either extension. If we can prove the input has been truncated from something with enough sign bits, we can use sign_extend instead and let DAG combine optimize it out. A similar rule is used by type legalization in LegalizeIntegerTypes.
This gets rid of the movzx in PR42189. The immediate will still take 4 bytes instead of the 2 bytes plus 0x66 prefix a cmp di, 32767 would get, but it avoids a length changing prefix.
Reviewers: RKSimon, spatel, xbolva00
Reviewed By: xbolva00
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D63032
llvm-svn: 362920
Summary:
We can only use the memory form of cvtss2sd under optsize due to a partial register update. So previously we were emitting 2 instructions for extload when optimizing for speed. Also due to a late optimization in preprocessiseldag we had to handle (fpextend (loadf32)) under optsize.
This patch forces extload to expand so that it will always be in the (fpextend (loadf32)) form during isel. And when optimizing for speed we can just let each of those pieces select an instruction independently.
Reviewers: spatel, RKSimon
Reviewed By: RKSimon
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D62710
llvm-svn: 362919
This is a potentially large perf win for AVX1 targets because of the way we
auto-vectorize to 256-bit but then expect the backend to legalize/optimize
for the half-implemented AVX1 ISA.
On the motivating example from PR37428 (even though this patch doesn't solve
the vector shift issue):
https://bugs.llvm.org/show_bug.cgi?id=37428
...there's a 16% speedup when compiling with "-mavx" (perf tested on Haswell)
because we eliminate the remaining 256-bit vblendv ops.
I added comments on a couple of tests that require further work. If we have
256-bit logic ops separating the vselect and extract, we should probably narrow
everything to 128-bit, but that requires a larger pattern match.
Differential Revision: https://reviews.llvm.org/D62969
llvm-svn: 362797
This is intended to enable the use of an immediate blend or
more optimal instruction. But if the passthru is zero we don't
need any additional instructions.
llvm-svn: 362675
As suggested in D62498 - collectConcatOps() matches both
concat_vectors and insert_subvector patterns, and we see
more test improvements by using the more general match.
llvm-svn: 362620
We already handle the case where we combine shuffle(extractsubvector(x),extractsubvector(x)), this relaxes the requirement to permit different sources as long as they have the same value type.
This causes a couple of cases where the VPERMV3 binary shuffles occur at a wider width than before, which I intend to improve in future commits - but as only the subvector's mask indices are defined, these will broadcast so we don't see any increase in constant size.
llvm-svn: 362599
-Use early returns to reduce indentation
-Replace multipe ifs with a switch.
-Replace an assert with an llvm_unreachable default in the switch.
-Check that the FP type we're going to use for the
X86ISD::FAND/FOR/FXOR is legal rather than checking that the
integer type matches the width of a legal scalar fp type. This all
runs after legalization so it shouldn't really matter, but making
sure we're using a valid type in the X86ISD node is really
whats important.
llvm-svn: 362565
This shows up as a side issue to the main problem for the AVX target example from PR37428:
https://bugs.llvm.org/show_bug.cgi?id=37428 - https://godbolt.org/z/7tpRa3
But as we can see in the pile of existing test diffs, it's actually a widespread problem
that affects any AVX or later target. Apart from a couple of oddballs, I think these are
all improvements for the reasons stated in the code comment: we do not want to enable YMM
unnecessarily (avoid vzeroupper and frequency throttling) and some cores split 256-bit
stores anyway.
We could say that MergeConsecutiveStores() is going overboard on some of these examples,
but that won't solve the problem completely. But that is a reason I'm proposing this as
a lowering rather than a combine: we will infinite loop fighting the merge code if we try
this earlier.
Differential Revision: https://reviews.llvm.org/D62498
llvm-svn: 362524
As discussed on D62777 - we should be able to use this in more SSE41+ cases as well but that requires us to separate it from the OR(AND(),ANDN()) matcher.
llvm-svn: 362504
Move this combine from x86 into generic DAGCombine, which currently only manages cases where the bitcast is between types of the same scalarsize.
Differential Revision: https://reviews.llvm.org/D59188
llvm-svn: 362324
The LoadExt table defaults to all combinations being Legal. For
vector types, only src VTs with an i1 element type were ever changed.
So we don't need to mark them legal manually.
llvm-svn: 362170
We already have good codegen for (vXiY *ext(vXi1 bitcast(iX))) cases, this patch uses it for loads of vXi1 types as well - changing the load into a iX integer load, and bitcasting so that combineToExtendBoolVectorInReg can then use it.
Differential Revision: https://reviews.llvm.org/D62449
llvm-svn: 362081
This patch add the ISD::LRINT and ISD::LLRINT along with new
intrinsics. The changes are straightforward as for other
floating-point rounding functions, with just some adjustments
required to handle the return value being an interger.
The idea is to optimize lrint/llrint generation for AArch64
in a subsequent patch. Current semantic is just route it to libm
symbol.
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D62017
llvm-svn: 361875
This shows up as a side issue to the main problem for the AVX target example from PR37428:
https://bugs.llvm.org/show_bug.cgi?id=37428 - https://godbolt.org/z/7tpRa3
But as we can see in the pile of existing test diffs, it's actually a widespread problem
that affects any AVX or later target. Apart from a couple of oddballs, I think these are
all improvements for the reasons stated in the code comment: we do not want to enable YMM
unnecessarily (avoid vzeroupper and frequency throttling) and some cores split 256-bit
stores anyway.
We could say that MergeConsecutiveStores() is going overboard on some of these examples,
but that won't solve the problem completely. But that is the reason I'm proposing this as
a lowering rather than a combine: we will infinite loop fighting the merge code if we try
this earlier.
Differential Revision: https://reviews.llvm.org/D62498
llvm-svn: 361822
Forking this out of the discussion in D62498
(and assuming that will be committed later, so adding the helper function here).
The LangRef says:
"the backend should never split or merge target-legal volatile load/store instructions."
Differential Revision: https://reviews.llvm.org/D62506
llvm-svn: 361815
We were only testing for direct SETCC results - this allows us to peek through AND/OR/XOR combinations of the comparison results as well.
There's a missing SEXT(PACKSS) fold that I need to investigate for v8i1 cases before I can enable it there as well.
llvm-svn: 361716
If we have a known non-nan operand, place it in the second operand
of fmin/fmax that is returned if either operand is nan.
Differential Revision: https://reviews.llvm.org/D62448
llvm-svn: 361704
This patch adds the overridable TargetLowering::getTargetConstantFromLoad function which allows targets to return any constant value loaded by a LoadSDNode node - only X86 makes use of this so far but everything should be in place for other targets.
computeKnownBits then uses this function to improve codegen, notably vector code after legalization.
A future commit will do the same for ComputeNumSignBits but computeKnownBits sees the bigger benefit.
This required a couple of fixes:
* SimplifyDemandedBits must early-out for getTargetConstantFromLoad cases to prevent infinite loops of constant regeneration (similar to what we already do for BUILD_VECTOR).
* Fix a DAGCombiner::visitTRUNCATE issue as we had trunc(shl(v8i32),v8i16) <-> shl(trunc(v8i16),v8i32) infinite loops after legalization on AVX512 targets.
Differential Revision: https://reviews.llvm.org/D61887
llvm-svn: 361620
Fix for https://bugs.llvm.org/show_bug.cgi?id=41971. Make the
combineVectorSizedSetCCEquality() transform more conservative by
checking that the bitcast to the vector type will be cheap/free
for both operands. I'm considering it cheap if it's a constant,
a load or already a vector. I've dropped the explicit check for
f128 because it should fall out naturally (in the cases where
it'd be detrimental).
Differential Revision: https://reviews.llvm.org/D62220
llvm-svn: 361352
Same as what we do for vector reductions in combineHorizontalPredicateResult, use movmsk+cmp for scalar (and(extract(x,0),extract(x,1)) reduction patterns.
llvm-svn: 361052
Summary:
This refactors four pieces of code that create SDNodes for references to
symbols:
- normal global address lowering (LEA, MOV, etc)
- callee global address lowering (CALL)
- external symbol address lowering (LEA, MOV, etc)
- external symbol address lowering (CALL)
Each of these pieces of code need to:
- classify the reference
- lower the symbol
- emit a RIP wrapper if needed
- emit a load if needed
- add offsets if needed
I think handling them all in one place will make the code easier to
maintain in the future.
Reviewers: craig.topper, RKSimon
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D61690
llvm-svn: 360952
This patch add the ISD::LROUND and ISD::LLROUND along with new
intrinsics. The changes are straightforward as for other
floating-point rounding functions, with just some adjustments
required to handle the return value being an interger.
The idea is to optimize lround/llround generation for AArch64
in a subsequent patch. Current semantic is just route it to libm
symbol.
llvm-svn: 360889
They encode the same way, but OR32mi8Locked sets hasUnmodeledSideEffects set
which should be stronger than the mayLoad/mayStore on LOCK_OR32mi8. I think
this makes sense since we are using it as a fence.
This also seems to hide the operation from the speculative load hardening pass
so I've reverted r360511.
llvm-svn: 360747
This was the portion split off D58632 so that it could follow the redzone API cleanup. Note that I changed the offset preferred from -8 to -64. The difference should be very minor, but I thought it might help address one concern which had been previously raised.
Differential Revision: https://reviews.llvm.org/D61862
llvm-svn: 360719
D61068 handled vector shifts, this patch does the same for scalars where there are similar number of pipes for shifts as bit ops - this is true almost entirely for AMD targets where the scalar ALUs are well balanced.
This combine avoids AND immediate mask which usually means we reduce encoding size.
Some tests show use of (slow, scaled) LEA instead of SHL in some cases, but thats due to particular shift immediates - shift+mask generate these just as easily.
Differential Revision: https://reviews.llvm.org/D61830
llvm-svn: 360684
This is a follow on to D58632, with the same logic. Given a memory operation which needs ordering, but doesn't need to modify any particular address, prefer to use a locked stack op over an mfence.
Differential Revision: https://reviews.llvm.org/D61863
llvm-svn: 360649
Returning SDValue() makes the caller think that nothing happened and it will
end up executing the Expand path. This generates extra nodes that will need to
be pruned as dead code.
Returning an ISD::MERGE_VALUES will tell the caller that we'd like to make a
change and it will take care of replacing uses. This will prevent falling into
the Expand path.
llvm-svn: 360627
These are updates to match how isel table would emit a LOCK_OR32mi8 node.
-Use i32 for the immediate zero even though only 8 bits are encoded.
-Use i16 for segment register.
-Use LOCK_OR32mi8 for idempotent atomic operations in 32-bit mode to match
64-bit mode. I'm not sure why OR32mi8Locked and LOCK_OR32mi8 both exist. The
only difference seems to be that OR32mi8Locked is marked as UnmodeledSideEffects=1.
-Emit an extra i32 result for the flags output.
I don't know if the types here really matter just noticed it was inconsistent
with normal behavior.
llvm-svn: 360619
Summary:
X86TargetLowering::LowerAsmOperandForConstraint had better support than
TargetLowering::LowerAsmOperandForConstraint for arbitrary depth
getelementpointers for "i", "n", and "s" extended inline assembly
constraints. Hoist its support from the derived class into the base
class.
Link: https://github.com/ClangBuiltLinux/linux/issues/469
Reviewers: echristo, t.p.northover
Reviewed By: t.p.northover
Subscribers: t.p.northover, E5ten, kees, jyknight, nemanjai, javed.absar, eraman, hiraditya, jsji, llvm-commits, void, craig.topper, nathanchance, srhines
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D61560
llvm-svn: 360604
Fixes the regression noted in D61782 where a VZEXT_MOVL was being inserted because we weren't discriminating between 'zeroable' and 'all undef' for the upper elts.
Differential Revision: https://reviews.llvm.org/D61782
llvm-svn: 360596
Now that we can use HADD/SUB for scalar additions from any pair of extracted elements (D61263), we can relax the one use limit as we will be able to merge multiple uses into using the same HADD/SUB op.
This exposes a couple of missed opportunities in LowerBuildVectorv4x32 which will be committed separately.
Differential Revision: https://reviews.llvm.org/D61782
llvm-svn: 360594
See if we can simplify the demanded vector elts from the extraction before trying to simplify the demanded bits.
This helps us with target shuffles and hops in particular.
llvm-svn: 360535
If we only use the lower xmm of a ymm hop, then extract the xmm's (for free), perform the xmm hop and then insert back into a ymm (for free).
Fixes some of the regressions noted in D61782
llvm-svn: 360435
The current lowering uses an mfence. mfences are substaintially higher latency than the locked operations originally requested, but we do want to avoid contention on the original cache line. As such, use a locked instruction on a cache line assumed to be thread local.
Differential Revision: https://reviews.llvm.org/D58632
llvm-svn: 360393
As reported on PR39920, "slow horizontal ops" targets tend to internally expand to 2*shuffle+add/sub - so if we can reduce 2*shuffle+add/sub to a hadd/sub then we should do it - similar port usage but reduced instruction count.
This works out in most cases, although the "PR22377" regression in vector-shuffle-combining.ll is annoying - going from 2*shuffle+add+shuffle to hadd+2*shuffle - I've opened PR41813 to cover this.
Differential Revision: https://reviews.llvm.org/D61308
llvm-svn: 360360
Summary:
A COFF stub indirects the reference to a symbol through memory. A
.refptr.$sym global variable pointer is created to refer to $sym.
Typically mingw uses these for external global variable declarations,
but we can use them for weak function declarations as well.
Updates the dso_local classification to add a special case for
extern_weak symbols on COFF in both clang and LLVM.
Fixes PR37598
Reviewers: smeenai, mstorsjo
Subscribers: hiraditya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D61615
llvm-svn: 360207
Basic "revectorization" combine, we can probably do more opcodes here but it can be a tricky cost-benefit depending on where the subvectors came from - but this case helps shuffle combining.
llvm-svn: 360134
The FR32/FR64/VR128/VR256 register classes don't contain the upper 16 registers. For most cases we use the default implementation which will find any register class that contains the register in question if the VT is legal for the register class. But if the VT is i32 or i64, we won't find a matching register class and will instead up in the code modified in this patch.
If the requested register is x/y/zmm16-31 we weren't returning a register class that contains those registers and will hit an assertion in the caller.
To fix this, I've changed to use the extended register class instead. I don't believe we need a subtarget check to see if avx512 is enabled. The default implementation just pick whatever register class it finds first. I checked and we currently pick FR32X for XMM0 with an f32 type using the default implementation regardless of whether avx512 is enabled. So I assume its it is ok to do the same for i32.
Differential Revision: https://reviews.llvm.org/D61457
llvm-svn: 360102
Summary:
1. Enable infrastructure of AVX512_BF16, which is supported for BFLOAT16 in Cooper Lake;
2. Enable VCVTNE2PS2BF16, VCVTNEPS2BF16 and DPBF16PS instructions, which are Vector Neural Network Instructions supporting BFLOAT16 inputs and conversion instructions from IEEE single precision.
VCVTNE2PS2BF16: Convert Two Packed Single Data to One Packed BF16 Data.
VCVTNEPS2BF16: Convert Packed Single Data to Packed BF16 Data.
VDPBF16PS: Dot Product of BF16 Pairs Accumulated into Packed Single Precision.
For more details about BF16 isa, please refer to the latest ISE document: https://software.intel.com/en-us/download/intel-architecture-instruction-set-extensions-programming-reference
Author: LiuTianle
Reviewers: craig.topper, smaslov, LuoYuanke, wxiao3, annita.zhang, RKSimon, spatel
Reviewed By: craig.topper
Subscribers: kristina, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D60550
llvm-svn: 360017
The default impementation in the base class for TargetLowering::getRegForInlineAsmConstraint doesn't work for mask registers when the VT is a scalar type integer types since the only legal mask types are vXi1. So we end up just getting whatever the first register class that contains the register. Currently this appears to be VK1, but its really dependent on the order tablegen outputs the register classes.
Some code in the caller ends up looking up the type for this register class and find v1i1 then generates a copyfromreg from the physical k-register with the v1i1 type. Then it generates an any_extend from v1i1 to the scalar VT which isn't legal. This bad any_extend sticks around until isel where it selects a MOVZX32rr8 with a v1i1 input or maybe a i8 input. Not sure but eventually we pick up a copy from VK1 to GR8 in MachineIR which isn't supported. This leads to a failure in physical register copying.
This patch uses the scalar type to find a VK class of the right size. In the attached test case this will be VK16. This causes a bitcast from vk16 to i16 to be generated instead of an any_extend. This will be properly iseled to a VK16 to GR32 copy and a GR32->GR16 extract_subreg.
Fixes PR41678
Differential Revision: https://reviews.llvm.org/D61453
llvm-svn: 359837
Limiting scalar hadd/hsub generation to the lowest xmm looks to be unnecessary - we will be extracting one upper xmm whatever, and we can remove a shuffle by using the hop which is inline with what shouldUseHorizontalOp expects to happen anyway.
Testing on btver2 (the main target for fast-hops) shows this is beneficial even for float ops where we have a 'shuffle' to extract the float result:
https://godbolt.org/z/0R-U-K
Differential Revision: https://reviews.llvm.org/D61426
llvm-svn: 359786
We already perform horizontal add/sub if we extract from elements 0 and 1, this patch extends it to non-0/1 element extraction indices (as long as they are from the lowest 128-bit vector).
Differential Revision: https://reviews.llvm.org/D61263
llvm-svn: 359707
This is an alternative to D59669 which more aggressively extracts i1 elements from vXi1 bool vectors using a MOVMSK.
Differential Revision: https://reviews.llvm.org/D61189
llvm-svn: 359666
Current LLVM uses pxor+pinsrb on SSE4+ for INSERT_VECTOR_ELT(ZeroVec, 0, Elt) insead of much simpler movd.
INSERT_VECTOR_ELT(ZeroVec, 0, Elt) is idiomatic construct which is used e.g. for _mm_cvtsi32_si128(Elt) and for lowest element initialization in _mm_set_epi32.
So such inefficient lowering leads to significant performance digradations in ceratin cases switching from SSSE3 to SSE4.
https://bugs.llvm.org/show_bug.cgi?id=41512
Here INSERT_VECTOR_ELT(ZeroVec, 0, Elt) is simply converted to SCALAR_TO_VECTOR(Elt) when applicable since latter is closer match to desired behavior and always efficiently lowered to movd and alike.
Committed on behalf of @Serge_Preis (Serge Preis)
Differential Revision: https://reviews.llvm.org/D60852
llvm-svn: 359545
The MachineFunction wasn't used in getOptimalMemOpType, but more importantly,
this allows reuse of findOptimalMemOpLowering that is calling getOptimalMemOpType.
This is the groundwork for the changes in D59766 and D59787, that allows
implementation of TTI::getMemcpyCost.
Differential Revision: https://reviews.llvm.org/D59785
llvm-svn: 359537
Add target shuffle decoding to isHorizontalBinOp as well as ISD::VECTOR_SHUFFLE support.
This does mean we can go through bitcasts so we need to bitcast the extracted args to ensure they are the correct type
Fixes PR39936 and should help with PR39920/PR39921
Differential Revision: https://reviews.llvm.org/D61245
llvm-svn: 359491
Some of the combines might be further improved if we lower more shuffles with X86ISD::VPERMV3 directly, instead of waiting to combine the results.
llvm-svn: 359400
An xor reduction of a bool vector can be optimized to a parity check of the MOVMSK/BITCAST'd integer - if the population count is odd return 1, else return 0.
Differential Revision: https://reviews.llvm.org/D61230
llvm-svn: 359396
As predicate masks are legal on AVX512 targets, we avoid MOVMSK in these cases, but we can just bitcast the bool vector to the integer equivalent directly - avoiding expansion of the reduction to a shuffle pattern.
llvm-svn: 359386
Fixes PR40332 in the limited case where we're selecting between a target shuffle and a zero vector.
We can extend this in the future to handle more opcodes and non-zero selections.
llvm-svn: 359378
Summary: If we have SSE2 we can use a MOVQ to store 64-bits and avoid falling back to a cmpxchg8b loop. If its a seq_cst store we need to insert an mfence after the store.
Reviewers: spatel, RKSimon, reames, jfb, efriedma
Reviewed By: RKSimon
Subscribers: hiraditya, dexonsmith, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D60546
llvm-svn: 359368
Create a matchBitOpReduction helper that checks for the pattern with any opcode.
First step towards reusing this code to recognize other scalar reduction patterns.
llvm-svn: 359296
As detailed on PR40758, Bobcat/Jaguar can perform vector immediate shifts on the same pipes as vector ANDs with the same latency - so it doesn't make sense to replace a shl+lshr with a shift+and pair as it requires an additional mask (with the extra constant pool, loading and register pressure costs).
Differential Revision: https://reviews.llvm.org/D61068
llvm-svn: 359293
A small step towards combining shuffles across vector sizes - this recognizes when a shuffle's operands are all extracted from the same larger source and tries to combine to an unary shuffle of that source instead. Fixes one of the test cases from PR34380.
Differential Revision: https://reviews.llvm.org/D60512
llvm-svn: 359292
Truncate the movmsk scalar integer result to the equivalent scalar integer width as before but then bitcast to the requested type.
We still have the issue identified in PR41594 but D61114 should handle this.
llvm-svn: 359176
If the types don't match, we can't just remove the shuffle.
There may be some other opportunity for optimization here,
but this should prevent the crashing seen in:
https://bugs.llvm.org/show_bug.cgi?id=41414
llvm-svn: 359095
Circling back to a leftover bit from PR39859:
https://bugs.llvm.org/show_bug.cgi?id=39859#c1
...we have this counter-intuitive (based on the test diffs) opportunity to use 'psubus'.
This appears to be the better perf option for both Haswell and Jaguar based on llvm-mca.
We already do this transform for the SETULT predicate, so this makes the code more
symmetrical too. If we have pminub/pminuw, we prefer those, so this should not affect
anything but pre-SSE4.1 subtargets.
$ cat before.s
movdqa -16(%rip), %xmm2 ## xmm2 = [32768,32768,32768,32768,32768,32768,32768,32768]
pxor %xmm0, %xmm2
pcmpgtw -32(%rip), %xmm2 ## xmm2 = [255,255,255,255,255,255,255,255]
pand %xmm2, %xmm0
pandn %xmm1, %xmm2
por %xmm2, %xmm0
$ cat after.s
movdqa -16(%rip), %xmm2 ## xmm2 = [256,256,256,256,256,256,256,256]
psubusw %xmm0, %xmm2
pxor %xmm3, %xmm3
pcmpeqw %xmm2, %xmm3
pand %xmm3, %xmm0
pandn %xmm1, %xmm3
por %xmm3, %xmm0
$ llvm-mca before.s -mcpu=haswell
Iterations: 100
Instructions: 600
Total Cycles: 909
Total uOps: 700
Dispatch Width: 4
uOps Per Cycle: 0.77
IPC: 0.66
Block RThroughput: 1.8
$ llvm-mca after.s -mcpu=haswell
Iterations: 100
Instructions: 700
Total Cycles: 409
Total uOps: 700
Dispatch Width: 4
uOps Per Cycle: 1.71
IPC: 1.71
Block RThroughput: 1.8
Differential Revision: https://reviews.llvm.org/D60838
llvm-svn: 358999
This was supposed to be NFC, but the change in SDLoc
definitions causes instruction scheduling changes.
There's nothing x86-specific in this code, and it can
likely be used from DAGCombiner's simplifyVBinOp().
llvm-svn: 358930
Summary:
If you pass two 1024 bit vectors in IR with AVX2 on Windows 64. Both vectors will be split in four 256 bit pieces. The four pieces of the first argument will be passed indirectly using 4 gprs. The second argument will get passed via pointers in memory.
The PartOffsets stored for the second argument are all in terms of its original 1024 bit size. So the PartOffsets for each piece are 32 bytes apart. So if we consider it for copy elision we'll only load an 8 byte pointer, but we'll move the address 32 bytes. The stack object size we create for the first part is probably wrong too.
This issue was encountered by ISPC. I'm working on getting a reduce test case, but wanted to go ahead and get feedback on the fix.
Reviewers: rnk
Reviewed By: rnk
Subscribers: dbabokin, llvm-commits, hiraditya
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D60801
llvm-svn: 358817
combineVectorTruncationWithPACKUS is currently splitting the upper bit bit masking into 128-bit subregs and then concatenating them back together.
This was originally done to avoid regressions that caused existing subregs to be concatenated to the larger type just for the AND masking before being extracted again. This was fixed by @spatel (notably rL303997 and rL347356).
This also lets SimplifyDemandedBits do some further improvements before it hits the recursive depth limit.
My only annoyance with this is that we were broadcasting some xmm masks but we seem to have lost them by moving to ymm - but that's a known issue as the logic in lowerBuildVectorAsBroadcast isn't great.
Differential Revision: https://reviews.llvm.org/D60375#inline-539623
llvm-svn: 358692
This replaces the MOVMSK combine introduced at D52121/rL342326
(movmsk (setne (and X, (1 << C)), 0)) -> (movmsk (X << C))
with the more general icmp lowering so it can pick up more cases through bitcasts - notably vXi8 cases which use vXi16 shifts+masks, this patch can remove the mask and use pcmpgtb(0,x) for the sra.
Differential Revision: https://reviews.llvm.org/D60625
llvm-svn: 358651
As discussed on PR41359, this patch renames the pair of shift-mask target feature functions to make their purposes more obvious.
shouldFoldShiftPairToMask -> shouldFoldConstantShiftPairToMask
preferShiftsToClearExtremeBits -> shouldFoldMaskToVariableShiftPair
llvm-svn: 358526
Improves codegen demonstrated by D60512 - instructions represented by X86ISD::PERMV/PERMV3 can never memory fold the operand used for their index register.
This patch updates the 'isUseOfShuffle' helper into the more capable 'isFoldableUseOfShuffle' that recognises that the op is used for a X86ISD::PERMV/PERMV3 index mask and can't be folded - allowing us to use broadcast/subvector-broadcast ops to reduce the size of the mask constant pool data.
Differential Revision: https://reviews.llvm.org/D60562
llvm-svn: 358516
Currently combineHorizontalPredicateResult only handles anyof/allof reduction patterns of legal types, which can be tricky to match as type legalization of bools can introduce bitcasts/truncs/extensions.
This patch extends combineHorizontalPredicateResult to recognise vXi1 bool reductions as well and uses the existing combineBitcastvxi1 helper to create the MOVMSK necessary to then compare the signmask result.
This ensures the accuracy of the reduction costs added in D60403 which assume the MOVMSK generation.
Differential Revision: https://reviews.llvm.org/D60610
llvm-svn: 358286
If the vector setcc has been legalized then we will need to convert a vector boolean of 0 or -1 to a scalar boolean of 0 or 1.
The added test case previously crashed in 32-bit mode by creating a setcc with an i64 condition that type legalization couldn't expand.
llvm-svn: 358218
This patch adds patterns for turning bitcasted atomic load/store into movss/sd.
It also removes the pseudo instructions for atomic RMW fadd. Instead just adding isel patterns for folding an atomic load into addss/sd. And relying on the new movss/sd store pattern to handle the write part.
This also makes the fadd patterns use VEX and EVEX instructions when AVX or AVX512F are enabled.
Differential Revision: https://reviews.llvm.org/D60394
llvm-svn: 358215
With correct test checks this time.
If we have X87, but not SSE2 we can atomicaly load an i64 value into the significand of an 80-bit extended precision x87 register using fild. We can then use a fist instruction to convert it back to an i64 integ
This matches what gcc and icc do for this case and removes an existing FIXME.
llvm-svn: 358214
If we have X87, but not SSE2 we can atomicaly load an i64 value into the significand of an 80-bit extended precision x87 register using fild. We can then use a fist instruction to convert it back to an i64 integer and store it to a stack temporary. From there we can do two 32-bit loads to get the value into integer registers without worrying about atomicness.
This matches what gcc and icc do for this case and removes an existing FIXME.
Differential Revision: https://reviews.llvm.org/D60156
llvm-svn: 358211
Certain optimisations from ConstantHoisting and CGP rely on Selection DAG not
seeing through to the constant in other blocks. Revert this patch while we come
up with a better way to handle that.
I will try to follow this up with some better tests.
llvm-svn: 358113
Returning SDValue() makes the caller think custom lowering was unsuccessful and then it will fall back to trying to expand the original node. This expanded code will end up with no users and end up being pruned later. But it was useless unnecessary work to create it.
Instead return a MERGE_VALUES with all the results so the caller knows something changed. The caller can handle the replacements.
For one of the cases I had to use UNDEF has a dummy value for a result we know is unused. This should get pruned later.
llvm-svn: 357935
I was looking at a potential DAGCombiner fix for 1 of the regressions in D60278, and it caused severe regression test pain because x86 TLI lies about the desirability of 8-bit shift ops.
We've hinted at making all 8-bit ops undesirable for the reason in the code comment:
// TODO: Almost no 8-bit ops are desirable because they have no actual
// size/speed advantages vs. 32-bit ops, but they do have a major
// potential disadvantage by causing partial register stalls.
...but that leads to massive diffs and exposes all kinds of optimization holes itself.
Differential Revision: https://reviews.llvm.org/D60286
llvm-svn: 357912
Previously LowerOperationWrapper took the number of results from the original
node and counted that many results from the new node. This was intended to drop
chain operands from FP_TO_SINT lowering that uses X87 with memory operations to
stack temporaries. The final load had an extra chain output that needs to be
ignored.
Unfortunately, it didn't work with scatter which has 2 result operands, the
mask output which is discarded and a chain output. The chain output is the one
that is needed but it comes second and it would be dropped by the previous
logic here. To workaround this we were doing a ReplaceAllUses in the lowering
code so that the generic legalization code wouldn't see any uses to replace
since it had been given the wrong result/type.
After this change we take the LowerOperation result directly if the original
node has one result. This allows us to directly return the chain from scatter
or the load data from the FP_TO_SINT case. When the original node has multiple
results we'll ensure the returned node has the same number and copy them over.
For cases where the original node has multiple results and the new code for some
reason has even more results, MERGE_VALUES can be used to pass only the needed
results.
llvm-svn: 357887
In the case where we only want the sign bit (e.g. when using PACKSS truncation of comparison results for MOVMSK) then we can just demand the sign bit of the source operands.
This makes use of the fact that PACKSS saturates out of range values to the min/max int values - so the sign bit is always preserved.
Differential Revision: https://reviews.llvm.org/D60333
llvm-svn: 357859
Summary:
This avoids needing an isel pattern for each condition code. And it removes translation switches for converting between Jcc instructions and condition codes.
Now the printer, encoder and disassembler take care of converting the immediate. We use InstAliases to handle the assembly matching. But we print using the asm string in the instruction definition. The instruction itself is marked IsCodeGenOnly=1 to hide it from the assembly parser.
Reviewers: spatel, lebedev.ri, courbet, gchatelet, RKSimon
Reviewed By: RKSimon
Subscribers: MatzeB, qcolombet, eraman, hiraditya, arphaman, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D60228
llvm-svn: 357802
Summary:
Teach SelectionDAG how to compute known bits of ISD::CopyFromReg if
the virtual reg used has one def only.
This can be particularly useful when calling isBaseWithConstantOffset()
with the ISD::CopyFromReg argument, as more optimizations may get enabled
in the result.
Also add a missing truncation on X86, found by testing of this patch.
Change-Id: Id1c9fceec862d118c54a5b53adf72ada5d6daefa
Reviewers: bogner, craig.topper, RKSimon
Reviewed By: RKSimon
Subscribers: lebedev.ri, nemanjai, jvesely, nhaehnle, javed.absar, jsji, jdoerfert, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D59535
llvm-svn: 357745
We already promote SRL and SHL to i32.
This will introduce sign extends sometimes which might be harder to deal with than the zero we use for promoting SRL. I ran this through some of our internal benchmark lists and didn't see any major regressions.
I think there might be some DAG combine improvement opportunities in the test changes here.
Differential Revision: https://reviews.llvm.org/D60278
llvm-svn: 357743
It unnecessarily breaks previously-working code which used varargs,
but didn't pass any float/double arguments (such as EDK2).
Also revert the fixup on top of that:
Revert [X86] Fix a test from r357317
This reverts r357317 (git commit d413f41de6)
This reverts r357380 (git commit 7af32444b9)
llvm-svn: 357718
These inserters inserted some instructions to zero some registers and copied from virtual registers to physical registers.
This change instead inserts the zeros directly into the DAG at lowering time using new ISD opcodes
that take the extra zeroes as inputs. The zeros will then go through isel on their own to select
the MOV32r0 pseudo. Then we just need to mention the physical registers directly
in the isel patterns and the isel table and InstrEmitter will take care of inserting the necessary
copies to/from physical registers.
llvm-svn: 357659
This custom inserter existed so we could do a weird thing where we pretended that the instructions support
a full address mode instead of taking a pointer in EAX/RAX. I think was largely so we could be pointer
size agnostic in the isel pattern.
To make this work we would then put the address into an LEA into EAX/RAX in front of the instruction after
isel. But the LEA is overkill when we just have a base pointer. So we end up using the LEA as a slower MOV
instruction.
With this change we now just do custom selection during isel instead and just assign the incoming address
of the intrinsic into EAX/RAX based on its size. After the intrinsic is selected, we can let isel take
care of selecting an LEA or other operation to do any address computation needed in this basic block.
I've also split the instruction into a 32-bit mode version and a 64-bit mode version so the implicit
use is properly sized based on the pointer. Without this we get comments in the assembly output about
killing eax and defing rax or vice versa depending on whether we define the instruction to use EAX/RAX.
llvm-svn: 357652
This pattern would show up as a regression if we more
aggressively convert vector FP ops to scalar ops.
There's still a missed optimization for the v4f64 legal
case (AVX) because we create that h-op with an undef operand.
We should probably just duplicate the operands for that
pattern to avoid trouble.
llvm-svn: 357642
One motivation for making this change is that the lack of using movmsk is likely
a main source of perf difference between clang and gcc on the C-Ray benchmark as
shown here:
https://www.phoronix.com/scan.php?page=article&item=gcc-clang-2019&num=5
...but this change alone isn't enough to solve that problem.
The 'all-of' examples show what is likely the worst case trade-off: we end up with
an extra instruction (or 2 if we count the 'xor' register clearing). The 'any-of'
examples look clearly better using movmsk because we've traded 2 vector instructions
for 2 scalar instructions, and movmsk may have better timing than the generic 'movq'.
If we examine the llvm-mca output for these cases, it appears that even though the
'all-of' movmsk variant looks worse on paper, it would perform better on both
Haswell and Jaguar.
$ llvm-mca -mcpu=haswell no_movmsk.s -timeline
Iterations: 100
Instructions: 400
Total Cycles: 504
Total uOps: 400
Dispatch Width: 4
uOps Per Cycle: 0.79
IPC: 0.79
Block RThroughput: 1.0
$ llvm-mca -mcpu=haswell movmsk.s -timeline
Iterations: 100
Instructions: 600
Total Cycles: 358
Total uOps: 600
Dispatch Width: 4
uOps Per Cycle: 1.68
IPC: 1.68
Block RThroughput: 1.5
$ llvm-mca -mcpu=btver2 no_movmsk.s -timeline
Iterations: 100
Instructions: 400
Total Cycles: 407
Total uOps: 400
Dispatch Width: 2
uOps Per Cycle: 0.98
IPC: 0.98
Block RThroughput: 2.0
$ llvm-mca -mcpu=btver2 movmsk.s -timeline
Iterations: 100
Instructions: 600
Total Cycles: 311
Total uOps: 600
Dispatch Width: 2
uOps Per Cycle: 1.93
IPC: 1.93
Block RThroughput: 3.0
Finally, there may be CPUs where movmsk is horribly slow (old AMD small cores?), but if
that's true, then we're also almost certainly making the wrong transform already for
reductions with >2 elements, so that should be fixed independently.
Differential Revision: https://reviews.llvm.org/D59997
llvm-svn: 357367
Fixes PR41316 where the expanded PAVG intrinsic had had one of its ADDs turned into an OR due to its operands having no conflicting bits.
llvm-svn: 357351
We need XMM registers to handle varargs with the Win64 ABI. Before we would
silently generate bad code resulting in an assertion failure elsewhere in the
backend.
llvm-svn: 357317
This is probably the least important of our movmsk problems, but I'm starting
at the bottom to reduce distractions.
We were creating a select_cc which bypasses the select and bitmask codegen
optimizations that we have now. If we produce a compare+negate instead, we
allow things like neg/sbb carry bit hacks, and in all cases we avoid a cmov.
There's no partial register update danger in these sequences because we always
produce the zero-register xor ahead of the 'set' if needed.
There seems to be a missing fold for sext of a bool bit here:
negl %ecx
movslq %ecx, %rax
...but that's an independent transform.
Differential Revision: https://reviews.llvm.org/D59818
llvm-svn: 357172
If we know the 2 halves of an oversized zext-in-reg are the same,
don't create those halves independently.
I tried several different approaches to fold this, but it's difficult
to get right during legalization. In the default path, we are creating
a generic shuffle that looks like an unpack high, but it can get
transformed into a different mask (a blend), so it's not
straightforward to match that. If we try to fold after it actually
becomes an X86ISD::UNPCKH node, we can't be sure what the operand node
is - it might be a generic shuffle, or it could be some x86-specific op.
From the test output, we should be doing something like this for SSE4.1
as well, but I'd rather leave that as a follow-up since it involves
changing lowering actions.
Differential Revision: https://reviews.llvm.org/D59777
llvm-svn: 357129
This is not exactly NFC because it should make further combines
of MOVMSK easier to match, but there should be no outward differences
because we have isel patterns in place specifically to allow this. See:
// Also support integer VTs to avoid a int->fp bitcast in the DAG.
llvm-svn: 357128
Enable SSE41 ZERO_EXTEND_VECTOR_INREG shuffle combines - for the PMOVZX(PSHUFD(V)) -> UNPCKH(V,0) pattern we reduce the shuffles (port5-bottleneck on Intel) at the expense of creating a zero (pxor v,v) and an extra register move - which is a good trade off as these are pretty cheap and in most cases it doesn't increase register pressure.
This also exposed a missed opportunity to use combine to ZERO_EXTEND_VECTOR_INREG with folded loads - even if we're in the float domain.
........
Causes PR41249
llvm-svn: 357057
Enable SSE41 ZERO_EXTEND_VECTOR_INREG shuffle combines - for the PMOVZX(PSHUFD(V)) -> UNPCKH(V,0) pattern we reduce the shuffles (port5-bottleneck on Intel) at the expense of creating a zero (pxor v,v) and an extra register move - which is a good trade off as these are pretty cheap and in most cases it doesn't increase register pressure.
This also exposed a missed opportunity to use combine to ZERO_EXTEND_VECTOR_INREG with folded loads - even if we're in the float domain.
llvm-svn: 356864
Just enable this for AVX for now as SSE41 introduces extra register moves for the PMOVZX(PSHUFD(V)) -> UNPCKH(V,0) pattern (but otherwise helps reduce port5 usage on Intel targets).
Only AVX support is required for PR40685 as the issue is due to 8i8->8i32 zext shuffle leftovers.
llvm-svn: 356858
This is yet another step towards solving PR14613:
https://bugs.llvm.org/show_bug.cgi?id=14613
uaddsat X, Y --> (X >u (X + Y)) ? -1 : X + Y
usubsat X, Y --> (X >u Y) ? X - Y : 0
We can't count on a sane vector ISA, so override the default (umin/umax)
expansion of unsigned add/sub saturate in cases where we do not have umin/umax.
Differential Revision: https://reviews.llvm.org/D59006
llvm-svn: 356855
On 32-bit targets without popcnt, we currently expand 64-bit popcnt to sequences of arithmetic and logic ops for each 32-bit half and then add the 32 bit halves together. If we have xmm registers we can use use those to implement the operation instead. This results in less instructions then doing two separate 32-bit popcnt sequences.
This mitigates some of PR41151 for the i64 on i686 case when we have SSE2.
Differential Revision: https://reviews.llvm.org/D59662
llvm-svn: 356808
We used a lock cmpxchg8b to do i64 atomic loads. But if we have SSE2 we can do better and use a plain movq to do the load instead.
I tried to just use an f64 atomic load and add isel patterns to MOVSD(which the domain fixing pass can turn to MOVQ), but the atomic_load SDNode in TargetSelectionDAG.td requires the type to be integer.
So I've emitted VZEXT_LOAD instead which should be selected by isel to a MOVQ. Hopefully we don't need a specific atomic flavor of this. I kept the memory operand from the original AtomicSDNode. I wasn't sure if I might need to set the MOVolatile flag?
I've left some FIXMEs for improvements we can do without SSE2.
Differential Revision: https://reviews.llvm.org/D59679
llvm-svn: 356807
CMPXCHG8B was introduced on i586/pentium generation.
If its not enabled, limit the atomic width to 32 bits so the AtomicExpandPass will expand to lib calls. Unclear if we should be using a different limit for other configs. The default is 1024 and experimentation shows that using an i256 atomic will cause a crash in SelectionDAG.
Differential Revision: https://reviews.llvm.org/D59576
llvm-svn: 356631
This patch enables the use of lowerShuffleAsBitMask for 512-bit blends before
falling back to move immedate, GPR to k-register, and masked op.
I had to make some changes to support v8i64 when i64 is not a legal type. And to
support floating point types.
This trades a load for the move immediate and GPR move which is higher latency.
But its probably better for register pressure not having to hop through other
register classes. The load+and should play better with LICM and
rematerialization I think.
Differential Revision: https://reviews.llvm.org/D59479
llvm-svn: 356618
This patch removes the following dag node opcodes from namespace X86ISD:
RDTSC_DAG,
RDTSCP_DAG,
RDPMC_DAG
The logic that expands RDTSC/RDPMC/XGETBV intrinsics is basically the same. The
only differences are:
RDTSC/RDTSCP don't implicitly read ECX.
RDTSCP also implicitly writes ECX.
I moved the common expansion logic into a helper function with the goal to get
rid of code repetition. That helper is now used for the expansion of
RDTSC/RDTSCP/RDPMC/XGETBV intrinsics.
No functional change intended.
Differential Revision: https://reviews.llvm.org/D59547
llvm-svn: 356546
These changes are related to PR37743 and include:
SelectionDAGBuilder::visitSelect handles the unary SelectPatternFlavor::SPF_ABS case to build ABS node.
Delete the redundant recognizer of the integer ABS pattern from the DAGCombiner.
Add promoting the integer ABS node in the LegalizeIntegerType.
Expand-based legalization of integer result for the ABS nodes.
Expand-based legalization of ABS vector operations.
Add some integer abs testcases for different typesizes for Thumb arch
Add the custom ABS expanding and change the SAD pattern recognizer for X86 arch: The i64 result of the ABS is expanded to:
tmp = (SRA, Hi, 31)
Lo = (UADDO tmp, Lo)
Hi = (XOR tmp, (ADDCARRY tmp, hi, Lo:1))
Lo = (XOR tmp, Lo)
The "detectZextAbsDiff" function is changed for the recognition of pattern with the ABS node. Given a ABS node, detect the following pattern:
(ABS (SUB (ZERO_EXTEND a), (ZERO_EXTEND b))).
Change integer abs testcases for codegen with the ABS node support for AArch64.
Indicate that the ABS is legal for the i64 type when the NEON is supported.
Change the integer abs testcases to show changing of codegen.
Add combine and legalization of ABS nodes for Thumb arch.
Extend 'matchSelectPattern' to recognize the ABS patterns with ICMP_SGE condition.
For discussion, see https://bugs.llvm.org/show_bug.cgi?id=37743
Patch by: @ikulagin (Ivan Kulagin)
Differential Revision: https://reviews.llvm.org/D49837
llvm-svn: 356468
This allows better code size for aarch64 floating point materialization
in a future patch.
Reviewers: evandro
Differential Revision: https://reviews.llvm.org/D58690
llvm-svn: 356389
Summary:
As noted by @andreadb in https://reviews.llvm.org/D59035#inline-525780
If we have `sext (trunc (cmov C0, C1) to i8)`,
we can instead do `cmov (sext (trunc C0 to i8)), (sext (trunc C1 to i8))`
Reviewers: craig.topper, andreadb, RKSimon
Reviewed By: craig.topper
Subscribers: llvm-commits, andreadb
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D59412
llvm-svn: 356301
The asm parser generates the immediate without the SAE bit. So for consistency we should generate the MCInst the same way from CodeGen.
Since they are now both the same, remove the masking from the printer and replace with an llvm_unreachable.
Use a target constant since we're rebuilding the node anyway. Then we don't have to have isel convert it. Saves about 500 bytes from the isel table.
llvm-svn: 356294
Reduce the size of an any-extended i64 scalar_to_vector source to i32 - the any_extend nodes are often introduced by SimplifyDemandedBits.
llvm-svn: 356292
This is an immediate fix for:
https://bugs.llvm.org/show_bug.cgi?id=41066
...but as noted there and the code comments, we should do better
by stubbing this out sooner.
llvm-svn: 356158
The feature flag alone can't be trusted since it can be passed via -mattr. Need to ensure 64-bit mode as well.
We had a 64 bit mode check on the instruction to make the assembler work correctly. But we weren't guarding any of our lowering code or the hooks for the AtomicExpandPass.
I've added 32-bit command lines to atomic128.ll with and without cx16. The tests there would all previously fail if -mattr=cx16 was passed to them. I had to move one test case for f128 to a new file as it seems to have a different 32-bit mode or possibly sse issue.
Differential Revision: https://reviews.llvm.org/D59308
llvm-svn: 356078
Attempt to combine CONCAT_VECTORS nodes, which we only really have pre-legalization.
This encourages a lot of X86ISD::SUBV_BROADCAST generation, so I've added SimplifyDemandedVectorEltsForTargetNode handling for this at the same time.
The X86ISD::VTRUNC regression in shuffle-vs-trunc-256-widen.ll will be handled in a future commit.
llvm-svn: 356064
A fuzzer found the crasher:
https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=13700
The bug was introduced recently here:
rL355741
This is the quick fix. If we need to do this transform
later, then we'd have to extend/truncate the vector setcc
element type to the scalar setcc type (i8).
llvm-svn: 356053
AVX1 broadcasts were failing as we were adding bitcasts that caused MayFoldLoad's hasOneUse to return false.
This patch stops introducing bitcasts so early and also replaces the broadcast index scaling through bitcasts (which can't succeed in some cases) to instead just keep track of the bitoffset which can be converted back to the broadcast index later on.
Differential Revision: https://reviews.llvm.org/D58888
llvm-svn: 356043
Instead I plan to have dedicated nodes for FROUND_CURRENT and FROUND_NO_EXC.
This patch starts with FADDS/FSUBS/FMULS/FDIVS/FMAXS/FMINS/FSQRTS.
llvm-svn: 355799
Many of our tests were not using valid rounding mode immediates. Clang verifies this in the frontend when it creates the intrinsics from builtins, but the backend would still lower invalid immediates.
With this change we will now leave them as intrinsics if the immediate is invalid. This will cause an isel selection failure.
llvm-svn: 355789
An extension of D58282 noted in PR39665:
https://bugs.llvm.org/show_bug.cgi?id=39665
This doesn't answer the request to use movmsk, but that's an
independent problem. We need this and probably still need
scalarization of FP selects because we can't do that as a
target-independent transform (although it seems likely that
targets besides x86 should have this transform).
llvm-svn: 355741
Move the x86 combine from D58974 into the DAGCombine VSELECT code and update the SELECT version to use the isBooleanFlip helper as well.
Requested by @spatel on D59006
llvm-svn: 355533
As noticed on D58965
DAGCombiner::visitSELECT has something similar, so we should be able to move this to DAGCombiner and support VSELECT as well at some point.
Differential Revision: https://reviews.llvm.org/D58974
llvm-svn: 355494
X86TargetLowering::EmitLoweredSelect presently detects sequences of CMOV pseudo
instructions without accounting for debug intrinsics. This leads to different
codegen with and without option -g, if a DBG_VALUE instruction lands in the
middle of several lowered selects.
Work around this by skipping over debug instructions when looking for CMOV
sequences, and sinking those debug insts into the EmitLoweredSelect sunk block.
This might slightly shift where variables appear in the instruction sequence,
but won't re-order assignments.
Differential Revision: https://reviews.llvm.org/D58672
llvm-svn: 355307
Summary:
This extends the variety of pattern that can generate a SHLD instead of using two shifts.
This fixes a regression that would be introduced by D57367 or D33587
Reviewers: RKSimon, craig.topper
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D57389
llvm-svn: 355260
This is another step towards ensuring that we produce the optimal code for reductions,
but there are other potential benefits as seen in the tests diffs:
1. Memory loads may get scalarized resulting in more efficient code.
2. Memory stores may get scalarized resulting in more efficient code.
3. Complex ops like fdiv/sqrt get scalarized which may be faster instructions depending on uarch.
4. Even simple ops like addss/subss/mulss/roundss may result in faster operation/less frequency throttling when scalarized depending on uarch.
The TODO comment suggests 1 or more follow-ups for opcodes that can currently result in regressions.
Differential Revision: https://reviews.llvm.org/D58282
llvm-svn: 355130
We don't have any combines that can look through a bitcast to truncate a build vector of constants. So the truncate will stick around and give us something like this pattern (binop (trunc X), (trunc (bitcast (build_vector)))) which has two truncates in it. Which will be reversed by hoistLogicOpWithSameOpcodeHands in the generic DAG combiner. Thus causing an infinite loop.
Even if we had a combine for (truncate (bitcast (build_vector))), I think it would need to be implemented in getNode otherwise DAG combiner visit ordering would probably still visit the binop first and reverse it. Or combineTruncatedArithmetic would need to do its own constant folding.
Differential Revision: https://reviews.llvm.org/D58705
llvm-svn: 355116
Summary:
The description of KnownBits::zext() and
KnownBits::zextOrTrunc() has confusingly been telling
that the operation is equivalent to zero extending the
value we're tracking. That has not been true, instead
the user has been forced to explicitly set the extended
bits as known zero afterwards.
This patch adds a second argument to KnownBits::zext()
and KnownBits::zextOrTrunc() to control if the extended
bits should be considered as known zero or as unknown.
Reviewers: craig.topper, RKSimon
Reviewed By: RKSimon
Subscribers: javed.absar, hiraditya, jdoerfert, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D58650
llvm-svn: 355099
A lot of the INSERT_SUBVECTOR combines can be more generally handled as if they have come from a CONCAT_VECTORS node.
I've been investigating adding a CONCAT_VECTORS combine to X86, but this is a much easier first step that avoids the issue of handling a number of pre-legalization issues that I've encountered.
Differential Revision: https://reviews.llvm.org/D58583
llvm-svn: 355015
Summary:
Use a custom calling convention handler for interrupts instead of fixing
up the locations in LowerMemArgument. This way, the offsets are correct
when constructed and we don't need to account for them in as many
places.
Depends on D56883
Replaces D56275
Reviewers: craig.topper, phil-opp
Subscribers: hiraditya, llvm-commits
Differential Revision: https://reviews.llvm.org/D56944
llvm-svn: 354837
Avoid ADD/SUB instruction duplication by reusing the X86ISD::ADD/SUB results.
Includes ADD commutation - I tried to include NEG+SUB SUB commutation as well but this causes regressions as we don't have good combine coverage to simplify X86ISD::SUB.
Differential Revision: https://reviews.llvm.org/D58597
llvm-svn: 354771
Its proving tricky to combine shuffles across multiple vector sizes, so for now I'm adding this more specific combine - the pattern is common enough to be worth it as a first step.
llvm-svn: 354757
Summary:
When promoting the over flow vector for these ops we should use the target's desired setcc result type. This way a v8i32 result type will use a v8i32 overflow vector instead of a v8i16 overflow vector. A v8i16 overflow vector will cause LegalizeDAG/LegalizeVectorOps to have to use v8i32 and truncate to v8i16 in its expansion. By doing this in type legalization instead, we get the truncate into the DAG earlier and give DAG combine more of a chance to optimize it.
We also have to fix unrolling to use the scalar setcc result type for the scalarized operation, and convert it to the required vector element type after the scalar operation. We have to observe the vector boolean contents when doing this conversion. The previous code was just taking the scalar result and putting it in the vector. But for X86 and AArch64 that would have only put a the boolean value in bit 0 of the element and left all other bits in the element 0. We need to ensure all bits in the element are the same. I'm using a select with constants here because that's what setcc unrolling in LegalizeVectorOps used.
Reviewers: spatel, RKSimon, nikic
Reviewed By: nikic
Subscribers: javed.absar, kristof.beyls, dmgreen, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D58567
llvm-svn: 354753
r354648 was a follow up to fix a regression "[X86] Add a DAG combine for (aext_vector_inreg (aext_vector_inreg X)) -> (aext_vector_inreg X) to fix a regression from my previous commit."
These were reverted in r354713 as their context depended on other patches that were reverted for a bug.
llvm-svn: 354734
Even on AVX1 we can pretty cheaply (VPERM2F128+VSHUFPD) permute a single v4f64/v4i64 input (on AVX2 its just a single VPERMPD), followed by a BLENDPD.
llvm-svn: 354729
r354363 caused https://crbug.com/934963#c1, which has a plain C reduced
test case.
I also had to revert some dependent changes:
- r354648
- r354647
- r354640
- r354511
llvm-svn: 354713
If the the input type will be promoted to 128 bits its better to put a sign_extend_inreg/and in the 128 bit register before the split occurs. Otherwise we end up doing it on each half in the wider register.
Some of the overflow arithmetic tests are regressions, but I think we can make some improvement using getSetccResultType in DAG combine and/or type legalization.
llvm-svn: 354709
As discussed in:
D56864
D58197
Always use the narrow (128-bit) instruction when possible.
We already had the signed int version of this transform.
llvm-svn: 354675
Only the 1st fold is attempted pre-legalization, but it requires
legal (simple) types too, so we don't need an EVT in any of the code.
llvm-svn: 354674
This is a follow-up to D56864.
If we're extracting from a non-zero index before casting to FP,
then shuffle the vector and optionally narrow the vector before doing the cast:
cast (extelt V, C) --> extelt (cast (extract_subv (shuffle V, [C...]))), 0
This might be enough to close PR39974:
https://bugs.llvm.org/show_bug.cgi?id=39974
Differential Revision: https://reviews.llvm.org/D58197
llvm-svn: 354619
We currently bail if the target shuffle decodes to more than 2 input vectors, this change alters the input index to work for any number of inputs for when we drop that requirement.
llvm-svn: 354575
Second part of https://bugs.llvm.org/show_bug.cgi?id=40442.
This adds an extra UnrollVectorOverflowOp() method to SDAG, because
the general UnrollOverflowOp() method can't deal with multiple results.
Additionally we need to expand UMULO/SMULO during vector op
legalization, as it may result in unrolling, which may need additional
type legalization.
Differential Revision: https://reviews.llvm.org/D57997
llvm-svn: 354513
We currently bail if the target shuffle decodes to more than 2 input vectors, this is some initial cleanup that still has the limit but generalizes the opindices to an array that will be necessary when we drop the limit.
llvm-svn: 354489
D42042 introduced the ability for the ExecutionDomainFixPass to more easily change between BLENDPD/BLENDPS/PBLENDW as the domains required.
With this ability, we can avoid most bitcasts/scaling in the DAG that was occurring with X86ISD::BLENDI lowering/combining, blend with the vXi32/vXi64 vectors directly and use isel patterns to lower to the float vector equivalent vectors.
This helps the shuffle combining and SimplifyDemandedVectorElts be more aggressive as we lose track of fewer UNDEF elements than when we go up/down through bitcasts.
I've introduced a basic blend(bitcast(x),bitcast(y)) -> bitcast(blend(x,y)) fold, there are more generalizations I can do there (e.g. widening/scaling and handling the tricky v16i16 repeated mask case).
The vector-reduce-smin/smax regressions will be fixed in a future improvement to SimplifyDemandedBits to peek through bitcasts and support X86ISD::BLENDV.
Reapplied after reversion at rL353699 - AVX2 isel fix was applied at rL354358, additional test at rL354360/rL354361
Differential Revision: https://reviews.llvm.org/D57888
llvm-svn: 354363
This patch adds scalar/subvector BROADCAST handling to EltsFromConsecutiveLoads.
It mainly shows codegen changes to 32-bit code which failed to handle i64 loads, although 64-bit code is also using this new path to more efficiently combine to a broadcast load.
Differential Revision: https://reviews.llvm.org/D58053
llvm-svn: 354340
The motivating x86 cases for forming the intrinsic are shown in PR31754 and PR40487:
https://bugs.llvm.org/show_bug.cgi?id=31754https://bugs.llvm.org/show_bug.cgi?id=40487
..and those are shown in the IR test file and x86 codegen file.
Matching the usubo pattern is harder than uaddo because we have 2 independent values rather than a def-use.
This adds a TLI hook that should preserve the existing behavior for uaddo formation, but disables usubo
formation by default. Only x86 overrides that setting for now although other targets will likely benefit
by forming usbuo too.
Differential Revision: https://reviews.llvm.org/D57789
llvm-svn: 354298
Similar to D57867 - this is a small patch with lots of test diffs.
With half-vector-width narrowing potential, using an extract + 128-bit vshufps
is a win because it replaces a 256-bit shuffle with a 128-bit shufle.
This seems like it should be a win even for targets with 'fast-variable-shuffle',
but we are intentionally deferring that to an independent change to make sure
that is true.
Differential Revision: https://reviews.llvm.org/D58181
llvm-svn: 354279
No need for a separate stack slot. The lifetimes don't overlap.
Also fix the MachinePointerInfo for the final load after the integer conversion to indicate it came from the stack slot.
llvm-svn: 354234
No need to manually split everything. We can let the type legalizer work for us.
The test change seems to be caused by some DAG ordering issue that was previously circumventing a one use check in LowerSELECT where FP selects are turned into blends if the setcc has one use. But it was running after an integer select and the same setcc had been legalized to cmov and X86SISD::CMP. This dropped the use count of the setcc, but wasn't what was intended.
llvm-svn: 354197
When we need to do an fp->int conversion using x87 instructions, we need to temporarily change the rounding mode to 0b11 and perform a store. To do this we save the old value of the fpcw to the stack, then set the fpcw to 0xc7f, do the store, then restore fpcw. But the 0xc7f value forces the exception mask bits 1. While this is what they would be in the default FP environment, as we move to support changing the FP environments, we shouldn't make this assumption.
This patch changes the code to explicitly OR 0xc00 with the old value so that only the rounding mode is changed. Unfortunately, this requires two stack temporaries instead of one. One to hold the old value and one to hold the new value. Without two stack temporaries we would need an additional GPR. We already need one to do the OR operation in. This is similar to what gcc and icc do for this operation. Though they are both better at reusing the stack temporaries when there are multiple truncates in a function(or at least in a basic block)
Differential Revision: https://reviews.llvm.org/D57788
llvm-svn: 354178
These checks aren't needed on the call to FP_TO_INTHelper from the type legalizer for splitting i64. We always want to use X87 FIST/FISTT to memory there.
Moving up the SSE checks will allow this routine to focus on what it cares about and makes its return semantics cleaner.
llvm-svn: 354161
As detailed on PR40730, we are not correctly filling in the lane shuffle mask (D53148/rL344446) - we fill in for the correct src lane but don't add it to the correct mask element, so any reference to the correct element is likely to see an UNDEF mask index.
This allows constant folding to propagate UNDEFs prior to the lane mask being (correctly) lowered to vperm2f128.
This patch fixes the issue by fully populating the lane shuffle mask - this is more than is necessary (if we only filled in the required mask elements we might be able to match other shuffle instructions - broadcasts etc.), but its the most cautious approach as this needs to be cherrypicked into the 8.0.0 release branch.
Differential Revision: https://reviews.llvm.org/D58237
llvm-svn: 354117
When SSE is enabled sint_to_fp with i16 is blindly promoted to i32, but that changes the behavior of f80 conversion.
Move the promotion to i16 to LowerFP_TO_INT so we can limit it based on the floating point type.
llvm-svn: 354003
In 64-bit mode prior to avx512 we use Expand, but with avx512 we need to make f32/f64 conversions Legal so we use Custom and then do our own expansion for f80. But this seems to produce codegen differences relative to avx2. This patch corrects this.
llvm-svn: 353921
-Pull the final stack load creation from the two callers into the helper.
-Return a single SDValue instead of a std::pair.
-Remove the Replace flag which isn't really needed.
llvm-svn: 353920
A more limited version of rL352997 that had to be disabled in rL353198 - allow extension of any 128/256/512 bit vector that at least uses byte sized scalars.
llvm-svn: 353860
We were using DstTy, but that represents the integer type we are converting to which is i64 in this
case. The FLD is part of an intermediate step to get from the SSE registers to the x87 registers.
If the floating point type is f32, the memory operand should reflect a 4 byte access not an 8 byte
access. The store we used to get from SSE to the stack is using the corect size.
While there, consistenly use TheVT in place of Op.getOperand(0).getValueType() throughout the function.
llvm-svn: 353745
256-bit horizontal math ops are an x86 monstrosity (and thankfully have
not been extended to 512-bit AFAIK).
The two 128-bit halves operate on separate halves of the inputs. So if we
don't demand anything in the upper half of the result, we can extract the
low halves of the inputs, do the math, and then insert that result into a
256-bit output.
All of the extract/insert is free (ymm<-->xmm), so we're left with a
narrower (cheaper) version of the original op.
In the affected tests based on:
https://bugs.llvm.org/show_bug.cgi?id=33758https://bugs.llvm.org/show_bug.cgi?id=38971
...we see that the h-op narrowing can result in further narrowing of other
math via existing generic transforms.
I originally drafted this patch as an exact pattern match starting from
extract_vector_elt, but I thought we might see diffs starting from
extract_subvector too, so I changed it to a more general demanded elements
solution. There are no extra existing regression test improvements from
that switch though, so we could go back.
Differential Revision: https://reviews.llvm.org/D57841
llvm-svn: 353641
As discussed on D57389, this is a first step towards moving the SHLD/SHRD matching code to DAGCombiner using FSHL/FSHR instead.
There's a bit of work to do before I can do that, so this just folds to FSHL/FSHR in the existing code (handling the different SHRD/FSHR argument ordering), which fixes the issue we had with i16 shift amounts not being correctly masked.
llvm-svn: 353626
D42042 introduced the ability for the ExecutionDomainFixPass to more easily change between BLENDPD/BLENDPS/PBLENDW as the domains required.
With this ability, we can avoid most bitcasts/scaling in the DAG that was occurring with X86ISD::BLENDI lowering/combining, blend with the vXi32/vXi64 vectors directly and use isel patterns to lower to the float vector equivalent vectors.
This helps the shuffle combining and SimplifyDemandedVectorElts be more aggressive as we lose track of fewer UNDEF elements than when we go up/down through bitcasts.
I've introduced a basic blend(bitcast(x),bitcast(y)) -> bitcast(blend(x,y)) fold, there are more generalizations I can do there (e.g. widening/scaling and handling the tricky v16i16 repeated mask case).
The vector-reduce-smin/smax regressions will be fixed in a future improvement to SimplifyDemandedBits to peek through bitcasts and support X86ISD::BLENDV.
Differential Revision: https://reviews.llvm.org/D57888
llvm-svn: 353610
Move the (add (umax X, C), -C) --> (usubsat X, C) X86 combine into generic DAGCombiner
First of a number of saturated arithmetic folds that can be moved out of X86-specific code for PR40111.
Differential Revision: https://reviews.llvm.org/D57754
llvm-svn: 353457
This is intentionally a small step because it's hard to know exactly
where we might introduce a conflicting transform with the code that
tries to form wider shuffles. But I think this is safe - if we have
a wide shuffle with 2 operands, then we should do better with an
extract + narrow shuffle.
Differential Revision: https://reviews.llvm.org/D57867
llvm-svn: 353427
combineExtractWithShuffle may leave a dangling bitcast which may
prevent further optimization in later passes. Avoid constructing it
unless it is used.
llvm-svn: 353333
The proposal in D56796 may cross the line because we're trying to avoid vectorization
transforms in generic DAG combining. So this is an alternate, later, x86-specific
translation of that patch.
There are several potential follow-ups to enhance this:
1. Allow extraction from non-zero element index.
2. Peek through extends of smaller width integers.
3. Support x86-specific conversion opcodes like X86ISD::CVTSI2P
Differential Revision: https://reviews.llvm.org/D56864
llvm-svn: 353302
rL352997 enabled ZERO_EXTEND from non-shuffle-able value types. I've disabled it for now to fix a regression identified by @asbirlea until I can fix this properly.
llvm-svn: 353198
If we have broadcasts of different vector widths, keep the longest vector width and extract subvectors for the shorter vectors (which should be free).
Differential Revision: https://reviews.llvm.org/D57663
llvm-svn: 353154
Summary:
We don't currently map these constraints to physical register numbers so they don't make it to the MachineIR representation of inline assembly.
This could have problems for proper dependency tracking in the machine schedulers though I don't have a test case that shows that.
Reviewers: rnk
Reviewed By: rnk
Subscribers: eraman, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D57641
llvm-svn: 353141
For PCMPGT(0, X) patterns where we only demand the sign bit (e.g. BLENDV or MOVMSK) then we can use X directly.
Differential Revision: https://reviews.llvm.org/D57667
llvm-svn: 353051
Summary:
Add an additional combine to combineCarryThroughADD to reverse it back to the C flag to avoid regressions.
I believe this catches the cases that D57547 got.
Reviewers: RKSimon, spatel
Reviewed By: spatel
Subscribers: javed.absar, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D57637
llvm-svn: 352984
Push the insert_subvector up through the shuffle operands to help find more cross-lane shuffles.
The is exposes a couple of minor issues that will be fixed shortly:
Missed broadcast folds - we have a mixture of vzext_load lengths that need cleaning up
combine-sdiv.ll - AVX1 SimplifyDemandedVectorElts failure (hits max depth due to a couple of extra bitcasts).
llvm-svn: 352963
We already have the getConstantOperandVal helper which returns a uint64_t, but along comes the fuzzer and inserts a i128 -1 constant or something and the whole thing asserts.......
I've updated a few obvious cases, and tried to make use of the const reference where possible, but there's more to do. A number of existing oss-fuzz tickets should be fixed if we start using APInt and perform value clamping where necessary.
llvm-svn: 352961
This cleans up all LoadInst creation in LLVM to explicitly pass the
value type rather than deriving it from the pointer's element-type.
Differential Revision: https://reviews.llvm.org/D57172
llvm-svn: 352911
This cleans up all CallInst creation in LLVM to explicitly pass a
function type rather than deriving it from the pointer's element-type.
Differential Revision: https://reviews.llvm.org/D57170
llvm-svn: 352909
As suggested on PR40318, this patch uses PSLLDQ/PSRLDQ to lower shuffles to zero out the ends of a vector, leaving a sequential inner section.
For pre-SSSE3 we do this for shuffles with zeros at either end (requiring up to 3 shifts), but once PSHUFB is available I've limited this to shuffles with a single zeroable end (2 shifts).
Differential Revision: https://reviews.llvm.org/D56784
llvm-svn: 352883
Enable peeking through one use bitcasts to the subvector shuffle.
This still depends on the subvector being the same scalar-size but D57514 has already helped with the more tricky patterns
llvm-svn: 352879
Recommit r352791 after tweaking DerivedTypes.h slightly, so that gcc
doesn't choke on it, hopefully.
Original Message:
The FunctionCallee type is effectively a {FunctionType*,Value*} pair,
and is a useful convenience to enable code to continue passing the
result of getOrInsertFunction() through to EmitCall, even once pointer
types lose their pointee-type.
Then:
- update the CallInst/InvokeInst instruction creation functions to
take a Callee,
- modify getOrInsertFunction to return FunctionCallee, and
- update all callers appropriately.
One area of particular note is the change to the sanitizer
code. Previously, they had been casting the result of
`getOrInsertFunction` to a `Function*` via
`checkSanitizerInterfaceFunction`, and storing that. That would report
an error if someone had already inserted a function declaraction with
a mismatching signature.
However, in general, LLVM allows for such mismatches, as
`getOrInsertFunction` will automatically insert a bitcast if
needed. As part of this cleanup, cause the sanitizer code to do the
same. (It will call its functions using the expected signature,
however they may have been declared.)
Finally, in a small number of locations, callers of
`getOrInsertFunction` actually were expecting/requiring that a brand
new function was being created. In such cases, I've switched them to
Function::Create instead.
Differential Revision: https://reviews.llvm.org/D57315
llvm-svn: 352827
This reverts commit f47d6b38c7 (r352791).
Seems to run into compilation failures with GCC (but not clang, where
I tested it). Reverting while I investigate.
llvm-svn: 352800
The FunctionCallee type is effectively a {FunctionType*,Value*} pair,
and is a useful convenience to enable code to continue passing the
result of getOrInsertFunction() through to EmitCall, even once pointer
types lose their pointee-type.
Then:
- update the CallInst/InvokeInst instruction creation functions to
take a Callee,
- modify getOrInsertFunction to return FunctionCallee, and
- update all callers appropriately.
One area of particular note is the change to the sanitizer
code. Previously, they had been casting the result of
`getOrInsertFunction` to a `Function*` via
`checkSanitizerInterfaceFunction`, and storing that. That would report
an error if someone had already inserted a function declaraction with
a mismatching signature.
However, in general, LLVM allows for such mismatches, as
`getOrInsertFunction` will automatically insert a bitcast if
needed. As part of this cleanup, cause the sanitizer code to do the
same. (It will call its functions using the expected signature,
however they may have been declared.)
Finally, in a small number of locations, callers of
`getOrInsertFunction` actually were expecting/requiring that a brand
new function was being created. In such cases, I've switched them to
Function::Create instead.
Differential Revision: https://reviews.llvm.org/D57315
llvm-svn: 352791
Similar to what we already do in DAGCombiner, but this version also handles bitcasts from types with different scalar sizes, which x86 is better at handling.
Differential Revision: https://reviews.llvm.org/D57514
llvm-svn: 352773
Enables 32/64-bit scalar load broadcasts on AVX1 targets
The extractelement-load.ll regression will be fixed shortly in a followup commit.
llvm-svn: 352743
If we're not inserting the broadcast into the lowest subvector then we can avoid the insertion by just performing a larger broadcast.
Avoids a regression when we enable AVX1 broadcasts in shuffle combining
llvm-svn: 352742
I believe this was there to handle avx512bw intrinsics that returned i64 type in 32-bit mode. But all those intrinsics have since been changed to v64i1 results or replaced with generic IR.
llvm-svn: 352698
Without the fix we get the following (with -Werror):
../lib/Target/X86/X86ISelLowering.cpp:14181:58: error: suggest braces around initialization of subobject [-Werror,-Wmissing-braces]
SmallVector<std::array<int, 2>, 2> LaneSrcs(NumLanes, {-1, -1});
^~~~~~
{ }
1 error generated.
llvm-svn: 352455
This did not cause the buildbot failure it was previously reverted for.
Original commit message:
I'm not sure why we were using SEXTLOAD. EXTLOAD seems more appropriate since we don't care about the upper bits.
This patch changes this and then modifies the X86 post legalization combine to emit a extending shuffle instead of a sign_extend_vector_inreg. Could maybe use an any_extend_vector_inre
On AVX512 targets I think we might be able to use a masked vpmovzx and not have to expand this at all.
llvm-svn: 352433
First step towards adding support for 64-bit unary "sublane" handling (a bit like lowerShuffleAsRepeatedMaskAndLanePermute).
This allows us to add lowerV64I8Shuffle handling.
llvm-svn: 352389
This is tricky to make optimal: sometimes we're better off using
a single wider op, but other times it makes more sense to combine
a narrow ops to achieve the same result.
This solves the case from:
https://bugs.llvm.org/show_bug.cgi?id=40434
There's potentially a similar change for vectors with 64-bit elements,
but it needs adjustments similar to rL352333 to avoid creating infinite
loops.
llvm-svn: 352380
This transform was added with rL351346, and we had
an escape for shufps, but we also want one for
unpckps vs. vpermps because vpermps doesn't take
an immediate shuffle index operand.
llvm-svn: 352333
Although this is longer code, this is no-functional-change-intended.
The goal is to untangle the conditions under which we bail out, so
that's easier to adjust.
llvm-svn: 352320
Fix issue noted in D57281 that only tested the one use for the SDValue (the result flag), not the entire SUB.
I've added the getNode() to make it clearer what is intended than just the -> redirection.
llvm-svn: 352291
As discussed on PR24545, we should try to commute X86::COND_A 'icmp ugt' cases to X86::COND_B 'icmp ult' to more optimally bind the carry flag output to a SBB instruction.
Differential Revision: https://reviews.llvm.org/D57281
llvm-svn: 352289
We often generate X86ISD::SBB(X, 0) for carry flag arithmetic.
I had tried to create test cases for the ADC equivalent (which often uses the same pattern) but haven't managed to find anything yet.
Differential Revision: https://reviews.llvm.org/D57169
llvm-svn: 352288
This reduces a bit of duplication between the combining and
lowering places that use it, but the primary motivation is
to make it easier to rearrange the lowering logic and solve
PR40434:
https://bugs.llvm.org/show_bug.cgi?id=40434
llvm-svn: 352280
Summary: We have isel patterns for this, but we're missing some load patterns and all broadcast patterns. A DAG combine seems like a better fit for this.
Reviewers: RKSimon, spatel
Reviewed By: RKSimon
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D56971
llvm-svn: 352260
Summary:
I'm not sure why we were using SEXTLOAD. EXTLOAD seems more appropriate since we don't care about the upper bits.
This patch changes this and then modifies the X86 post legalization combine to emit a extending shuffle instead of a sign_extend_vector_inreg. Could maybe use an any_extend_vector_inreg, but I just did what we already do in LowerLoad. I think we can actually get rid of this code entirely if we switch to -x86-experimental-vector-widening-legalization.
On AVX512 targets I think we might be able to use a masked vpmovzx and not have to expand this at all.
Reviewers: RKSimon, spatel
Reviewed By: RKSimon
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D57186
llvm-svn: 352255
We also need to combine to masked truncating with saturation stores, but I'm leaving that for a future patch.
This does regress some tests that used truncate wtih saturation followed by a masked store. Those now use a truncating store and use min/max to saturate.
Differential Revision: https://reviews.llvm.org/D57218
llvm-svn: 352230
This seems unnecessarily complicated because we gave names to
opposite polarity bools and have code comments that don't really
line up with the logic.
Step 1: remove UndefUpper and assert that it is the opposite of
UndefLower after the initial early exit.
llvm-svn: 352217
Simplify to the generic ISD::ADD/SUB if we don't make use of the result flag.
This mainly helps with ADDCARRY/SUBBORROW intrinsics which get expanded to X86ISD::ADD/SUB but could be simplified further.
Noticed in some of the test cases in PR31754
Differential Revision: https://reviews.llvm.org/D57234
llvm-svn: 352210
This isn't the final fix for our reduction/horizontal codegen, but it takes care
of a lot of the problems. After we narrow the shuffle, existing combines for
insert/extract and binops kick in, and we end up with cheaper 128-bit ops.
The avg and mul reduction tests show an existing shuffle lowering hole for
AVX2/AVX512. I think in its most minimal form this is:
https://bugs.llvm.org/show_bug.cgi?id=40434
...but we might need multiple fixes to get it right.
Differential Revision: https://reviews.llvm.org/D57156
llvm-svn: 352209
As noted in D57156, we want to check at least part of
this pattern earlier (in combining), so this will allow
the code to be shared instead of duplicated.
llvm-svn: 352127
For constant bit select patterns, replace one AND with a ANDNP, allowing us to reuse the constant mask. Only do this if the mask has multiple uses (to avoid losing load folding) or if we have XOP as its VPCMOV can handle most folding commutations.
This also requires computeKnownBitsForTargetNode support for X86ISD::ANDNP and X86ISD::FOR to prevent regressions in fabs/fcopysign patterns.
Differential Revision: https://reviews.llvm.org/D55935
llvm-svn: 351819
Summary:
Use X86ISD::VFPROUND in the instruction isel patterns. Add new patterns for ISD::FP_ROUND to maintain support for fptrunc in IR.
In the process I found a couple duplicate isel patterns which I also deleted in this patch.
Reviewers: RKSimon, spatel
Reviewed By: RKSimon
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D56991
llvm-svn: 351762
Summary:
For compress, a select node doesn't semantically reflect the behavior of the instruction. The mask would have holes in it, but the resulting write is to contiguous elements at the bottom of the vector.
Furthermore, as far as the compressing and expanding is concerned the behavior is depended on the mask. You can't just have an expand/compress node that only reads the input vector. That node would have no meaning by itself.
This all only works because we pattern match the compress/expand+select back to the instruction. But conceivably an optimization of the select could break the pattern and leave something meaningless.
This patch modifies the expand and compress node to take the mask and passthru as additional inputs and gets rid of the select all together.
Reviewers: RKSimon, spatel
Reviewed By: RKSimon
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D57002
llvm-svn: 351761
to reflect the new license.
We understand that people may be surprised that we're moving the header
entirely to discuss the new license. We checked this carefully with the
Foundation's lawyer and we believe this is the correct approach.
Essentially, all code in the project is now made available by the LLVM
project under our new license, so you will see that the license headers
include that license only. Some of our contributors have contributed
code under our old license, and accordingly, we have retained a copy of
our old license notice in the top-level files in each project and
repository.
llvm-svn: 351636
Summary:
Right now we include ${TGT}GenCallingConv.inc once per each instruction
selection method implemented by ${TGT}:
- ${TGT}ISelLowering.cpp
- ${TGT}CallLowering.cpp
- ${TGT}FastISel.cpp
Instead, add a mechanism to tablegen for marking a particular convention
as "External", which causes tablegen to emit into the ::llvm namespace,
instead of as a static helper. This allows us to provide a header to
forward declare it, so we can simply call the function from all the
places it is referenced. Typically the calling convention analyzer is
called indirectly, so it doesn't benefit from inlining.
This saves a bit of final binary size, but mostly just saves object file
size:
before after diff artifact
12852K 12492K -360K X86ISelLowering.cpp.obj
4640K 4280K -360K X86FastISel.cpp.obj
1704K 2092K +388K X86CallingConv.cpp.obj
52448K 52336K -112K llc.exe
I didn't collect before numbers for X86CallLowering.cpp.obj, which is
for GlobalISel, but we should save 360K there as well.
This patch applies the strategy to the X86 backend, but there is no
reason it couldn't be applied to the other backends that implement
multiple ISel strategies, like AArch64.
Reviewers: craig.topper, hfinkel, efriedma
Subscribers: javed.absar, kristof.beyls, hiraditya, llvm-commits
Differential Revision: https://reviews.llvm.org/D56883
llvm-svn: 351616
This sends these intrinsics through isel in a much more normal way. This should allow addressing mode matching in isel to make better use of the displacement field.
llvm-svn: 351583
This sends these intrinsics through isel in a much more normal way. This should allow addressing mode matching in isel to make better use of the displacement field.
Differential Revision: https://reviews.llvm.org/D56827
llvm-svn: 351570
Previously we used ISD::SHL and ISD::SRL to represent these in SelectionDAG. ISD::SHL/SRL interpret an out of range shift amount as undefined behavior and will constant fold to undef. While the intrinsics are defined to return 0 for out of range shift amounts. A previous patch added a special node for VPSRAV to produce all sign bits.
This was previously believed safe because undefs frequently get turned into 0 either from the constant pool or a desire to not have a false register dependency. But undef is treated specially in some optimizations. For example, its ignored in detection of vector splats. So if the ISD::SHL/SRL can be constant folded and all of the elements with in bounds shift amounts are the same, we might fold it to single element broadcast from the constant pool. This would not put 0s in the elements with out of bounds shift amounts.
We do have an existing InstCombine optimization to use shl/lshr when the shift amounts are all constant and in bounds. That should prevent some loss of constant folding from this change.
Patch by zhutianyang and Craig Topper
Differential Revision: https://reviews.llvm.org/D56695
llvm-svn: 351381
This cleans up the duplication we have with both intrinsic isel patterns and vselect isel patterns. This should also allow the intrinsics to get SimplifyDemandedBits support for the condition.
I've switched the canonical pattern in isel to use the X86ISD::BLENDV node instead of VSELECT. Since it always seemed weird to move from BLENDV with its relaxed rules on condition bits to VSELECT which has strict rules about all bits of the condition element being the same. Its more correct to go from VSELECT to BLENDV.
Differential Revision: https://reviews.llvm.org/D56771
llvm-svn: 351380
If we're going to generate a new inverted setcc, we should make sure we will be able to remove the old setcc.
Differential Revision: https://reviews.llvm.org/D56765
llvm-svn: 351378
Remove the existing assertion and just return false for unexpected shuffle value types (<X x i1> mainly....).
Found while updating combineX86ShufflesRecursively to run within SimplifyDemandedVectorElts/SimplifyDemandedBits.
llvm-svn: 351365
combineX86ShufflesRecursively is pretty cumbersome with a lot of arguments that only matter later in recursion.
This commit adds a wrapper version that only takes the initial root Op to simplify calls that don't need to worry about these.
An early, cleanup step towards merging combineX86ShufflesRecursively into SimplifyDemandedVectorElts/SimplifyDemandedBits.
llvm-svn: 351352
I was trying to prevent shuffle regressions while matching more horizontal ops
and ended up here:
shuf (extract X, 0), (extract X, 4), Mask --> extract (shuf X, undef, Mask'), 0
The affected tests were added for:
https://bugs.llvm.org/show_bug.cgi?id=34380
This patch won't change the examples in the bug report itself, but we should be
able to extend this to catch more types.
Differential Revision: https://reviews.llvm.org/D56756
llvm-svn: 351346
Summary:
Make recoverfp intrinsic target-independent so that it can be implemented for AArch64, etc.
Refer D53541 for the context. Clang counterpart D56748.
Reviewers: rnk, efriedma
Reviewed By: rnk, efriedma
Subscribers: javed.absar, kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D56747
llvm-svn: 351281
That's really what it is. If we didn't use intrinsics for BLENDVPS/BLENDVPD/PBLENDVB all the way to isel, this is the node we would use.
llvm-svn: 351278
We're trying to have the vXi1 types in IR as much as possible. This prevents the need for bitcasts when the producer of the mask was already a vXi1 value like an icmp. The bitcasts can be subject to code motion and interfere with basic block at a time isel in bad ways.
llvm-svn: 351275
In keeping with our general direction of having the vXi1 type present in IR, this patch converts the mask argument for avx512 gather to vXi1. This can avoid k-register to GPR to k-register transitions late in codegen.
I left the existing intrinsics behind because they have many out of tree users such as ISPC. They generate their own code and don't go through the autoupgrade path which only works for bitcode and ll parsing. Ideally we will get them to migrate to target independent intrinsics, but it might be easier for them to migrate to these new intrinsics.
I'll work on scatter and gatherpf/scatterpf next.
Differential Revision: https://reviews.llvm.org/D56527
llvm-svn: 351234
Modify getRegForInlineAsmConstraint to return special singleton
register class when a constraint references ST(7) not RFP80 for which
ST(7) is not a member.
llvm-svn: 351206
If we're shuffling with a zero vector, then we are better off not doing VECTOR_SHUFFLE(UNPCK()) as we lose track of those zero elements.
We were already doing this for SSSE3 targets as we have PSHUFB, but its worth doing for all targets.
llvm-svn: 351203
If we have PSHUFB and we're shuffling with a zero vector, then we are better off not doing VECTOR_SHUFFLE(UNPCK()) as we lose track of those zero elements.
llvm-svn: 351103
add (extractelt (X, 0), extractelt (X, 1)) --> extractelt (hadd X, X), 0
This is the integer sibling to D56011.
There's an additional restriction to only to do this transform in the
case where we don't have extra extracts from the source vector. Without
that, we can fail to match larger horizontal patterns that are more
beneficial than this minimal case. An improvement to the more general
h-op lowering may allow us to remove the restriction here in a follow-up.
llvm-svn: 351093
We can't represent this properly with vselect like we normally do. We also have to update the instruction definition to use a VK2WM mask instead of VK4WM to represent this.
Fixes another case from PR34877
llvm-svn: 351018
We can't represent this properly with vselect like we normally do. We also have to update the instruction definition to use a VK2WM mask instead of VK4WM to represent this.
Fixes another case from PR34877.
llvm-svn: 351017
This patch takes some of the code from D49837 to allow us to enable ISD::ABS support for all SSE vector types.
Differential Revision: https://reviews.llvm.org/D56544
llvm-svn: 350998
The 128-bit input produces 64-bits of output and fills the upper 64-bits with 0. The mask only applies to the lower elements. But we can't represent this with a vselect like we normally do.
This also avoids the need to have a special X86ISD::SELECT when avx512bw isn't enabled since vselect v8i16 isn't legal there.
Fixes another instruction for PR34877.
llvm-svn: 350994
We no longer need to extend mask scalars before bitcasting them to vXi1. This was only needed for the truncate intrinsics. And was really a bug in our lowering of them.
llvm-svn: 350991
We still use i8 for the load/store type. So we need to convert to/from i16 to around the mask type.
By doing this we get an i8->i16 extload which we can then pattern match to a KMOVW if the access is aligned.
llvm-svn: 350989
We can't properly represent this with a vselect since the upper elements of the result are supposed to be zeroed regardless of the mask.
This also reuses the new nodes even when the result type fits in 128 bits if the input is q/d and the result is w/b since vselect w/b using k-register condition isn't legal without avx512bw. Currently we're doing this even when avx512bw is enabled, but I might change that.
This fixes some of PR34877
llvm-svn: 350985
Previously, we limited this transform to cases where the
extraction into the build vector happens from vectors of
the same type as the build vector, but that's not required.
There's a slight potential regression seen in the AVX512
result for phadd -- we're using the 256-bit flavor of the
instruction now even though the 128-bit subset is sufficient.
The same problem could already be seen in the AVX2 result.
Follow-up patches will attempt to narrow that back down.
llvm-svn: 350928
We were lowering the last step extract_vector_elt to a bitcast+truncate. Change it to use an extract_vector_elt of index 0 instead. Add isel patterns to do the equivalent of what the bitcast would have done. Plus an isel pattern for an any_extend+extract to prevent some regressions.
Finally add a DAG combine to turn v1i1 scalar_to_vector+extract_vector_elt of 0 into an extract_subvector.
This fixes some of the regressions from D350800.
llvm-svn: 350918
This extends to combineVSelectToShrunkBlend to be able to resimplify SHRUNKBLENDS that have already been created.
This should help some of the regressions from D56387
Differential Revision: https://reviews.llvm.org/D56421
llvm-svn: 350875
Despite what the comment says, FCMP_UNE would be an OR not an AND. In the lowering code the first branch created still goes to the original destination. The second branch was exchanged to go to where the subsequent unconditional branch went. This is different than what we do for FCMP_OEQ where both branches that we create go to the original unconditional branch.
As far as I can tell, I think this means we don't need to exchange the branch target with the unconditional branch for FCMP_UNE at all.
Differential Revision: https://reviews.llvm.org/D56309
llvm-svn: 350873
When we use the partial-matching function on a 128-bit chunk, we must
account for the possibility that we've matched undef halves of the
original source vectors, so the outputs may need to be reset.
This should allow closing PR40243:
https://bugs.llvm.org/show_bug.cgi?id=40243
llvm-svn: 350830
This is a partial fix for:
https://bugs.llvm.org/show_bug.cgi?id=40243
...as seen in the integer test, we still need to correct the result when using the
existing (old) horizontal op matching function because it does not model the way
x86 256-bit horizontal ops return results (each 128-bit half is its own horizontal-op).
A potential follow-up change for that is discussed in the bug report - see also D56490.
This generally duplicates a lot of the existing matching code, but we can't just remove
that without introducing regressions, so the existing code is renamed and used less often.
Follow-ups may try to reduce that overlap.
Differential Revision: https://reviews.llvm.org/D56450
llvm-svn: 350826
Found while trying to figure out why my second version of D56421 worked better than the first version. We weren't deleting the vselect in a timely fashion and that caused SimplfyDemandedBit to see an additional user.
The new version doesn't have this problem so this fix isn't needed there, but seemed like the right thing to do.
llvm-svn: 350781
When the result type is v2i64/v2f64 and the index element size is i32, the index vector has two unused elements making the type v4i32. The mask VT should match the number of memory accesses that will be made.
This is consistent with the isel patterns used for the target independent gather/scatter intrinsic.
llvm-svn: 350687
Summary: AVX512VBMI2 supports a funnel shift by immediate and a funnel shift by a variable vector.
Reviewers: spatel, RKSimon
Reviewed By: RKSimon
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D56361
llvm-svn: 350498
This is especially helpful on targets without avx512bw since we don't have a good way to convert from v16i8/v32i8 to v16i1/v32i1 for the truncate anyway. If we're just going to convert it to a GPR we might as well use pmovmskb to accomplish both.
llvm-svn: 350480
The 1st try for this was at rL350369, but it caused IR-level diffs because
our cost models differentiate custom vs. legal/promote lowering. So that was
reverted at rL350373. The cost models were fixed independently at rL350403,
so this is effectively the same patch as last time.
Original commit message:
This would show up if we fix horizontal reductions to narrow as they go along,
but it's an improvement for size and/or Jaguar (fast-hops) independent of that.
We need to do this late to not interfere with other pattern matching of larger
horizontal sequences.
We can extend this to integer ops in a follow-up patch.
Differential Revision: https://reviews.llvm.org/D56011
llvm-svn: 350421
This would show up if we fix horizontal reductions to narrow as they go along,
but it's an improvement for size and/or Jaguar (fast-hops) independent of that.
We need to do this late to not interfere with other pattern matching of larger
horizontal sequences.
We can extend this to integer ops in a follow-up patch.
Differential Revision: https://reviews.llvm.org/D56011
llvm-svn: 350369
As noted in PR39973 and D55558:
https://bugs.llvm.org/show_bug.cgi?id=39973
...this is a partial implementation of a fold that we do as an IR canonicalization in instcombine:
// extelt (binop X, Y), Index --> binop (extelt X, Index), (extelt Y, Index)
We want to have this in the DAG too because as we can see in some of the test diffs (reductions),
the pattern may not be visible in IR.
Given that this is already an IR canonicalization, any backend that would prefer a vector op over
a scalar op is expected to already have the reverse transform in DAG lowering (not sure if that's
a realistic expectation though). The transform is limited with a TLI hook because there's an
existing transform in CodeGenPrepare that tries to do the opposite transform.
Differential Revision: https://reviews.llvm.org/D55722
llvm-svn: 350354
INC/DEC are pretty much the same as ADD/SUB except that they don't update the C flag.
This patch removes the special nodes and just pattern matches from ADD/SUB during isel if the C flag isn't being used.
I had to avoid selecting DEC is the result isn't used. This will become a SUB immediate which will turned into a CMP later by optimizeCompareInstr. This lead to the one test change where we use a CMP instead of a DEC for an overflow intrinsic since we only checked the flag.
This also exposed a hole in our RMW flag matching use of hasNoCarryFlagUses. Our root node for the match is a store and there's no guarantee that all the flag users have been selected yet. So hasNoCarryFlagUses needs to check copyToReg and machine opcodes, but it also needs to check for the pre-match SETCC, SETCC_CARRY, BRCOND, and CMOV opcodes.
Differential Revision: https://reviews.llvm.org/D55975
llvm-svn: 350245
Peek through shift modulo masks while matching double shift patterns.
I was hoping to delay this until I could remove the X86 code with generic funnel shift matching (PR40081) but this will do for now.
Differential Revision: https://reviews.llvm.org/D56199
llvm-svn: 350222
All of these use custom isel so we can pretty easily detect the differences in the custom code in X86ISelDAGToDAG. The ISD opcodes just need to express the desired semantics not the details of how they would be selected by isel. So unifying them lets us remove the special casing from lowering.
llvm-svn: 350206
These require a different X86ISD node to be created than i16/i32/i64. I guess no one wanted to add the special code for that except in LowerXALUO. But now LowerXALUO, LowerSELECT, and LowerBRCOND all use a common helper function so they all share the special code.
Unfortunately, there are no test changes because we seem to correct the miss in a DAG combine later. I did verify it manually using test cases from xmulo.ll
llvm-svn: 350205
This makes it easier to keep the LowerBRCOND and LowerSELECT code in sync with LowerXALUO so they always pick the same operation for overflowing instructions.
This is inspired by the helper functions used by ARM and AArch64 for the same purpose.
The test change is because LowerSELECT was not in sync with LowerXALUO with regard to INC/DEC for SADDO/SSUBO.
llvm-svn: 350198
This seems to be getting in the way more than its helping. This does mean we stop scalarizing some cases, but I'm not convinced the scalarization was really better.
Some of the changes to vsel-cmp-load.ll are a regression but D56156 should fix it.
llvm-svn: 350159
This allows us to sign extend to v4i32 first. And then share that extension to implement the final steps to v4i64 using a pcmpgt and punpckl and punpckh.
We already do something similar for SIGN_EXTEND with -x86-experimental-vector-widening-legalization.
llvm-svn: 350158
This was tricking us into making these operations and then letting them get scalarized later. But I can't prove that the scalarized version is actually better.
llvm-svn: 350141
Previously we emitted a multiply and some masking that was supposed to matched to PMULUDQ, but the masking could sometimes be removed before we got a chance to match it. So instead just emit the PMULUDQ directly.
Remove the DAG combine that was added when the ReplaceNodeResults code was originally added. Add a new DAG combine to avoid regressions in shrink_vmul.ll
Some of the shrink_vmul.ll test cases now pick PMULUDQ instead of PMADDWD/PMULLD, but I think this should be an improvement on most CPUs.
I think all of this can go away if/when we switch to -x86-experimental-vector-widening-legalization
llvm-svn: 350134
Create PMULDQ/PMULUDQ as long as the number of elements is a power of 2.
This seems to give some improvements in our ability to use SimplifyDemandedBits.
llvm-svn: 350084
Make each of the helper functions only return their comparison node and the condition code. Leave X86ISD::SETCC creation to the LowerSETCC function itself.
Looking into whether we can use this code directly in BRCOND and SELECT lowering instead of going through LowerSETCC which creates an X86ISD::SETCC node we need to look through.
llvm-svn: 350082
Only one of the 3 callers of LowerAndToBT need the SETCC node. Two of them have to look through it to find the operands they really need. Instead create it after the one call that needs it.
LowerAndToBT now returns both the BT node and the X86 specific condition code separately.
llvm-svn: 350081
This is an alternative to what I attempted in D56057.
GetDemandedBits is a special version of SimplifyDemandedBits that allows simplifications even when the operand has other uses. GetDemandedBits will only do simplifications that allow a node to be bypassed. It won't create new nodes or alter any of the other users.
I had to add support for bypassing SIGN_EXTEND_INREG to GetDemandedBits.
Based on a patch that Simon Pilgrim sent me in email.
Fixes PR40142.
llvm-svn: 350059
Remove the TESTmr isel patterns and add another postprocessing combine for TESTrr+ANDrm->TESTmr. We already have a postprocessing combine for TESTrr+ANDrr->TESTrr. With this we can give ANDN a chance to match first. And clean it up during post processing if we ended up with just a regular AND.
This is another step towards my plan to gut EmitTest and do more flag handling during isel matching or by using optimizeCompare.
llvm-svn: 350038
This is admittedly a narrow fix for the problem:
https://bugs.llvm.org/show_bug.cgi?id=37502
...but as the XOP restriction shows, it's a maze to get this right.
In the motivating example, note that we have movddup before SSE4.1 and
again with AVX2. That's because insertps isn't available pre-SSE41 and
vbroadcast is (more generally) available with AVX2 (and the splat is
reduced to movddup via isel pattern).
Differential Revision: https://reviews.llvm.org/D55898
llvm-svn: 349937
Migrate the X86 backend from X86ISD opcodes ADDS and SUBS to generic
ISD opcodes SADDSAT and SSUBSAT. This also improves scodegen for
@llvm.sadd.sat() and @llvm.ssub.sat() intrinsics.
This is a followup to D55787 and part of PR40056.
Differential Revision: https://reviews.llvm.org/D55833
llvm-svn: 349520
InstCombine seems to canonicalize or PSUB patter into a max with the cosntant and an add with an inverse of the constant.
This patch recognizes this pattern and turns it into PSUBUS. Future work could improve undef element handling.
Fixes some of PR40053
Differential Revision: https://reviews.llvm.org/D55780
llvm-svn: 349519
Replace the X86ISD opcodes ADDUS and SUBUS with generic ISD opcodes
UADDSAT and USUBSAT. As a side-effect, this also makes codegen for
the @llvm.uadd.sat and @llvm.usub.sat intrinsics reasonable.
This only replaces use in the X86 backend, and does not move any of
the ADDUS/SUBUS X86 specific combines into generic codegen.
Differential Revision: https://reviews.llvm.org/D55787
llvm-svn: 349481
(VSRAI (VSHLI X, C1), C1) --> X iff NumSignBits(X) > C1
This works better as part of SimplifyDemandedBits than part of the general combine.
llvm-svn: 349462
This fold was incredibly specific - replace with a SimplifyDemandedBits fold to remove a VSRAI if only the original sign bit is demanded (its guaranteed to stay the same).
Test change is merely a rescheduling.
llvm-svn: 349459
Convert VSRAI to VSRLI is the sign bit is known zero and improve KnownBits output for all shift instruction.
Fixes the poor codegen comments in D55768.
llvm-svn: 349407
This allows a TEST to be used and can be combined with any AND that may already exist as an input to the shift.
This was already done in EmitTest, but was easily tricked by multiple uses because the setcc might be used by multiple instructions. Once the SETCC and users are legalized then we can look for the shift to be used by a single CMP, but the CMP itself can have multiple users.
This appears to fix the case in PR39968.
llvm-svn: 349385
This is an initial patch to add the necessary support for a DemandedElts argument to SimplifyDemandedBits, more closely matching computeKnownBits and to help improve vector codegen.
I've added only a small amount of the changes necessary to get at least one test to update - a lot more can be done but I'd like to add these methodically with proper test coverage, at the same time the hope is to slowly move some/all of SimplifyDemandedVectorElts into SimplifyDemandedBits as well.
Differential Revision: https://reviews.llvm.org/D55768
llvm-svn: 349374
I'd like to try to move a lot of the flag matching out of EmitTest and push it to isel or isel preprocessing. This is a step towards that.
The test-shrink-bug.ll changie is an improvement because we are no longer interfering with test shrink handling in isel.
The pr34137.ll change is a regression, but the IR came from -O0 and was not reduced by InstCombine. So it contains a lot of redundancies like duplicate loads that made it combine poorly.
llvm-svn: 349315
Use consistent rules for when to lower to SHLD/SHRD for slow machines - fixes a weird issue where funnel shift gets expanded but then X86ISelLowering's combineOr sees the optsize and combines to SHLD/SHRD, but now with the modulo amount guard......
llvm-svn: 349285
Summary:
If the setcc already has the target desired type we can reach the getSetCC/getSExtOrTrunc after the MatchingVecType check with the exact same types as the nodes we started with. This causes those causes VsetCC to be CSEd to N0 and the getSExtOrTrunc will CSE to N. When we return N, the caller will think that meant we called CombineTo and did our own worklist management. But that's not what happened. This prevents target hooks from being called for the node.
To fix this, I've now returned SDValue if the setcc is already the desired type. But to avoid some regressions in X86 I've had to disable one of the target combines that wasn't being reached before in the case of a (sext (setcc)). If we get vector widening legalization enabled that entire function will be deleted anyway so hopefully this is only for the short term.
Reviewers: RKSimon, spatel
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D55459
llvm-svn: 349137
This requires the two callers to manifest a 0 to make EmitCmp call EmitTest.
I'm looking into changing how we combine TEST and flag setting instructions to not be part of lowering. And instead be part of DAG combine or isel. Which will mean EmitTest will probably become gutted and maybe disappear entirely.
llvm-svn: 349094
There's still a couple of minor SimplifyDemandedElts regressions in some of the shift amount splats that will be fixed in future patches.
llvm-svn: 349052
Move existing rotation expansion code into TargetLowering and set it up for vectors as well.
Ideally this would share more of the funnel shift expansion, but we handle the shift amount modulo quite differently at the moment.
Begun removing x86 vector rotate custom lowering to use the expansion.
llvm-svn: 349025
I'm hoping we can just replace SETCC_CARRY with SBB. This is another step towards that.
I've explicitly used zero as the input to the setcc to avoid a false dependency that we've had with the SETCC_CARRY. I changed one of the patterns that used NEG to instead use an explicit compare with 0 on the LHS. We needed the zero anyway to avoid the false dependency. The negate would clobber its input register. By using a CMP we can avoid that which could be useful.
Differential Revision: https://reviews.llvm.org/D55414
llvm-svn: 348959
This patch introduces a generic function to determine whether a given vector type is known to be a splat value for the specified demanded elements, recursing up the DAG looking for BUILD_VECTOR or VECTOR_SHUFFLE splat patterns.
It also keeps track of the elements that are known to be UNDEF - it returns true if all the demanded elements are UNDEF (as this may be useful under some circumstances), so this needs to be handled by the caller.
A wrapper variant is also provided that doesn't take the DemandedElts or UndefElts arguments for cases where we just want to know if the SDValue is a splat or not (with/without UNDEFS).
I had hoped to completely remove the X86 local version of this function, but I'm seeing some regressions in shift/rotate codegen that will take a little longer to fix and I hope to get this in sooner so I can continue work on PR38243 which needs more capable splat detection.
Differential Revision: https://reviews.llvm.org/D55426
llvm-svn: 348953
This should really be generalized to allow increment and/or
we should replace it by using ISD::matchUnaryPredicate().
See D55515 for context.
llvm-svn: 348776
Previously we had to take the carry in and add -1 to it to set the carry flag so we could use it with ADC/SBB. But if we know its 0 then we don't need to bother.
This should go a long way towards fixing PR24545.
llvm-svn: 348727
This addresses a FIXME and avoids depending on an isel pattern match I think. I've remove the isel patterns too since he have no lit tests left that cover them. Hopefully that really means they are unused.
I'm trying to decide if we need SETCC_CARRY. This removes one of its usages.
Differential Revision: https://reviews.llvm.org/D55355
llvm-svn: 348536
Initial step towards making the function more generic (and probably move into SelectionDAG).
This is necessary to avoid massive codegen bloat for PR38243 (Add modulo rotate support to LowerRotate).
llvm-svn: 348498
Prep work for PR38243 - mainly adding comments on where we need to add modulo support (doing so at the moment causes massive codegen regressions).
I've also consistently added support for modulo folding for uniform constants (although at the moment we have no way to trigger this) and removed the old assertions.
llvm-svn: 348366
This is an initial patch to add a minimum level of support for funnel shifts to the SelectionDAG and to begin wiring it up to the X86 SHLD/SHRD instructions.
Some partial legalization code has been added to handle the case for 'SlowSHLD' where we want to expand instead and I've added a few DAG combines so we don't get regressions from the existing DAG builder expansion code.
Differential Revision: https://reviews.llvm.org/D54698
llvm-svn: 348353
PR17686 demonstrates that for some targets FP exceptions can fire in cases where the FP_TO_UINT is expanded using a FP_TO_SINT instruction.
The existing code converts both the inrange and outofrange cases using FP_TO_SINT and then selects the result, this patch changes this for 'strict' cases to pre-select the FP_TO_SINT input and the offset adjustment.
The X87 cases don't need the strict flag but generates much nicer code with it....
Differential Revision: https://reviews.llvm.org/D53794
llvm-svn: 348251
We only needed this because it provided really aggressive constant folding even through constant pool entries created from build_vectors. The main case was for vXi8 MULH legalization which was happening as part of legalize DAG instead of as part of legalize vector ops. Now its part of vector op legalization and we've added special handling for build vectors of all constants there. This has removed the need for this code on the list tests we have.
llvm-svn: 348237
This is the smallest vector enhancement I could find to D54640.
Here, we're allowing narrowing to only legal vector ops because we'll see
regressions without that. All of the test diffs are wins from what I can tell.
With AVX/AVX512, we can shrink ymm/zmm ops to xmm.
x86 vector multiplies are the problem case that we're avoiding due to the
patchwork ISA, and it's not clear to me if we can dance around those
regressions using TLI hooks or if we need preliminary patches to plug those
holes.
Differential Revision: https://reviews.llvm.org/D55126
llvm-svn: 348195
Summary:
We need to unpackl and unpackh the operands to use two vXi16 multiplies. Previously it looks like the low unpack would get constant folded at least in the 128-bit case after shuffle lowering turned the unpackl into ZERO_EXTEND_VECTOR_INREG and X86 custom DAG combined it. The same doesn't happen for the high half. So we'd load a constant and then shuffle it. But the low half would just be loaded and used by the multiply directly.
After this patch we now end up with a constant pool entry for the low and high unpacks separately with no shuffle operations.
This is a step towards removing custom constant folding for ZERO_EXTEND_VECTOR_INREG/SIGN_EXTEND_VECTOR_INREG in the X86 backend.
Reviewers: RKSimon, spatel
Reviewed By: RKSimon
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D55165
llvm-svn: 348159
Summary:
Under -x86-experimental-vector-widening-legalization, fp_to_uint/fp_to_sint with a smaller than 128 bit vector type results are custom type legalized by promoting the result to a 128 bit vector by promoting the elements, inserting an assertzext/assertsext, then truncating back to original type. The truncate will be further legalizdd to a pack shuffle. In the case of a v8i8 result type, we'll end up with a v8i16 fp_to_sint. This will need to be further legalized during vector op legalization by promoting to v8i32 and then truncating again. Under avx2 this produces good code with two pack instructions, but Under avx512 this will result in a truncate instruction and a packuswb instruction. But we should be able to get away with a single truncate instruction.
The other option is to promote all the way to vXi32 result type during the first type legalization. But in some experimentation that seemed to require more work to produce good code for other configurations.
Reviewers: RKSimon, spatel
Reviewed By: RKSimon
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D54836
llvm-svn: 348158
Previously this code generated its own extracts and build_vector. But we can use a simpler concat_vectors or scalar_to_vector operation and let type legalization do additional legalization of those operations.
llvm-svn: 348087
The generic legalizer will fall back to a stack spill that uses a truncating store. That store will get expanded into a shuffle and non-truncating store on pre-avx512 targets. Once that happens the stack store/load pair will be combined away leaving behind the shuffle and bitcasts. On avx512 targets the truncating store is legal so doesn't get folded away.
By custom legalizing it we can avoid this churn and maybe produce better code.
llvm-svn: 348085
Summary: With sse4.1 we use two zero_extend_vector_inreg and a pshufd to expand the v16i8 input into two v8i16 vectors for the multiply. That's 3 shuffles to extend one operand. The other operand is usually constant as this is mostly used by division by constant optimization. Pre sse4.1 we use a punpckhbw and a punpcklbw with a zero vector. That's two shuffles and an xor and a copy due to tied register constraints. That seems maybe better than the 3 shuffles. With AVX we avoid the copy so that's obviously better.
Reviewers: spatel, RKSimon
Reviewed By: RKSimon
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D55138
llvm-svn: 348079
This reduces the number of shuffle operations that need to be done. The splitting strategy requires the shuffle unit for the extraction and the extension. With the unpack strategy the unpacks accomplish a splitting and extending in one operation.
llvm-svn: 348019
This does require a constant pool load instead of loading an immediate into a gpr, moving to a k register and masking. But its less instructions and more consistent with previous ISAs. It probably opens up more combine opportunities as one of the test cases demonstrates.
llvm-svn: 348018
Previously we emitted a punpcklbw/punpckhbw to move the byte elements into the upper half of 16 bit elements then shifted right by 8 to zero the upper bits. After DAG combine we end up with punpcklbw/punpckhbw into the lower bits with zeros in the uppers bits and no shifts. So just emit that directly.
llvm-svn: 347966
We had a EVT variable capturing the result of getSimpleValueType which returns an MVT. Another place using EVT that could have been MVT. And an 'int' that should be 'unsigned'.
llvm-svn: 347959
I believe we should be legalizing these with the rest of vector binary operations. If any custom lowering is required for these nodes, this will give the DAG combine between LegalizeVectorOps and LegalizeDAG to run on the custom code before constant build_vectors are lowered in LegalizeDAG.
I've moved MULHU/MULHS handling in AArch64 from Lowering to isel. Moving the lowering earlier caused build_vector+extract_subvector simplifications to kick in which made the generated code worse.
Differential Revision: https://reviews.llvm.org/D54276
llvm-svn: 347902
This is another patch for -x86-experimental-vector-widening. This pre widens narrow division by constants so that we can get pass the legal type check in the generic DAG combiner. Otherwise we end up scalarizing.
I've restricted this to splats for now because it was easy to just call DAG.getConstant. Not sure what we should do for non-splat? Increase the element size?Widen the constant vector by padding with 1?
Differential Revision: https://reviews.llvm.org/D54919
llvm-svn: 347898
This failed to select (which might be a separate bug) in
X86ISelDAGToDAG because we try to create a select node
that can be simplified away after rL347227.
This change avoids the problem by simplifying the SHRUNKBLEND
node sooner. In the test case, we manage to realize that the
true/false values of the select (SHRUNKBLEND) are the same thing,
so it simplifies away completely.
llvm-svn: 347818
Expansion of SIGN_EXTEND_INREG can create a VSRAI instruction. If there is already a VSRAI after it, we should combine them into a larger VSRAI
Differential Revision: https://reviews.llvm.org/D54959
llvm-svn: 347784
We're already mixing this APInt with other 'unsigned' variables. This allows us to use regular comparison operators instead of needing to use APInt::ult or APInt::uge. And it removes a later conversion from APInt to unsigned.
I might be adding another combine to this function and this will probably simplify the logic required for that.
llvm-svn: 347684
If we fold the bitcast into the store we'll end up creating a truncating store to vXi1 that will get scalarized. Instead allow the bitcast to be turned into a movmsk.
We probably need to do something if the store itself is a vXi1 type, but I'll leave that til a testcase appears.
llvm-svn: 347632
SplitVecOp_TruncateHelper tries to promote the result type while splitting FP_TO_SINT/UINT. It then concatenates the result and introduces a truncate to the original result type. But it does this without inserting the AssertZExt/AssertSExt that the regular result type promotion would insert. Nor does it turn FP_TO_UINT into FP_TO_SINT the way normal result type promotion for these operations does. This is bad on X86 which doesn't support FP_TO_SINT until AVX512.
This patch disables the use of SplitVecOp_TruncateHelper for these operations and just lets normal promotion handle it. I've tweaked a couple things in X86ISelLowering to avoid a few obvious regressions there. I believe all the changes on X86 are improvements. The other targets look neutral.
Differential Revision: https://reviews.llvm.org/D54906
llvm-svn: 347593
We have these 2 "isDesirable" promotion hooks (I'm not sure why we need both of them, but that's
independent of this patch), and we can adjust them to promote "mul i8 X, C" to i32. Then, all of
our existing LEA and other multiply expansion magic happens as it would for i32 ops.
Some of the test diffs show that we could end up with an actual 32-bit mul instruction here
because we choose not to expand to simpler ops. That instruction could be slower depending on the
subtarget. On the plus side, this means we don't need a separate instruction to load the constant
operand and possibly an extra instruction to move the result. If we need to tune mul i32 further,
we could add a later transform that tries to shrink it back to i8 based on subtarget timing.
I did not bother to duplicate all of the 32-bit test file RUNs and target settings that exist to
test whether LEA expansion is cheap or not. The diffs here assume a default target, so that means
LEA is generally cheap.
Differential Revision: https://reviews.llvm.org/D54803
llvm-svn: 347557
This should likely be adjusted to limit this transform
further, but these diffs should be clear wins.
If we have blendv/conditional move, then we should assume
those are cheap ops. The loads become independent of the
compare, so those can be speculated before we need to use
the values in the blend/mov.
llvm-svn: 347526
We can't guarantee that demanded bits passing through the vector shuffle won't cause the AND in front of this to be removed. This would prevent the PACKUS from being matched during shuffle lowering.
Unfortunately, this adds a packuswb to one of the vector-reduce-mul.ll tests since we were removing the shuffle via SimplifyDemandedVectorElts. We appear to have similar issues with vpmovwb on the same test case on other targets.
llvm-svn: 347361
Previously we emitted to separate shuffles, one for unpcklbw and one for unpcklwd. Instead emit a single shuffle equivalent to both of the original shuffles. Shuffle lowering seems able to handle it. This avoids a bitcast between the two shuffles which seems helpful to DAG combine.
Remove the custom type legalization for v8i8->v8i32. I had put that in to avoid some almost duplicate punpcklbw instructions I was seeing, but this lowering change seems to fix that. It also fixes some duplicate shuffles seen in vector-sext.ll
llvm-svn: 347348
Pull out getPackDemandedElts demanded elts remapping helper from computeKnownBitsForTargetNode and use in computeKnownBits/ComputeNumSignBits.
llvm-svn: 347303
Previously if V2 was unused we ended up using V1 for both inputs as part of the code that follows the new code. By using lowerVectorShuffleWithUNPCK we keep the undef nature of V2 in the output.
As near as I can tell this makes v16i8 behavior consistent with every other VT now.
This does mean that we give the register allocator freedom to fill in random registers now and create false dependencies. But like I said we're already doing that for other types.
llvm-svn: 347296
getZeroVector produces a specifically canonicalized zero vector, but we can just let DAG legalization take care of it.
The test changes are because MULH lowering happens later than it should and this change gave us the opportunity to constant fold away a multiply during a DAG combine before the build_vector got legalized with a bitcast.
llvm-svn: 347290
SSE PSHUFB vector ctlz lowering works at the i4 nibble level. As detailed in PR39703, we were masking the lower nibble off but we only actually use it in the case where the upper nibble is known to be zero, making it safe to remove the mask and save an instruction.
Differential Revision: https://reviews.llvm.org/D54707
llvm-svn: 347242
Previously we split the vectors in half to allow the two halves to be any extended then concatenated the results back together.
This patch instead instead extends the v16i8 sse algorithm to extend half of each 128-bit lane using punpcklbw/punpckhbw. Multiplies all the low half lanes and high half lanes together in separate operations. Then merges the half lane results back together using packuswb.
Unfortunately, some of the cases in vector-reduce-mul.ll regress because we aren't narrowing the vector width of the multiplies as we reduce. The splitting was somewhat making up for that before by causing halves to be discarded after the split.
Differential Revision: https://reviews.llvm.org/D54668
llvm-svn: 347240
The shift requires a copy to avoid clobbering a register. Comparing with 0 uses an xor to produce 0 that will be overwritten with the compare results. So still requires 2 instructions, but should be one byte shorter since it doesn't need to encode an immediate.
llvm-svn: 347185
Previously we used an arithmetic shift right by 31, but that requires a copy to preserve the input. So we might as well materialize a zero and compare to it since the comparison will overwrite the register that contains the zeros. This should be one byte shorter.
llvm-svn: 347181
Leave just the v4i8->v4i64 and v8i8->v8i64, but only enable them on pre-sse4.1 targets when 64-bit mode is enabled. In those cases we end up creating sext loads that get scalarized to code that looks better than what we get from loading into a vector register and doing a multiple step sign extend using unpacks and shifts.
llvm-svn: 347180
Pre-SSE4.1 sext_invec for v2i64 is complicated because we don't have a v2i64 sra instruction. So instead we sign extend to i32 using unpack and sra, then copy the elements and do a v4i32 sra to fill with sign bits, then interleave the i32 sign extend and the sign bits. So really we're doing to two sign extends but only using half of the v4i32 intermediate result.
When the result is more than 128 bits, default type legalization would prefer to split the destination type all the way down to v2i64 with shuffles followed by v16i8/v8i16->v2i64 sext_inreg operations. This results in more instructions than necessary because we are only utilizing the lower 2 elements of the v4i32 intermediate result. Instead we can custom split a v4i8/v4i16->v4i64 sign_extend. Then we can sign extend v4i8/v4i16->v4i32 invec producing a full v4i32 result. Create the sign bit vector as a v4i32 then split and interleave with the sign bits using an punpackldq and punpackhdq.
llvm-svn: 347176
If we widen illegal types instead of promoting, we should be able to rely on the type legalizer to create the vector_inreg operations for us with some caveats.
This patch disables combineToExtendVectorInReg when we are using widening.
I've enabled custom legalization for v8i8->v8i64 extends under avx512f since the type legalizer would want to create a vector_inreg with a v64i8 input type which isn't legal without avx512bw. So we go to v16i8 with custom code using the relaxation of rules we get from D54346.
I've also enable custom legalization of v8i64 and v16i32 operations with with AVX. When the input type is 128 bits, the default splitting legalization would extend first 128->256, then do the a split to two 128 pieces. Extend each half to 256 and then concat the result. The custom legalization I've added instead uses a 128->256 bit vector_inreg extend that only reads the lower 64-bits for the low half of the split. Then shuffles the high 64-bits to the low 64-bits and does another vector_inreg extend.
llvm-svn: 347172
Summary: This is an improvement over the two pshufbs and punpcklqdq we'd get otherwise.
Reviewers: RKSimon, spatel
Reviewed By: RKSimon
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D54671
llvm-svn: 347171
Refactor towards making this recursive (necessary for PR38243 rotation splat detection).
IsSplatVector returns the original vector source of the splat and the splat index.
GetSplatValue returns the scalar splatted value as an extraction from IsSplatVector.
llvm-svn: 347168
We were using the 'normalized' shuffle mask from resolveTargetShuffleInputs, which replaces zero/undef inputs with sentinel values. For SimplifyDemandedVectorElts we need the raw mask so we can correctly demand those 'zero' inputs that got normalized away, this requires an extra bit of logic to locally normalize undef inputs.
llvm-svn: 347158
The zero extend will require two stages of unpacks to implement. So its better to shrink the multiply using pmullw and then extend that result back to v4i32 using a single unpack.
llvm-svn: 347149
This tries to force the result type to vXi32 followed by a truncate. This can help avoid scalarization that would otherwise occur.
There's some annoying examples of an avx512 truncate instruction followed by a packus where we should really be able to just use one truncate. But overall this is still a net improvement.
llvm-svn: 347105
By early promoting the multiply to use an i16 element type we can avoid op legalization emit a second multiply for the 8 upper elements of the v16i8 type we would otherwise get.
llvm-svn: 347032
We aren't going to use the upper bits of the multiply result that the extend would effect. So we don't need a specific type of extend.
This makes some reduction test cases shorter because we were previously trying to sign_extend a truncate which we can't eliminate.
llvm-svn: 347011
Removing this code doesn't affect any lit tests so it doesn't appear to be tested anymore. I assume it was when it was added, but I guess something else changed? Code coverage report also says its unused.
I mostly didn't like that it seemed to count the sign bits as if it was a sign_extend, but then set isPositive as if it was a zero_extend. It feels like we should have picked one interpretation?
Differential Revision: https://reviews.llvm.org/D54596
llvm-svn: 346995
Use unsigned to calculate the subvector index to avoid a cast.
Remove an unnecessary condition and replace it with a stronger assert.
Use the InVT variable we updated when we extracted instead of grabbing it from the In SDValue.
llvm-svn: 346983
In reduceVMULWidth, we no longer need to worry about extending the vector to 128 bits first. Regular widening of extends, muls and shuffles will take care of that for us.
In combineMulToPMADDWD, we can handle v2i32 multiplies and allow the VPMADDWD to be widened to v4i32 during type legalization by adding custom widening like we do have for AVG/ADDUS/SUBUS. I had to modify that code a little to allow different and output VTs.
Differential Revision: https://reviews.llvm.org/D54512
llvm-svn: 346980
This avoids some nasty shuffles when we have avx512. It will also prevent using zmm truncate instructions when a ymm instruction that zeroes part of an xmm register will do. Also avoid using avx512 truncate instructions when the input is 128 bits or less. These instructions are 2 uops on skx so we can probably find a better single uop shuffle like pshufb.
llvm-svn: 346936
The narrow types end up requesting widening, but generic legalization will end up scalaring and using a build_vector to do the widening.
llvm-svn: 346916
On 64-bit targets the type legalizer will use i64 to legalize these. But when i64 isn't legal, the type legalizer won't try an FP type. So do it manually instead.
There are a few regressions in here due to some v2i32 operations like mul and div now being reassembled into a full vector just to store instead of storing the pieces. But this was already occuring in 64-bit mode so its not a new issue.
llvm-svn: 346908
Narrower vectors will be widened to 128 bits without changing the element size. And generic type legalization can already handle widening mulhu/mulhs.
Differential Revision: https://reviews.llvm.org/D54513
llvm-svn: 346879
This patch removes the last use of the constant pool shuffle decode helper and consistently uses the 'getTargetShuffleMaskIndices' versions instead. The constant pool versions are now purely used for assembly comments.
The avx512vbmi intrinsic upgrades had to be altered as they were being decoded as broadcasts, similar to what I fixed in rL346032. I don't think the change is critical - although its annoying that we lose the {k}{z} instruction test coverage as they are tricky to generate....
Differential Revision: https://reviews.llvm.org/D54083
llvm-svn: 346850
Previously, the extend_vector_inreg opcode required their input register to be the same total width as their output. But this doesn't match up with how the X86 instructions are defined. For X86 the input just needs to be a legal type with at least enough elements to cover the output.
This patch weakens the check on these nodes and allows them to be used as long as they have more input elements than output elements. I haven't changed type legalization behavior so it will still create them with matching input and output sizes.
X86 will custom legalize these nodes by shrinking the input to be a 128 bit vector and once we've done that we treat them as legal operations. We still have one case during type legalization where we must custom handle v64i8 on avx512f targets without avx512bw where v64i8 isn't a legal type. In this case we will custom type legalize to a *extend_vector_inreg with a v16i8 input. After that the input is a legal type so type legalization should ignore the node and doesn't need to know about the relaxed restriction. We are no longer allowed to use the default expansion for these nodes during vector op legalization since the default expansion uses a shuffle which required the widths to match. Custom legalization for all types will prevent us from reaching the default expansion code.
I believe DAG combine works correctly with the released restriction because it doesn't check the number of input elements.
The rest of the patch is changing X86 to use either the vector_inreg nodes or the regular zero_extend/sign_extend nodes. I had to add additional isel patterns to handle any_extend during isel since simplifydemandedbits can create them at any time so we can't legalize to zero_extend before isel. We don't yet create any_extend_vector_inreg in simplifydemandedbits.
Differential Revision: https://reviews.llvm.org/D54346
llvm-svn: 346784
This patch adds the ability to use a PALIGNR to rotate a pair of inputs to select a range containing all the referenced elements, followed by a single input permute to put them in the right location.
Differential Revision: https://reviews.llvm.org/D54267
llvm-svn: 346706
Truncate and shuffle lowering are already capable of matching to PACKUS using known bits analysis.
This features one test change where we now prefer to extend v16i16->v16i32 then trunc v16i32->v16i8 over extract_subvector+packus when avx512f is available, but avx512bw is not.
llvm-svn: 346697
getConstant will create a BUILD_VECTOR for us and use a legal type if necessary. So just create the simple node and let BUILD_VECTOR legalization do the canonicalization.
llvm-svn: 346603
This gives shuffle lowering the freedom to use zero_extend_vector_inreg for the unpckl shuffle. Shuffle combining usually makes this swap later, but not when AVX512 is enabled it seems.
While there also use DAG.getConstant to create a 0 vector instead of using the helper the forces a specific BUILD_VECTOR. I don't think that helper is usually needed. We're basically free to create a constant build_vector anytime and it will be legalized on its own.
llvm-svn: 346574
With avx512f but not avx512bw we need to extend to v16i32 then truncate that to to v16i8. Previously we emitted both nodes during lowering, but I'm trying to switch to using target independent nodes and with that switched the extend+truncate wou
This patch changes the implementation to what will be necessary with that patch which helps minimize test diffs.
llvm-svn: 346552
This makes X86ISD::VSEXT more similar to ISD::SIGN_EXTEND and ISD::ZERO_EXTEND.
I'm hoping to replace X86ISD::VSEXT/VZEXT with target independent nodes. Making the target specific nodes similar to the target independent nodes helps minimize test diffs in that patch.
llvm-svn: 346539
I noticed that we weren't generating broadcasts as much I thought we would with
D54271, and this is part of the problem.
Widening the shuffle elements means adding bitcasts and hiding the relationship
between a splatted scalar and the vector. If we can form a broadcast, do that
before going through the rest of the shuffle lowering because broadcasts should
be cheap and can often be load-folded.
Differential Revision: https://reviews.llvm.org/D54280
llvm-svn: 346498
As discussed in D54073, we have a potential regression from more aggressive vector narrowing here, so let's try to avoid that by changing build-vector lowering slightly.
Insert-vector-element lowering always does this since there's no "pinsr" for ymm/zmm:
// If the vector is wider than 128 bits, extract the 128-bit subvector, insert
// into that, and then insert the subvector back into the result.
...but we can sometimes do better for insert-into-constant-vector by using shuffle lowering.
Differential Revision: https://reviews.llvm.org/D54271
llvm-svn: 346433
The main caller of this already has an MVT and several targets called getSimpleVT inside without checking isSimple. This makes the simpleness explicit.
llvm-svn: 346180
SimplifyDemandedBits can turn a sign_extend back into an any_extend and trigger an infinite loop. So instead legalize it the same way as a sign_extend, but preserve the opcode. Then just pattern match it the same as sign_extend during isel.
I don't have a reduced test case for such an infinite loop yet.
llvm-svn: 346170
v2i8/v2i16/v2i32 are promoted to v2i64. pmuludq takes a v2i64 input and produces a v2i64 output. Since we don't about the upper bits of the type legalized multiply we can use the pmuludq to produce the multiply result for the bits we do care about.
llvm-svn: 346115
Summary: This also enables some constant folding from KnownBits propagation. This helps on some cases vXi64 case in 32-bit mode where constant vectors appear as vXi32 and a bitcast. This can prevent getNode from constant folding sra/shl/srl.
Reviewers: RKSimon, spatel
Reviewed By: spatel
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D54069
llvm-svn: 346102
These methods were just wrappers around getNode with additional asserts (identical and repeated 3 times). But getNode already has a switch that can be used to hold these asserts that allows them to be shared for all 3 opcodes. This also enables checking on the places that create these nodes without using the wrappers.
The rest of the patch is just changing all callers to use getNode directly.
llvm-svn: 346087
The majority of the changes are because the rest of shuffle lowering/combining prefers to replace the undef input with the other operand. Using UNPCKL directly seemed to avoid this and just grabbed a randomish register for the undef which can create false dependencies.
llvm-svn: 346050
We already have custom lowering for the AVX case in LegalizeVectorOps. So its better to keep the regular extend op around as long as possible.
I had to qualify one place in DAG combine that created illegal vector extending load operations. This change by itself had no effect on any tests which is why its included here.
I've made a few cleanups to the custom lowering. The sign extend code no longer creates an identity shuffle with undef elements. The zero extend code now emits a zero_extend_vector_inreg instead of an unpckl with a zero vector.
For the high half of the custom lowering of zero_extend/any_extend, we're now using an unpckh with a zero vector or undef. Previously we used used a pshufd to move the upper 64-bits to the lower 64-bits and then used a zero_extend_vector_inreg. I think the zero vector should require less execution resources and be smaller code size.
Differential Revision: https://reviews.llvm.org/D54024
llvm-svn: 346043
This patch adds support for expanding vector CTPOP instructions and removes the x86 'bitmath' lowering which replicates the same expansion.
Differential Revision: https://reviews.llvm.org/D53258
llvm-svn: 345869
Reapplying an updated version of rL345395 (reverted in rL345451), now the issues noticed in PR39483 have been fixed.
This patch allows resolveTargetShuffleInputs to remove UNDEF inputs from cases where we have more than 2 inputs.
llvm-svn: 345824
The CONCAT_VECTORS case was using the original mask element count to determine how to adjust the broadcast index. But if we looked through a bitcast the original mask size doesn't tell us anything about the concat_vectors.
This patch switchs to using the concat_vectors input element count directly instead.
Differential Revision: https://reviews.llvm.org/D53823
llvm-svn: 345626
Summary: Previously if we had a bitcast vector output type that needs promotion and a vector input type that needs widening we would just do a stack store and load to handle the conversion. We can do a little better if we can widen the bitcast to a legal vector type the same size as the widened input type. Then we can do the bitcast between this widened type and the widened input type. Afterwards we can extract_subvector back to the original output and any_extend that. Type legalization will then circle back and handle promotion of the extract_subvector and the any_extend will just be removed. This will avoid going through the stack and allows us to remove a custom version of this legalization from X86.
Reviewers: efriedma, RKSimon
Reviewed By: efriedma
Subscribers: javed.absar, llvm-commits
Differential Revision: https://reviews.llvm.org/D53229
llvm-svn: 345567
Use SelectionDAG::EVTToAPFloatSemantics. Make the LogicVT calculation in LowerFABSorFNEG similar to LowerFCOPYSIGN. Use APInt::getSignedMaxValue instead of ~APInt::getSignMask.
llvm-svn: 345565
Add vector support to TargetLowering::expandFP_TO_UINT.
This exposes an issue in X86TargetLowering::LowerVSELECT which was assuming that the select mask was the same width as the LHS/RHS ops - as long as the result is a sign splat we can easily sext/trunk this.
llvm-svn: 345473
Makes no difference to actual shuffle decoding yet, but merges all the existing limits in one place for when proper support is fixed.
........
Its been reported that this is causing out of trunk failures.
llvm-svn: 345451
These promotions add additional bitcasts to the SelectionDAG that can pessimize computeKnownBits/computeNumSignBits. It also seems to interfere with broadcast formation.
This patch removes the promotion and adds isel patterns instead.
The increased table size is more than I would like, but hopefully we can find some canonicalizations or other tricks to start pruning out patterns going forward.
Differential Revision: https://reviews.llvm.org/D53268
llvm-svn: 345408
This is a narrow fix for 1 of the problems mentioned in PR27780:
https://bugs.llvm.org/show_bug.cgi?id=27780
I looked at more general solutions, but it's a mess. We canonicalize shuffle masks
based on the number of elements accessed from each operand, and that's not optional.
If you remove that, we'll crash because we fail to match isel patterns. So I'm
waiting until we're sure that we have blendvb with constant condition and then
commuting based on the load potential. Other cases like blend-with-immediate are
already handled elsewhere, so this is probably not a common problem anyway.
I didn't use "MayFoldLoad" because that checks for one-use and in these cases, we've
screwed that up by creating a temporary PSHUFB using these operands that we're counting
on to be killed later. Undoing that didn't look like a simple task because it's
intertwined with determining if we actually use both operands of the shuffle or not.a
Differential Revision: https://reviews.llvm.org/D53737
llvm-svn: 345390
Multiply a is complex operation so just because some bit of the output isn't used doesn't mean that bit of the input isn't used.
We might able to bound it, but it will require some more thought.
llvm-svn: 345241
Add X86 SimplifyDemandedBitsForTargetNode and use it to simplify PMULDQ/PMULUDQ target nodes.
This enables us to repeatedly simplify the node's arguments after the previous approach had to be reverted due to PR39398.
Differential Revision: https://reviews.llvm.org/D53643
llvm-svn: 345182
When implementing memset's today we often see this pattern:
$x0 = MOV 0xXYXYXYXYXYXYXYXY
store $x0, ...
$w1 = MOV 0xXYXYXYXY
store $w1, ...
We first create a 64bit constant in a 64bit register with all bytes the
same and then create a 32bit constant with all bytes the same in a 32bit
register. In many targets we could just access the lower byte of the
64bit register instead.
- Ideally this would be handled by the ConstantHoist pass but it runs
too early when memset isn't expanded yet.
- The memset expansion code already had this optimization implemented,
however SelectionDAG constantfolding would constantfold the
"trunc(bigconstnat)" pattern to "smallconstant".
- This patch makes the memset expansion mark the constant as Opaque and
stop DAGCombiner from constant folding in this situation. (Similar to
how ConstantHoisting marks things as Opaque to avoid folding
ADD/SUB/etc.)
Differential Revision: https://reviews.llvm.org/D53181
llvm-svn: 345102
We can't add the MULDQ node back to the worklist after the demanded bits change has been committed in case the node has been removed entirely. This will have to wait until we have SimplifyDemandedBitsForTargetNode.
llvm-svn: 345070
Matches the approach taken in the constant pool shuffle decoders, and uses an UndefElts mask instead of uint64_t(-1) raw mask values, which doesn't work safely for i32/i64 shuffle mask sizes (as the -1 value is legal).
This allows us to remove the constant pool shuffle decoders from most of the getTargetShuffleMask variable shuffle cases (X86ISD::VPERMV3 will be handled in a future commit).
llvm-svn: 345018
I've included a fix to DAGCombiner::ForwardStoreValueToDirectLoad that I believe will prevent the previous miscompile.
Original commit message:
Theoretically this was done to simplify the amount of isel patterns that were needed. But it also meant a substantial number of our isel patterns have to match an explicit bitcast. By making the vXi32/vXi16/vXi8 types legal for loads, DAG combiner should be able to change the load type to rem
I had to add some additional plain load instruction patterns and a few other special cases, but overall the isel table has reduced in size by ~12000 bytes. So it looks like this promotion was hurting us more than helping.
I still have one crash in vector-trunc.ll that I'm hoping @RKSimon can help with. It seems to relate to using getTargetConstantFromNode on a load that was shrunk due to an extract_subvector combine after the constant pool entry was created. So we end up decoding more mask elements than the lo
I'm hoping this patch will simplify the number of patterns needed to remove the and/or/xor promotion.
Reviewers: RKSimon, spatel
Reviewed By: RKSimon
Subscribers: llvm-commits, RKSimon
Differential Revision: https://reviews.llvm.org/D53306
llvm-svn: 344965
We can't safely assume that certain RawMask entries are UNDEF as most variable shuffles ignore non-index bits - PSHUFB only works on i8 elts so it'd be safe to use but I'm intending to come up with an alternative approach that works for all.
........
Enable this for PSHUFB constant mask decoding and remove the ConstantPool DecodePSHUFBMask
llvm-svn: 344937
We can't safely assume that certain RawMask entries are UNDEF as most variable shuffles ignore non-index bits.
........
Add support for UNDEF raw mask elements and remove the ConstantPool DecodeVPERMILPMask usage in X86ISelLowering.cpp
llvm-svn: 344936
Summary:
As discussed in D52304 / IRC, we now have pattern matching for
'bit extract' in two places - tablegen and `X86DAGToDAGISel`.
There are 4 patterns.
And we will have a problem with `x & (-1 >> (32 - y))` pattern.
* If the mask is one-use, then it is always unfolded into `x << (32 - y) >> (32 - y)` first.
Thus, the existing test coverage is already broken.
* If it is not one-use, then it is not unfolded, and is matched as BZHI.
* If it is not one-use, we will not match it as BEXTR. And if it is one-use, it will have been unfolded already.
So we will either not handle that pattern for BEXTR, or not have test coverage for it.
This is bad.
As discussed with @craig.topper, let's unify this matching, and do everything in `X86DAGToDAGISel`.
Then we will not have code duplication, and will have proper test coverage.
This indeed does not affect any tests, and this is great.
It means that for these two patterns, the `X86DAGToDAGISel` is identical to the tablegen version.
Please review carefully, i'm not fully sure about that intrinsic change, and introduction of the new `X86ISD` opcode.
Reviewers: craig.topper, RKSimon, spatel
Reviewed By: craig.topper
Subscribers: llvm-commits, craig.topper
Differential Revision: https://reviews.llvm.org/D53164
llvm-svn: 344904
Summary:
Theoretically this was done to simplify the amount of isel patterns that were needed. But it also meant a substantial number of our isel patterns have to match an explicit bitcast. By making the vXi32/vXi16/vXi8 types legal for loads, DAG combiner should be able to change the load type to remove the bitcast.
I had to add some additional plain load instruction patterns and a few other special cases, but overall the isel table has reduced in size by ~12000 bytes. So it looks like this promotion was hurting us more than helping.
I still have one crash in vector-trunc.ll that I'm hoping @RKSimon can help with. It seems to relate to using getTargetConstantFromNode on a load that was shrunk due to an extract_subvector combine after the constant pool entry was created. So we end up decoding more mask elements than the load size.
I'm hoping this patch will simplify the number of patterns needed to remove the and/or/xor promotion.
Reviewers: RKSimon, spatel
Reviewed By: RKSimon
Subscribers: llvm-commits, RKSimon
Differential Revision: https://reviews.llvm.org/D53306
llvm-svn: 344877
Summary:
These nodes exist to overcome an isel problem where we can generate a zero extend of an AH register followed by an extract subreg, and another zero extend. The first zero extend exists to avoid a partial register update copying the AH register into the low 8-bits. The second zero extend exists if the user wanted the remainder zero extended.
To make this work we had a DAG combine to morph the DIVREM opcode to a special opcode that included the extend. But then we had to add the new node to computeKnownBits and computeNumSignBits to process the extension portion.
This patch instead removes all of that and adds a late peephole to detect the two extends.
Reviewers: RKSimon, spatel
Reviewed By: RKSimon
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D53449
llvm-svn: 344874
D53306 exposes an issue where we sometimes use constant pool data from bigger vectors than the target shuffle mask. This should be safe to do, but we have to be certain that we're using the bottom most part of the vector as the shuffle mask decoders have no way to peek into subvectors with non-zero offsets.
llvm-svn: 344867
Summary:
I've noticed that the bitcasts we introduce for these make computeKnownBits and computeNumSignBits not work well in LegalizeVectorOps. LegalizeVectorOps legalizes bottom up while LegalizeDAG legalizes top down. The bottom up strategy for LegalizeVectorOps means operands are legalized before their uses. So we promote and/or/xor before we legalize the operands that use them making computeKnownBits/computeNumSignBits in places like LowerTruncate suboptimal. I looked at changing LegalizeVectorOps to be top down as well, but that was more disruptive and caused some regressions. I also looked at just moving promotion of binops to LegalizeDAG, but that had a few issues one around matching AND,ANDN,OR into VSELECT because I had to create ANDN as vXi64, but the other nodes hadn't legalized yet, I didn't look too hard at fixing that.
This patch seems to produce better results overall than my other attempts. We now form broadcasts of constants better in some cases. For at least some of them the AND was being introduced in LegalizeDAG, promoted to vXi64, and the BUILD_VECTOR was also legalized there. I think we got bad ordering of that. Now the promotion is out of the legalizer so we handle this better.
In the longer term I think we really should evaluate whether we should be doing this promotion at all. It's really there to reduce isel pattern count, but I'm wondering if we'd be better served just eating the pattern cost or doing C++ based isel for vector and/or/xor in X86ISelDAGToDAG. The masked and/or/xor will definitely be difficult in patterns if a bitcast gets between the vselect and the and/or/xor node. That becomes a lot of permutations to cover.
Reviewers: RKSimon, spatel
Reviewed By: RKSimon
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D53107
llvm-svn: 344487
Summary: This is similar to what D52528 did for loads. It should match what generic type legalization does in 64-bit mode where it uses a v2i64 cast and an i64 store.
Reviewers: RKSimon, spatel
Reviewed By: RKSimon
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D53173
llvm-svn: 344470
There is one remnant - AVX1 custom splitting of 256-bit vectors - which is due to a regression where the X86ISD::ANDNP is still performed as a YMM.
I've also tightened the CTLZ or CTPOP lowering in SelectionDAGLegalize::ExpandBitCount to require a legal CTLZ - it doesn't affect existing users and fixes an issue with AVX512 codegen.
llvm-svn: 344457
Use isConstantSplat instead of ISD::isConstantSplatVector to let us us peek through to illegal types (in this case for i686 targets to recognise i64 constants)
llvm-svn: 344452
If we have better CTLZ support than CTPOP, then use cttz(x) = width - ctlz(~x & (x - 1)) - and remove the CTTZ_ZERO_UNDEF handling as it no longer gives better codegen.
Similar to rL344447, this is also closer to LegalizeDAG's approach
llvm-svn: 344448
This patch changes the vector CTTZ lowering from:
cttz(x) = ctpop((x & -x) - 1)
to:
cttz(x) = ctpop(~x & (x - 1))
Not only does this make better use of the PANDN instruction, but it also matches the LegalizeDAG method which should allow us to remove the x86 specific code at some point in the future (we need to fix some issues with the bitcasted logic ops and CTPOP lowering first).
Differential Revision: https://reviews.llvm.org/D53214
llvm-svn: 344447
Add shuffle lowering for the case where we can shuffle the lanes into place followed by an in-lane permute.
This is mainly for cases where we can have non-repeating permutes in each lane, but for now I've just enabled it for v4f64 unary shuffles to fix PR39161 - there is no test coverage for other shuffles that might benefit yet.
We now have several cross-lane shuffle lowering methods that all do something similar - I've looked at merging some of these (notably by making the repeated mask mechanism in lowerVectorShuffleByMerging128BitLanes optional), but there is a lot of assertions/assumptions in the way that makes this tricky - I ended up going for adding yet another relatively simple method instead.
Differential Revision: https://reviews.llvm.org/D53148
llvm-svn: 344446
Generic legalization should be able to finish legalizing the EXTRACT_SUBVECTOR probably by turning it into a BUILD_VECTOR. But we should emit the simplest sequence.
llvm-svn: 344424
The algorithm we would do previously was identical to generic legalization. If we ever switch to legalizing integer vectors via widening we'll be able to kill off the code since it now only runs for promotion.
llvm-svn: 344423
This is the planned follow-up to D52997. Here we are reducing horizontal vector math codegen
by default. AMD Jaguar (btver2) should have no difference with this patch because it has
fast-hops. (If we want to set that bit for other CPUs, let me know.)
The code changes are small, but there are many test diffs. For files that are specifically
testing for hops, I added RUNs to distinguish fast/slow, so we can see the consequences
side-by-side. For files that are primarily concerned with codegen other than hops, I just
updated the CHECK lines to reflect the new default codegen.
To recap the recent horizontal op story:
1. Before rL343727, we were producing hops for all subtargets for a variety of patterns.
Hops were likely not optimal for all targets though.
2. The IR improvement in r343727 exposed a hole in the backend hop pattern matching, so
we reduced hop codegen for all subtargets. That was bad for Jaguar (PR39195).
3. We restored the hop codegen for all targets with rL344141. Good for Jaguar, but
probably bad for other CPUs.
4. This patch allows us to distinguish when we want to produce hops, so everyone can be
happy. I'm not sure if we have the best predicate here, but the intent is to undo the
extra hop-iness that was enabled by r344141.
Differential Revision: https://reviews.llvm.org/D53095
llvm-svn: 344361
Pull out repeated byte sum stage for popcount of vector elements > 8bits.
This allows us to simplify the LUT/BITMATH popcnt code to always assume vXi8 vectors, and also improves avx512bitalg codegen which only has access to vpopcntb/vpopcntw.
llvm-svn: 344348
Fixes PR32160 by reducing the size of PSHUFB if we only use one of the lanes.
This approach can probably be generalized to handle any target shuffle (and any subvector index) but we have no test coverage at the moment.
llvm-svn: 344336
On 64-bit targets the generic legalize will use an i64 load and a scalar_to_vector for us. But on 32-bit targets i64 isn't legal and the generic legalizer will end up emitting two 32-bit loads. We have DAG combines that try to put those two loads back together with pretty good success.
This patch instead uses f64 to avoid the splitting entirely. I've made it do the same for 64-bit mode for consistency and to keep the load in the fp domain.
There are a few things in here that look like regressions in 32-bit mode, but I believe they bring us closer to the 64-bit mode codegen. And that the 64-bit mode code could be better. I think those issues should be looked at separately.
Differential Revision: https://reviews.llvm.org/D52528
llvm-svn: 344291
This is an alternative to D53080 since I think using a BEXTR for a shifted mask is definitely an improvement when the shl can be absorbed into addressing mode. The other cases I'm less sure about.
We already have several tricks for handling an and of a shift in address matching. This adds a new case for BEXTR.
I've moved the BEXTR matching code back to X86ISelDAGToDAG to allow it to match. I suppose alternatively we could directly emit a X86ISD::BEXTR node that isel could pattern match. But I'm trying to view BEXTR matching as an isel concern so DAG combine can see 'and' and 'shift' operations that are well understood. We did lose a couple cases from tbm_patterns.ll, but I think there are ways to recover that.
I've also put back the manual load folding code in matchBEXTRFromAnd that I removed a few months ago in r324939. This gives us some more freedom to make decisions based on the ability to fold a load. I haven't done anything with that yet.
Differential Revision: https://reviews.llvm.org/D53126
llvm-svn: 344270
Summary:
As discussed in [[ https://bugs.llvm.org/show_bug.cgi?id=38938 | PR38938 ]],
we fail to emit `BEXTR` if the mask is shifted.
We can't deal with that in `X86DAGToDAGISel` `before the address mode for the inc is selected`,
and we can't really do it in the normal DAGCombine, because we don't have generic `ISD::BitFieldExtract` node,
and if we simply turn the shifted mask into a normal mask + shift-left, it will be folded back.
So it would seem X86ISelLowering is the place to handle this.
This patch only moves the matchBEXTRFromAnd()
from X86DAGToDAGISel to X86ISelLowering.
It does not add support for the 'shifted mask' pattern.
Reviewers: RKSimon, craig.topper, spatel
Reviewed By: RKSimon
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D52426
llvm-svn: 344179
This is intended to restore horizontal codegen to what it looked like before IR demanded elements improved in:
rL343727
As noted in PR39195:
https://bugs.llvm.org/show_bug.cgi?id=39195
...horizontal ops can be worse for performance than a shuffle+regular binop, so I've added a TODO. Ideally, we'd
solve that in a machine instruction pass, but a quicker solution will be adding a 'HasFastHorizontalOp' feature
bit to deal with it here in the DAG.
Differential Revision: https://reviews.llvm.org/D52997
llvm-svn: 344141
Similar to what already happens in the DAGCombiner wrappers, this patch adds the root nodes back onto the worklist if the DCI wrappers' SimplifyDemandedBits/SimplifyDemandedVectorElts were successful.
Differential Revision: https://reviews.llvm.org/D53026
llvm-svn: 344132
This may give slightly better opportunities for DAG combine to simplify with the operations before the setcc. It also matches the type the xors will eventually be promoted to anyway so it saves a legalization step.
Almost all of the test changes are because our constant pool entry is now v2i64 instead of v4i32 on 64-bit targets. On 32-bit targets getConstant should be emitting a v4i32 build_vector and a v4i32->v2i64 bitcast.
There are a couple test cases where it appears we now combine a bitwise not with one of these xors which caused a new constant vector to be generated. This prevented a constant pool entry from being shared. But if that's an issue we're concerned about, it seems we need to address it another way that just relying a bitcast to hide it.
This came about from experiments I've been trying with pushing the promotion of and/or/xor to vXi64 later than LegalizeVectorOps where it is today. We run LegalizeVectorOps in a bottom up order. So the and/or/xor are promoted before their users are legalized. The bitcasts added for the promotion act as a barrier to computeKnownBits if we try to use it during vector legalization of a later operation. So by moving the promotion out we can hopefully get better results from computeKnownBits/computeNumSignBits like in LowerTruncate on AVX512. I've also looked at running LegalizeVectorOps in a top down order like LegalizeDAG, but thats showing some other issues.
llvm-svn: 344071
As noted in D52747, if we prefer IR to use trunc for bool vectors rather
than and+icmp, we can expose codegen shortcomings as seen here with masked store.
Replace a hard-coded PCMPGT simplification with the more general demanded bits call
to improve things.
Differential Revision: https://reviews.llvm.org/D52964
llvm-svn: 344048
As discussed on D52964, this adds 256-bit *_EXTEND_VECTOR_INREG lowering support for AVX1 targets to help improve SimplifyDemandedBits handling.
Differential Revision: https://reviews.llvm.org/D52980
llvm-svn: 344019
Simple types are a superset of what all in tree targets in LLVM could possibly have a legal type. This means the behavior of using isSimple to check for a supported type for X86 could change over time. For example, this could would change if a v256i1 type was added to MVT in the future.
llvm-svn: 343995
Some necessary yak shaving before lowering *_EXTEND_VECTOR_INREG 256-bit vectors on AVX1 targets as suggested by D52964.
Differential Revision: https://reviews.llvm.org/D52970
llvm-svn: 343991
The instructions are complicated, so this code will
probably never be very obvious, but hopefully this
makes it better.
As shown in PR39195:
https://bugs.llvm.org/show_bug.cgi?id=39195
...we need to improve the matching to not miss cases
where we're h-opping on 1 source vector, and that
should be a small patch after this rearranging.
llvm-svn: 343989
Prevents missing other simplifications that may occur deep in the operand chain where CommitTargetLoweringOpt won't add the PMULDQ back to the worklist itself
llvm-svn: 343922
Attempt to simplify PSHUFB masks (even non-constant ones) - we should probably be able to simplify other variable shuffles as well as the need arises.
llvm-svn: 343919
rL343853 didn't limit the number of subinputs, but we don't currently support faux shuffles with more than 2 total inputs, so put a limiter in place until this is fixed.
Found by Artem Dergachev.
llvm-svn: 343891
The comments in this code say we were trying to avoid 16-bit immediates, but if the immediate fits in 8-bits this isn't an issue. This avoids creating a zero extend that probably won't go away.
The movmskb related changes are interesting. The movmskb instruction writes a 32-bit result, but fills the upper bits with 0. So the zero_extend we were previously emitting was free, but we turned a -1 immediate that would fit in 8-bits into a 32-bit immediate so it was still bad.
llvm-svn: 343871
Decode subvector shuffles from INSERT_SUBVECTOR(SRC0, SHUFFLE(EXTRACT_SUBVECTOR(SRC1))
This was found necessary while investigating PR39161
llvm-svn: 343853
Previously we replaced the chain use ourself and return the data result. LegalizeVectorOps then detected that we'd done this and assumed the chain had already been handled.
This commit instead returns a MERGE_VALUES node with two results joined from nodes. This allows LegalizeVectorOps to do all the replacements for us without any special casing. The MERGE_VALUES will be removed by DAG combine.
llvm-svn: 343817
The additional patterns needed for this aren't overwhelming and introducing extra bitcasts during lowering limits our ability to do computeNumSignBits. Not that I have a good example of that for select. I'm just becoming increasingly grumpy about promotion of AND/OR/XOR. SELECT was just a lot easier to fix.
llvm-svn: 343723
Fix use of SSE1 registers for f32 ops in no-x87 mode.
Notably, allow use of SSE instructions for f32 operations in 64-bit
mode (but not 32-bit which is disallowed by callign convention).
Also avoid translating memset/memcopy/memmove into SSE registers
without X87 for 32-bit mode.
This fixes PR38738.
Reviewers: nickdesaulniers, craig.topper
Subscribers: hiraditya, llvm-commits
Differential Revision: https://reviews.llvm.org/D52555
llvm-svn: 343689
By removing demanded target shuffles that simplify to zero/undef/identity before simplifying its inputs we improve chances of further simplification, as only the immediate parent user of the combined is added back to the work list - this still doesn't help us if its passed through other ops though (bitcasts....).
llvm-svn: 343390
The shift amount might have peeked through a extract_subvector, altering the number of vector elements in the 'Amt' variable - so we were incorrectly calculating the ratio when peeking through bitcasts, resulting in incorrectly detecting splats.
llvm-svn: 343373
Similar to the existing ISD::SRL constant vector shifts from D49562, this patch adds ISD::SRA support with ISD::MULHS.
As we're dealing with signed values, we have to handle shift by zero and shift by one special cases, so XOP+AVX2/AVX512 splitting/extension is still a better solution - really we should still use ISD::MULHS if one of the special cases are used but for now I've just left a TODO and filtered by isKnownNeverZero.
Differential Revision: https://reviews.llvm.org/D52171
llvm-svn: 343093
This removes an int->fp bitcast between the surrounding code and the movmsk. I had already added a hack to combineMOVMSK to try to look through this bitcast to improve the SimplifyDemandedBits there.
But I found an additional issue where the bitcast was preventing combineMOVMSK from being called again after earlier nodes in the DAG are optimized. The bitcast gets revisted, but not the user of the bitcast. By using integer types throughout, the bitcast doesn't get in the way.
llvm-svn: 343046
This is the final (I hope!) problem pattern mentioned in PR37749:
https://bugs.llvm.org/show_bug.cgi?id=37749
We are trying to avoid an AVX1 sinkhole caused by having 256-bit bitwise logic ops but no other 256-bit integer ops.
We've already solved the simple logic ops, but 'andn' is an x86 special. I looked at alternative solutions like
extending the generic DAG combine or trying to wait until the ANDNP node is created, but those are bigger patches
that can over-reach. Ie, splitting to 128-bit does not look like a win in most cases with >1 256-bit op.
The pattern matching is cluttered with bitcasts because of our i64 element canonicalization. For the affected test,
we have this vector-type-legalized sequence:
t29: v8i32 = concat_vectors t27, t28
t30: v4i64 = bitcast t29
t18: v8i32 = BUILD_VECTOR Constant:i32<-1>, Constant:i32<-1>, ...
t31: v4i64 = bitcast t18
t32: v4i64 = xor t30, t31
t9: v8i32 = BUILD_VECTOR Constant:i32<255>, Constant:i32<255>, ...
t34: v4i64 = bitcast t9
t35: v4i64 = and t32, t34
t36: v8i32 = bitcast t35
t37: v4i32 = extract_subvector t36, Constant:i64<0>
t38: v4i32 = extract_subvector t36, Constant:i64<4>
Differential Revision: https://reviews.llvm.org/D52318
llvm-svn: 343008
This is an alternative to https://reviews.llvm.org/D37896. We can't decompose
multiplies generically without a target hook to tell us when it's profitable.
ARM and AArch64 may be able to remove some existing code that overlaps with
this transform.
This extends D52195 and may resolve PR34474:
https://bugs.llvm.org/show_bug.cgi?id=34474
(still an open question about transforming legal vector multiplies, but we
could open another bug report for those)
llvm-svn: 342844
We don't have a vXi8 shift left so we need to bitcast to a vXi16 vector to perform the shift. If we let lowering legalize the vXi8 shift we get an extra and that we don't need and fail to remove.
llvm-svn: 342795
x86 had 2 versions of peekThroughBitcast. DAGCombiner had 1. Plus, it had a 1-off implementation for the one-use variant.
Move the x86 versions of the code to SelectionDAG, so we don't have different copies of the code.
No functional change intended.
I'm putting this next to isBitwiseNot() because I am planning to use it in there. Another option is next to the
helpers in the ISD namespace (eg, ISD::isConstantSplatVector()). But if there's no good reason for those to be
there, I'd prefer to pull other helpers over to SelectionDAG in follow-up steps.
Differential Revision: https://reviews.llvm.org/D52285
llvm-svn: 342669
As the code comments suggest, these are about splitting, and they
are not necessarily limited to lowering, so that misled me.
There's nothing that's actually x86-specific in these either, so
they might be better placed in a common header so any target can
use them.
llvm-svn: 342575
This patch adds an initial x86 SimplifyDemandedVectorEltsForTargetNode implementation to handle target shuffles.
Currently the patch only decodes a target shuffle, calls SimplifyDemandedVectorElts on its input operands and removes any shuffle that reduces to undef/zero/identity.
Future work will need to integrate this with combineX86ShufflesRecursively, add support for other x86 ops, etc.
NOTE: There is a minor regression that appears to be affecting further (extractelement?) combines which I haven't been able to solve yet - possibly something to do with how nodes are added to the worklist after simplification.
Differential Revision: https://reviews.llvm.org/D52140
llvm-svn: 342564
This is an alternative to D37896. I don't see a way to decompose multiplies
generically without a target hook to tell us when it's profitable.
ARM and AArch64 may be able to remove some duplicate code that overlaps with
this transform.
As a first step, we're only getting the most clear wins on the vector examples
requested in PR34474:
https://bugs.llvm.org/show_bug.cgi?id=34474
As noted in the code comment, it's likely that the x86 constraints are tighter
than necessary, but it may not always be a win to replace a pmullw/pmulld.
Differential Revision: https://reviews.llvm.org/D52195
llvm-svn: 342554
Summary:
The IR reference for the `byval` attribute states:
```
This indicates that the pointer parameter should really be passed by value
to the function. The attribute implies that a hidden copy of the pointee is
made between the caller and the callee, so the callee is unable to modify
the value in the caller. This attribute is only valid on LLVM pointer arguments.
```
However, on Win64, this attribute is unimplemented and the raw pointer is
passed to the callee instead. This is problematic, because frontend authors
relying on the implicit hidden copy (as happens for every other calling
convention) will see the passed value silently (if mutable memory) or
loudly (by means of a crash) modified because the callee treats the
location as scratch memory space it is allowed to mutate.
At this point, it's worth taking a step back to understand the context.
In most calling conventions, aggregates that are too large to be passed
in registers, instead get *copied* to the stack at a fixed (computable
from the signature) offset of the stack pointer. At the LLVM, we hide
this hidden copy behind the byval attribute. The caller passes a pointer
to the desired data and the callee receives a pointer, but these pointers
are not the same. In particular, the pointer that the callee receives
points to temporary stack memory allocated as part of the call lowering.
In most calling conventions, this pointer is never realized in registers
or memory. The temporary memory is simply defined by an implicit
offset from the stack pointer at function entry.
Win64, uniquely, works differently. The structure is still passed in
memory, but instead of being stored at an implicit memory offset, the
caller computes a pointer to the temporary memory and passes it to
the callee as a regular pointer (taking up a register, or if all
registers are taken up, an additional stack slot). Presumably, this
was done to allow eliding the copy when passing aggregates through
several functions on the stack.
This explains why ignoring the `byval` attribute mostly works on Win64.
The argument simply gets passed as a pointer and as long as we're ok
with the callee trampling all over that memory, there are no ill effects.
However, it does contradict the documentation of the `byval` attribute
which specifies that there is to be an implicit copy.
Frontends can of course work around this by never emitting the `byval`
attribute for Win64 and creating `alloca`s for the requisite temporary
stack slots (and that does appear to be what frontends are doing).
However, the presence of the `byval` attribute is not a trap for
frontend authors, since it seems to work, but silently modifies the
passed memory contrary to documentation.
I see two solutions:
- Disallow the `byval` attribute in the verifier if using the Win64
calling convention.
- Make it work by simply emitting a temporary stack copy as we would
with any other calling convention (frontends can of course always
not use the attribute if they want to elide the copy).
This patch implements the second option (make it work), though I would
be fine with the first also.
Ref: https://github.com/JuliaLang/julia/issues/28338
Reviewers: rnk
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D51842
llvm-svn: 342402
For constant non-uniform cases we'll never introduce more and/andn/or selects than already occur in generic pre-SSE41 ISD::SRL lowering.
llvm-svn: 342352
https://bugs.llvm.org/show_bug.cgi?id=38949
It's not clear to me that we even need a one-use check in this fold.
Ie, 2 independent loads might be better than a load+dependent shuffle.
Note that the existing re-use tests are not affected. We actually do form a
broadcast node in those tests now because there's no extra use of the
insert_subvector node in those cases. But something later in isel pattern
matching decides that it is not worth using a broadcast for the full load in
those tests:
Legalized selection DAG: %bb.0 'test_broadcast_2f64_4f64_reuse:'
t7: v2f64,ch = load<(load 16 from %ir.p0)> t0, t2, undef:i64
t4: i64,ch = CopyFromReg t0, Register:i64 %1
t10: ch = store<(store 16 into %ir.p1)> t7:1, t7, t4, undef:i64
t18: v4f64 = insert_subvector undef:v4f64, t7, Constant:i64<0>
t20: v4f64 = insert_subvector t18, t7, Constant:i64<2>
Becomes:
t7: v2f64,ch = load<(load 16 from %ir.p0)> t0, t2, undef:i64
t4: i64,ch = CopyFromReg t0, Register:i64 %1
t10: ch = store<(store 16 into %ir.p1)> t7:1, t7, t4, undef:i64
t21: v4f64 = X86ISD::SUBV_BROADCAST t7
ISEL: Starting selection on root node: t21: v4f64 = X86ISD::SUBV_BROADCAST t7
...
Created node: t27: v4f64 = INSERT_SUBREG IMPLICIT_DEF:v4f64, t7, TargetConstant:i32<7>
Morphed node: t21: v4f64 = VINSERTF128rr t27, t7, TargetConstant:i8<1>
llvm-svn: 342347
Summary: This unfortunately adds a move, but isn't that better than going to the int domain and back?
Reviewers: RKSimon
Reviewed By: RKSimon
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D52134
llvm-svn: 342327
Summary:
MOVMSK only care about the sign bit so we don't need the setcc to fill the whole element with 0s/1s. We can just shift the bit we're looking for into the sign bit. This saves a constant pool load.
Inspired by PR38840.
Reviewers: RKSimon, spatel
Reviewed By: RKSimon
Subscribers: lebedev.ri, llvm-commits
Differential Revision: https://reviews.llvm.org/D52121
llvm-svn: 342326
Attempt to lower a shuffle as an unpack of elements from two inputs followed by a single-input (wider) permutation.
As long as the permutation is wider this is a win - there may be some circumstances where same size permutations would also be useful but I've left that for future work.
Differential Revision: https://reviews.llvm.org/D52043
llvm-svn: 342257
When replacing a named register input to the appropriately sized
sub/super-register. In the case of a 64-bit value being assigned to a
register in 32-bit mode, match GCC's assignment.
Reviewers: eli.friedman, craig.topper
Subscribers: nickdesaulniers, llvm-commits, hiraditya
Differential Revision: https://reviews.llvm.org/D51502
llvm-svn: 342175
Summary:
Previously we type legalized v2i32 div/rem by promoting to v2i64. But we don't support div/rem of vectors so op legalization would then scalarize it using i64 scalar ops since it doesn't know about the original promotion. 64-bit scalar divides on Intel hardware are known to be slow and in 32-bit mode they require a libcall.
This patch switches type legalization to do the scalarizing itself using i32.
It looks like the division by power of 2 optimization is still kicking in and leaving the code as a vector. The division by other constant optimization doesn't kick in pre type legalization since it ignores illegal types. And previously, after type legalization we scalarized the v2i64 since we don't have v2i64 MULHS/MULHU support.
Another option might be to widen v2i32 to v4i32 so we could do division by constant optimizations, but we'd have to be careful to only do that for constant divisors or we risk scalaring to 4 scalar divides.
Reviewers: RKSimon, spatel
Reviewed By: spatel
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D51325
llvm-svn: 342114
MOVMSKPS and MOVMSKPD both take FP types, but likely the operations before it are on integer types with just a int->fp bitcast between them. If the bitcast isn't used by anything else and doesn't change the element width we can look through it to simplify the integer ops.
llvm-svn: 341915
I'm having a hard time finding a test case for this, but we should be consistent here. The fact that we canonicalize all zeros and all ones constants to vXi32 and all other constants to loads makes this hard to hit the easy DAG combine infinite loop we get for some of the other types.
llvm-svn: 341859
We have isel patterns for v4i32/v4f64 that artificially widen to v8i32/v8f64 so just use that.
If x86-experimental-vector-widening-legalization is enabled, we don't need any custom legalization and can just return. I've modified the test RUN lines to cover this case.
llvm-svn: 341765
Summary:
This patch allows vectors with a power of 2 number of elements and i8/i16 element type to select paddus/psubus instructions. ReplaceNodeResults has been updated to custom widen these operations up to 128 bits like we already do for PAVG.
Another step towards fixing PR38691
Reviewers: RKSimon, spatel
Reviewed By: RKSimon, spatel
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D51818
llvm-svn: 341753
The generic type legalizer will scalarize vXi1 instructions getting rid of the vector entirely. Creating wider vector instructions is just going to prevent that.
llvm-svn: 341705
The type legalizer will try to scalarize this and fail.
It looks like there's some other v1iX oddities out there too since we still generated some vector instructions.
llvm-svn: 341704
Similar to what was recently done for addcarry/subborrow and has been done for rdrand/rdseed for a while. It's better to use two results and an explicit store in IR when the store isn't part of the semantics of the instruction. This allows store->load forwarding to happen in the middle end. Or the store to be removed if its never loaded.
Differential Revision: https://reviews.llvm.org/D51803
llvm-svn: 341698
We should represent the store directly in IR instead. This gives the middle end a chance to remove it if it can see a load from the same address.
Differential Revision: https://reviews.llvm.org/D51769
llvm-svn: 341677
These intrinsics use the same implementation as PTEST intrinsics, but use vXi1 vectors.
New clang builtins will be accompanying them shortly.
llvm-svn: 341259
This patch recognizes shuffles that shift elements and fill with zeros. I've copied and modified the shift matching code we use for normal vector registers to do this. I'm not sure if there's a good way to share more of this code without making the existing function more complex than it already is.
This will be used to enable kshift intrinsics in clang.
Differential Revision: https://reviews.llvm.org/D51401
llvm-svn: 341227
This assert tried to check that AND constants are only on the RHS. But its possible for both operands to be constants if one is opaque which will prevent the AND from being constant folded.
Fixes PR38771
llvm-svn: 341102
Noticed while looking at D49562 codegen - we can avoid a large constant mask load and a slow VPBLENDVB select op by using VPBLENDW+VPBLENDD instead.
TODO: As discussed on the patch, we should investigate adding VPBLENDVB handling to target shuffle combining as well, that will allow us to extend this to VPBLENDW+VPBLENDW+VPBLENDD.
Differential Revision: https://reviews.llvm.org/D50074
llvm-svn: 340913
Simon Pilgrim