A loaded value with multiple users compared with 0 will become a load and
test single instruction. The load is not folded in this case (multiple
users), but the compare instruction is eliminated.
This patch returns 0 cost for the icmp in these cases.
Review: Ulrich Weigand
https://reviews.llvm.org/D55111
llvm-svn: 348141
Fix ScalarEvolution/solve-quadratic.ll test to account for __func__
output listing the complete function prototype rather than just its
name, as it does on NetBSD.
Example Linux output:
GetQuadraticEquation: addrec coeff bw: 4
GetQuadraticEquation: equation -2x^2 + -2x + -4, coeff bw: 5, multiplied by 2
Example NetBSD output:
llvm::Optional<std::tuple<llvm::APInt, llvm::APInt, llvm::APInt, llvm::APInt, unsigned int> > GetQuadraticEquation(const llvm::SCEVAddRecExpr*): addrec coeff bw: 4
llvm::Optional<std::tuple<llvm::APInt, llvm::APInt, llvm::APInt, llvm::APInt, unsigned int> > GetQuadraticEquation(const llvm::SCEVAddRecExpr*): equation -2x^2 + -2x + -4, coeff bw: 5, multiplied by 2
Differential Revision: https://reviews.llvm.org/D55162
llvm-svn: 348096
We were adding the entire scalarization extraction cost for reductions, which returns the total cost of extracting every element of a vector type.
For reductions we don't need to do this - we just need to extract the 0'th element after the reduction pattern has completed.
Fixes PR37731
Rebased and reapplied after being reverted in rL347541 due to PR39774 - which was fixed by D54955/rL347759 and D55017/rL347997
Differential Revision: https://reviews.llvm.org/D54585
llvm-svn: 348076
Summary:
This is patch #3 of the new DivergenceAnalysis
<https://lists.llvm.org/pipermail/llvm-dev/2018-May/123606.html>
The GPUDivergenceAnalysis is intended to eventually supersede the existing
LegacyDivergenceAnalysis. The existing LegacyDivergenceAnalysis produces
incorrect results on unstructured Control-Flow Graphs:
<https://bugs.llvm.org/show_bug.cgi?id=37185>
This patch adds the option -use-gpu-divergence-analysis to the
LegacyDivergenceAnalysis to turn it into a transparent wrapper for the
GPUDivergenceAnalysis.
Reviewers: nhaehnle
Reviewed By: nhaehnle
Subscribers: jholewinski, jvesely, jfb, llvm-commits, alex-t, sameerds, arsenm, nhaehnle
Differential Revision: https://reviews.llvm.org/D53493
llvm-svn: 348048
Three minor changes to these extra costs:
* For ICmp instructions, instead of adding 2 all the time for extending each
operand, this is only done if that operand is neither a load or an
immediate.
* The operands extension costs for divides removed, because we now use a high
cost already for the divide (20).
* The costs for lhsr/ashr extra costs removed as this did not seem useful.
Review: Ulrich Weigand
https://reviews.llvm.org/D55053
llvm-svn: 347961
Unlike most cost model functions this code makes a lot of table lookups without using the results from getTypeLegalizationCost. This means 512-bit vectors can be looked up even when the type isn't legal.
This patch adds a check around the two tables that contain 512-bit types to make sure that neither of the types would be split by type legalization. Meaning 512 bit types are illegal. I wanted to write this in a somewhat generic way that uses type legalization query hooks. But if prefered, I can switch to just using is512BitVector and the subtarget feature.
Differential Revision: https://reviews.llvm.org/D54984
llvm-svn: 347786
This fixes some of scalarization costs reported for sext/zext using avx512bw. This does not fix all scalarization costs being reported. Just the worst.
I've restricted this only to combinations of types that are legal with avx512bw like v32i1/v64i1/v32i16/v64i8 and conversions between vXi1 and vXi8/vXi16 with legal vXi8/vXi16 result types.
Differential Revision: https://reviews.llvm.org/D54979
llvm-svn: 347785
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
CGF/CLGF compares an i64 register with a sign/zero extended loaded i32 value
in memory.
This patch makes such a load considered foldable and so gets a 0 cost.
Review: Ulrich Weigand
https://reviews.llvm.org/D54944
llvm-svn: 347735
AH, SH and MH costs are already covered in the cases where LHS is 32 bits and
RHS is 16 bits of memory sign-extended to i32.
As these instructions are also used when LHS is i16, this patch recognizes
that the loads will get folded then as well.
Review: Ulrich Weigand
https://reviews.llvm.org/D54940
llvm-svn: 347734
Single instructions exist for i8 and i16 comparisons of memory against a
small immediate.
This patch makes sure that if the load in these cases has a single user (the
ICmp), it gets a 0 cost (folded), and also that the ICmp gets a cost of 1.
Review: Ulrich Weigand
https://reviews.llvm.org/D54897
llvm-svn: 347733
Since byte-swapping loads and stores are supported, a 'load -> bswap' or
'bswap -> store' sequence should have the cost of one.
Review: Ulrich Weigand
https://reviews.llvm.org/D54870
llvm-svn: 347732
The check lines marked AVX256 in the zext256/sext256 functions should be closer to the AVX values which would take into account a splitting cost.
llvm-svn: 347722
Our sext/zext cost modeling was somewhat incomplete. And had no coverage for the fact that avx512bw v32i16/v64i8 types return a scalarization cost.
Truncates are a whole different mess because isTruncateFree is returning true for vectors when it shouldn't and that's the fall back for anything not in the tables.
llvm-svn: 347719
Summary:
IPA is implemented as module pass which produce map from Function or Alias to
StackSafetyInfo for a single function.
From prototype by Evgenii Stepanov and Vlad Tsyrklevich.
Reviewers: eugenis, vlad.tsyrklevich, pcc, glider
Subscribers: hiraditya, mgrang, llvm-commits
Differential Revision: https://reviews.llvm.org/D54543
llvm-svn: 347611
Summary:
Analysis produces StackSafetyInfo which contains information with how allocas
and parameters were used in functions.
From prototype by Evgenii Stepanov and Vlad Tsyrklevich.
Reviewers: eugenis, vlad.tsyrklevich, pcc, glider
Subscribers: hiraditya, llvm-commits
Differential Revision: https://reviews.llvm.org/D54504
llvm-svn: 347603
Add support for funnel shifts to the DemandedBits analysis. The
demanded bits of the first two operands can be determined if the
shift amount is constant. The demanded bits of the third operand
(shift amount) can be determined if the bitwidth is a power of two.
This is basically the same functionality as implemented in D54869
and D54478, but for DemandedBits rather than InstCombine.
Differential Revision: https://reviews.llvm.org/D54876
llvm-svn: 347561
This reverts commit r346970.
It was causing PR39774, a crash in slp-vectorizer on a rather simple loop
with just a bunch of 'and's in the body.
llvm-svn: 347541
Implement getIntrinsicInstrCost() and return costs reflecting that bswap can
be done with a vperm per vector register.
Review: Ulrich Weigand
https://reviews.llvm.org/D54789
llvm-svn: 347445
LVI was symbolically executing binary operators only when the RHS was
constant, missing the case where we have a ConstantRange for the RHS,
but not an actual constant. Tested using check-all and by
bootstrapping. Compile time is not impacted measurably.
Differential Revision: https://reviews.llvm.org/D19859
llvm-svn: 347379
Support saturating add/sub in constant folding, based on the APInt methods introduced in D54332.
Patch by: @nikic (Nikita Popov)
Differential Revision: https://reviews.llvm.org/D54531
llvm-svn: 347328
Summary:
Currently, when vectorizing stores to uniform addresses, the only
instance we prevent vectorization is if there are multiple stores to the
same uniform address causing an unsafe dependency.
This patch teaches LAA to avoid vectorizing loops that have an unsafe
cross-iteration dependency between a load and a store to the same uniform address.
Fixes PR39653.
Reviewers: Ayal, efriedma
Subscribers: rkruppe, llvm-commits
Differential Revision: https://reviews.llvm.org/D54538
llvm-svn: 347220
We were adding the entire scalarization extraction cost for reductions, which returns the total cost of extracting every element of a vector type.
For reductions we don't need to do this - we just need to extract the 0'th element after the reduction pattern has completed.
Fixes PR37731
Differential Revision: https://reviews.llvm.org/D54585
llvm-svn: 346970
Add support for the expansion of funnelshift/rotates to getIntrinsicInstrCost.
This also required us to move the X86 fshl/fshr costs to the same place as the rotates to avoid expansion and get correct scalarization vs vectorization costs.
llvm-svn: 346854
When we repeat the 2 shifting operands then this is a bit rotation - annoyingly this has to be done in the other getIntrinsicInstrCost than most intrinsics as we need to check the operands are the same.
llvm-svn: 346688
Improve getCastInstrCost() by respecting the different types of Src and Dst
for vector integer <-> fp conversions.
This means that extracting from integer becomes more expensive (by the
extraction penalty), and the extraction from fp becomes cheaper (no longer
has a false extraction penalty).
Review: Ulrich Weigand
https://reviews.llvm.org/D54423
llvm-svn: 346663
Instead of defaulting to a cost = 1, expand to element extract/insert like we do for other shuffles.
This exposes an issue in LoopVectorize which could call SK_ExtractSubvector with a scalar subvector type.
llvm-svn: 346656
The patch has been reverted because it ended up prohibiting propagation
of a constant to exit value. For such values, we should skip all checks
related to hard uses because propagating a constant is always profitable.
Differential Revision: https://reviews.llvm.org/D53691
llvm-svn: 346397
This reverts commit 2f425e9c7946b9d74e64ebbfa33c1caa36914402.
It seems that the check that we still should do the transform if we
know the result is constant is missing in this code. So the logic that
has been deleted by this change is still sometimes accidentally useful.
I revert the change to see what can be done about it. The motivating
case is the following:
@Y = global [400 x i16] zeroinitializer, align 1
define i16 @foo() {
entry:
br label %for.body
for.body: ; preds = %entry, %for.body
%i = phi i16 [ 0, %entry ], [ %inc, %for.body ]
%arrayidx = getelementptr inbounds [400 x i16], [400 x i16]* @Y, i16 0, i16 %i
store i16 0, i16* %arrayidx, align 1
%inc = add nuw nsw i16 %i, 1
%cmp = icmp ult i16 %inc, 400
br i1 %cmp, label %for.body, label %for.end
for.end: ; preds = %for.body
%inc.lcssa = phi i16 [ %inc, %for.body ]
ret i16 %inc.lcssa
}
We should be able to figure out that the result is constant, but the patch
breaks it.
Differential Revision: https://reviews.llvm.org/D51584
llvm-svn: 346198
Let i8/i16 uint/sint to fp conversions cost 1 if operand is a load.
Since the load already does the extension, there is no extra cost (previously
returned 2).
Review: Ulrich Weigand
https://reviews.llvm.org/D54028
llvm-svn: 346009
Summary:
The hot and cold count thresholds are derived from the summary, but for
debugging purposes it is convenient to provide the actual thresholds.
Reviewers: davidxl
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D54040
llvm-svn: 346005
Scalar i1 to fp conversions are done with a branch sequence, so it should
have a higher cost.
Review: Ulrich Weigand
https://reviews.llvm.org/D53924
llvm-svn: 345818
This factors out a new method getBoolVecToIntConversionCost() containing the
code for vector sext/zext of i1, in order to reuse it for i1 to double vector
conversions.
Review: Ulrich Weigand
https://reviews.llvm.org/D53923
llvm-svn: 345817
When rewriting loop exit values, IndVars considers this transform not profitable if
the loop instruction has a loop user which it believes cannot be optimized away.
In current implementation only calls that immediately use the instruction are considered
as such.
This patch extends the definition of "hard" users to any side-effecting instructions
(which usually cannot be optimized away from the loop) and also allows handling
of not just immediate users, but use chains.
Differentlai Revision: https://reviews.llvm.org/D51584
Reviewed By: etherzhhb
llvm-svn: 345814
When we calculate a product of 2 AddRecs, we end up making quite massive
computations to deduce the operands of resulting AddRec. This process can
be optimized by computing all args of intermediate sum and then calling
`getAddExpr` once rather than calling `getAddExpr` with intermediate
result every time a new argument is computed.
Differential Revision: https://reviews.llvm.org/D53189
Reviewed By: rtereshin
llvm-svn: 345813
Correct costings of SK_ExtractSubvector requires the SubTy argument to indicate the type/size of the extracted subvector.
Unlike the rest of the shuffle kinds this means that the main Ty argument represents the source vector type not the destination!
I've done my best to fix a number of vectorizer uses:
SLP - the reduction epilogue costs should be using a SK_PermuteSingleSrc shuffle as these all occur at the hardware vector width - we're not extracting (illegal) subvector types. This is causing the cost model diffs as SK_ExtractSubvector costs are poorly handled and tend to just return 1 at the moment.
LV - I'm not clear on what the SK_ExtractSubvector should represents for recurrences - I've used a <1 x ?> subvector extraction as that seems to match the VF delta.
Differential Revision: https://reviews.llvm.org/D53573
llvm-svn: 345617
Sub, SDiv and UDiv are not commutative, so only the RHS operand can fold a
load. This patch adds a check for this.
Review: Ulrich Weigand
https://reviews.llvm.org/D53791
llvm-svn: 345596
The SystemZ backend can do arithmetic of memory by loading and then extending
one of the operands. Similarly, a load + truncate can be folded into an
operand.
This patch improves the SystemZ TTI cost function to recognize this.
Review: Ulrich Weigand
https://reviews.llvm.org/D52692
llvm-svn: 345327
Enable the DAG optimization that converts vector div/rem with constants into
multiply+shifts sequences by expanding them early. This is needed since
ISD::SMUL_LOHI is 'Custom' lowered on SystemZ, and will therefore not be
available to BuildSDIV after legalization.
Better cost values for these instructions based on how they will be
implemented (a constant divisor is cheaper).
Review: Ulrich Weigand
https://reviews.llvm.org/D53196
llvm-svn: 345321
Non-uniform division/remainder handling was added back at D49248/D50765 - so share the 'mul+sub' costs that already exist for uniform cases.
llvm-svn: 345164
Summary:
Teach vectorizer about vectorizing variant value stores to uniform
address. Similar to rL343028, we do not allow vectorization if we have
multiple stores to the same uniform address.
Cost model already has the change for considering the extract
instruction cost for a variant value store. See added test cases for how
vectorization is done.
The patch also contains changes to the ORE messages.
Reviewers: Ayal, mkuper, anemet, hsaito
Subscribers: rkruppe, llvm-commits
Differential Revision: https://reviews.llvm.org/D52656
llvm-svn: 344613
SCEV's transform that turns `{A1,+,A2,+,...,+,An}<L> * {B1,+,B2,+,...,+,Bn}<L>` into
a single AddRec of size `2n+1` with complex combinatorial coefficients can easily
trigger exponential growth of the SCEV (in case if nothing gets folded and simplified).
We tried to restrain this transform using the option `scalar-evolution-max-add-rec-size`,
but its default value seems to be insufficiently small: the test attached to this patch
with default value of this option `16` has a SCEV of >3M symbols (when printed out).
This patch reduces the simplification limit. It is not a cure to combinatorial
explosions, but at least it reduces this corner case to something more or less
reasonable.
Differential Revision: https://reviews.llvm.org/D53282
Reviewed By: sanjoy
llvm-svn: 344584
Until mischeduler is clever enough to avoid spilling in a vectorized loop
with many (scalar) DLRs it is better to avoid high vectorization factors (8
and above).
llvm-svn: 344129
A new function getNumVectorRegs() is better to use for the number of needed
vector registers instead of getNumberOfParts(). This is to make sure that the
number of vector registers (and typically operations) required for a vector
type is accurate.
getNumberOfParts() which was previously used works by splitting the vector
type until it is legal gives incorrect results for types with a non
power of two number of elements (rare).
A new static function getScalarSizeInBits() that also checks for a pointer
type and returns 64U for it since otherwise it gets a value of 0). Used in a
few places where Ty may be pointer.
Review: Ulrich Weigand
llvm-svn: 344115
This is the third patch in a series intended to make
https://reviews.llvm.org/D44748 more easily reviewable. Please see that
patch for more context. The second being r344013.
The intent is to make the output of printing a LocationSize more
precise. The main motivation for this is that we plan to add a bit to
distinguish whether a given LocationSize is an upper-bound or is
precise; making that information available in pretty-printing is nice.
llvm-svn: 344108
Summary:
We are overly conservative in loop vectorizer with respect to stores to loop
invariant addresses.
More details in https://bugs.llvm.org/show_bug.cgi?id=38546
This is the first part of the fix where we start with vectorizing loop invariant
values to loop invariant addresses.
This also includes changes to ORE for stores to invariant address.
Reviewers: anemet, Ayal, mkuper, mssimpso
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D50665
llvm-svn: 343028
After recent improvements which makes better use of LOC instead of IPM, the
TTI cost functions also needs to be updated to reflect this.
This involves sext, zext and xor of i1.
The tests were updated so that for z13 the new costs are expected, while the
old costs are still checked for on zEC12.
Review: Ulrich Weigand
https://reviews.llvm.org/D51339
llvm-svn: 342207
This patch does the following things:
1. update SymbolicallyEvaluateGEP so that it bails out if it cannot preserve inrange arribute;
2. update llvm/test/Analysis/ConstantFolding/gep.ll to remove UB in it;
3. remove inaccurate comment above ConstantFoldInstOperandsImpl in llvm/lib/Analysis/ConstantFolding.cpp;
4. add a new regression test that makes sure that no optimizations change an inrange GEP in an unexpected way.
Patch by Zhaomo Yang!
Differential Revision: https://reviews.llvm.org/D51698
llvm-svn: 341888
The only point to this change is the test diffs. When I remove this code entirely (in favor of the recently added generic handling), I don't want there to be any confusion due to spurious test diffs.
As an aside, the fact out tests are AST construction order dependent is not great. I thought about fixing that, but the reasonable schemes I might want (e.g. sort by name) need the test diffs anyways.
Philip
llvm-svn: 341841
There were two combines not covered by the check before now, neither of which
actually differed from normal in the benefit analysis.
The most recent seems to be because it was just added at the top of the
function (naturally). The older is from way back in 2008 (r46687) when we just
didn't put those checks in so routinely, and has been diligently maintained
since.
llvm-svn: 341831
AliasSetTracker has special case handling for memset, memcpy and memmove which pre-existed argmemonly on functions and readonly and writeonly on arguments. This patch generalizes it using the AA infrastructure to any call correctly annotated.
The motivation here is to cut down on confusion, not performance per se. For most instructions, there is a direct mapping to alias set. However, this is not guaranteed by the interface and was not in fact true for these three intrinsics *and only these three intrinsics*. I kept getting myself confused about this invariant, so I figured it would be good to clearly distinguish between a instructions and alias sets. Calls happened to be an easy target.
The nice side effect is that custom implementations of memset/memcpy/memmove - including wrappers discovered by IPO - can now be optimized the same as builts by LICM.
Note: The actual removal of the memset/memtransfer specific handling will happen in a follow on NFC patch. It was originally part of this one, but separate for ease of review and rebase.
Differential Revision: https://reviews.llvm.org/D50730
llvm-svn: 341713
Summary:
This is patch 1 of the new DivergenceAnalysis (https://reviews.llvm.org/D50433).
The purpose of this patch is to free up the name DivergenceAnalysis for the new generic
implementation. The generic implementation class will be shared by specialized
divergence analysis classes.
Patch by: Simon Moll
Reviewed By: nhaehnle
Subscribers: jvesely, jholewinski, arsenm, nhaehnle, mgorny, jfb, llvm-commits
Differential Revision: https://reviews.llvm.org/D50434
Change-Id: Ie8146b11be2c50d5312f30e11c7a3036a15b48cb
llvm-svn: 341071
This patch issues an error message if Darwin ABI is attempted with the PPC
backend. It also cleans up existing test cases, either converting the test to
use an alternative triple or removing the test if the coverage is no longer
needed.
Updated Tests
-------------
The majority of test cases were updated to use a different triple that does not
include the Darwin ABI. Many tests were also updated to use FileCheck, in place
of grep.
Deleted Tests
-------------
llvm/test/tools/dsymutil/PowerPC/sibling.test was originally added to test
specific functionality of dsymutil using an object file created with an old
version of llvm-gcc for a Powerbook G4. After a discussion with @JDevlieghere he
suggested removing the test.
llvm/test/CodeGen/PowerPC/combine_loads_from_build_pair.ll was converted from a
PPC test to a SystemZ test, as the behavior is also reproducible there.
All other tests that were deleted were specific to the darwin/ppc ABI and no
longer necessary.
Phabricator Review: https://reviews.llvm.org/D50988
llvm-svn: 340795
Summary: This was inheriting the cost from the AVX table, but should be legal under AVX512.
Reviewers: RKSimon
Reviewed By: RKSimon
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D51267
llvm-svn: 340708
The way that PhiValues is integrated with BasicAA it is possible for a pass
which uses BasicAA to pick up an instance of BasicAA that uses PhiValues without
intending to, and then delete values from a function in a way that causes
PhiValues to return dangling pointers to these deleted values. Fix this by
having a set of callback value handles to invalidate values when they're
deleted.
llvm-svn: 340613
These changes expand the FunctionAttr logic in order to mark functions as
WriteOnly when appropriate. This is done through an additional bool variable
and extended logic.
Reviewers: hfinkel, jdoerfert
Differential Revision: https://reviews.llvm.org/D48387
llvm-svn: 340537
In general we can't assume flat loads are uniform, and cases where we can prove
they are should be handled through infer-address-spaces.
Differential Revision: https://reviews.llvm.org/D50991
llvm-svn: 340343
Volatility is not an aliasing property. We used to model volatile as if it had extremely conservative aliasing implications, but that hasn't been true for several years now. So, it doesn't make sense to be in AliasSet.
It also turns out the code is entirely a noop. Outside of the AST code to update it, there was only one user: load store promotion in LICM. L/S promotion doesn't need the check since it walks all the users of the address anyway. It already checks each load or store via !isUnordered which causes us to bail for volatile accesses. (Look at the lines immediately following the two remove asserts.)
There is the possibility of some small compile time impact here, but the only case which will get noticeably slower is a loop with a large number of loads and stores to the same address where only the last one we inspect is volatile. This is sufficiently rare it's not worth optimizing for..
llvm-svn: 340312
This is another step towards being able to canonicalize to the funnel shift
intrinsics in IR (see D49242 for the initial patch).
We should not have any loss of simplification power in IR between these and
the equivalent IR constructs.
Differential Revision: https://reviews.llvm.org/D50848
llvm-svn: 340022
The description of `isGuaranteedToExecute` does not correspond to its implementation.
According to description, it should return `true` if an instruction is executed under the
assumption that its loop is *entered*. However there is a sophisticated alrogithm inside
that tries to prove that the instruction is executed if the loop is *exited*, which is not the
same thing for infinite loops. There is an attempt to protect from dealing with infinite loops
by prohibiting loops without exit blocks, however an infinite loop can have exit blocks.
As result of that, MustExecute can falsely consider some blocks that are never entered as
mustexec, and LICM can hoist dangerous instructions out of them basing on this fact.
This may introduce UB to programs which did not contain it initially.
This patch removes the problematic algorithm and replaced it with a one which tries to
prove what is required in description.
Differential Revision: https://reviews.llvm.org/D50558
Reviewed By: reames
llvm-svn: 339984
Summary:
Profile count of a block is computed by multiplying its block frequency
by entry count and dividing the result by entry block frequency. Do
rounded division in the last step and update test cases appropriately.
Reviewers: davidxl
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D50822
llvm-svn: 339835
The `experimental_guard` intrinsic has memory write semantics to model the thread-exiting
logic, but does not do any actual writes to memory. Currently, `AliasSetTracker` treats it as a
normal memory write. As result, a loop-invariant load cannot be hoisted out of loop because
the guard may possibly alias with it.
This patch makes `AliasSetTracker` so that it doesn't treat guards as memory writes.
Differential Revision: https://reviews.llvm.org/D50497
Reviewed By: reames
llvm-svn: 339753
Summary:
Calls marked 'tail' cannot read or write allocas from the current frame
because the current frame might be destroyed by the time they run.
However, a tail call may use an alloca with byval. Calling with byval
copies the contents of the alloca into argument registers or stack
slots, so there is no lifetime issue. Tail calls never modify allocas,
so we can return just ModRefInfo::Ref.
Fixes PR38466, a longstanding bug.
Reviewers: hfinkel, nlewycky, gbiv, george.burgess.iv
Subscribers: hiraditya, llvm-commits
Differential Revision: https://reviews.llvm.org/D50679
llvm-svn: 339636
By using PhiValuesAnalysis we can get all the values reachable from a phi, so
we can be more precise instead of giving up when a phi has phi operands. We
can't make BaseicAA directly use PhiValuesAnalysis though, as the user of
BasicAA may modify the function in ways that PhiValuesAnalysis can't cope with.
For this optional usage to work correctly BasicAAWrapperPass now needs to be not
marked as CFG-only (i.e. it is now invalidated even when CFG is preserved) due
to how the legacy pass manager handles dependent passes being invalidated,
namely the depending pass still has a pointer to the now-dead dependent pass.
Differential Revision: https://reviews.llvm.org/D44564
llvm-svn: 338242
Summary:
In non-integral address spaces, we're not allowed to introduce inttoptr/ptrtoint
intrinsics. Instead, we need to expand any pointer arithmetic as geps on the
base pointer. Luckily this is a common task for SCEV, so all we have to do here
is hook up the corresponding helper function and add test case.
Fixes PR38290
Reviewers: sanjoy
Differential Revision: https://reviews.llvm.org/D49832
llvm-svn: 338073
Currently ComputeNumSignBits does early exit while processing some
of the operations (add, sub, mul, and select). This prevents the
function from using AssumptionCacheTracker if passed.
Differential Revision: https://reviews.llvm.org/D49759
llvm-svn: 337936
if the top level addition in (D + (C-D + x + ...)) could be proven to
not wrap, where the choice of D also maximizes the number of trailing
zeroes of (C-D + x + ...), ensuring homogeneous behaviour of the
transformation and better canonicalization of such expressions.
This enables better canonicalization of expressions like
1 + zext(5 + 20 * %x + 24 * %y) and
zext(6 + 20 * %x + 24 * %y)
which get both transformed to
2 + zext(4 + 20 * %x + 24 * %y)
This pattern is common in address arithmetics and the transformation
makes it easier for passes like LoadStoreVectorizer to prove that 2 or
more memory accesses are consecutive and optimize (vectorize) them.
Reviewed By: mzolotukhin
Differential Revision: https://reviews.llvm.org/D48853
llvm-svn: 337859
SCEV tries to constant-fold arguments of trunc operands in SCEVAddExpr, and when it does
that, it passes wrong flags into the recursion. It is only valid to pass flags that are proved for
narrow type into a computation in wider type if we can prove that trunc instruction doesn't
actually change the value. If it did lose some meaningful bits, we may end up proving wrong
no-wrap flags for sum of arguments of trunc.
In the provided test we end up with `nuw` where it shouldn't be because of this bug.
The solution is to conservatively pass `SCEV::FlagAnyWrap` which is always a valid thing to do.
Reviewed By: sanjoy
Differential Revision: https://reviews.llvm.org/D49471
llvm-svn: 337435
Summary:
This commit does two things:
1. modified the existing DivergenceAnalysis::dump() so it dumps the
whole function with added DIVERGENT: annotations;
2. added code to do that dump if the appropriate -debug-only option is
on.
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D47700
Change-Id: Id97b605aab1fc6f5a11a20c58a99bbe8c565bf83
llvm-svn: 336998
Summary:
This patch is crucial for proving equality laundered/stripped
pointers. eg:
bool foo(A *a) {
return a == std::launder(a);
}
Clang with -fstrict-vtable-pointers will emit something like:
define dso_local zeroext i1 @_Z3fooP1A(%struct.A* %a) {
entry:
%c = bitcast %struct.A* %a to i8*
%call = tail call i8* @llvm.launder.invariant.group.p0i8(i8* %c)
%0 = bitcast %struct.A* %a to i8*
%1 = tail call i8* @llvm.strip.invariant.group.p0i8(i8* %0)
%2 = tail call i8* @llvm.strip.invariant.group.p0i8(i8* %call)
%cmp = icmp eq i8* %1, %2
ret i1 %cmp
}
and because %2 can be replaced with @llvm.strip.invariant.group(%0)
and that %2 and %1 will produce the same value (because strip is readnone)
we can replace compare with true.
Reviewers: rsmith, hfinkel, majnemer, amharc, kuhar
Subscribers: llvm-commits, hiraditya
Differential Revision: https://reviews.llvm.org/D47423
llvm-svn: 336963
Summary:
Support for this option is needed for building Linux kernel.
This is a very frequently requested feature by kernel developers.
More details : https://lkml.org/lkml/2018/4/4/601
GCC option description for -fdelete-null-pointer-checks:
This Assume that programs cannot safely dereference null pointers,
and that no code or data element resides at address zero.
-fno-delete-null-pointer-checks is the inverse of this implying that
null pointer dereferencing is not undefined.
This feature is implemented in LLVM IR in this CL as the function attribute
"null-pointer-is-valid"="true" in IR (Under review at D47894).
The CL updates several passes that assumed null pointer dereferencing is
undefined to not optimize when the "null-pointer-is-valid"="true"
attribute is present.
Reviewers: t.p.northover, efriedma, jyknight, chandlerc, rnk, srhines, void, george.burgess.iv
Reviewed By: efriedma, george.burgess.iv
Subscribers: eraman, haicheng, george.burgess.iv, drinkcat, theraven, reames, sanjoy, xbolva00, llvm-commits
Differential Revision: https://reviews.llvm.org/D47895
llvm-svn: 336613
We penalize general SDIV/UDIV costs but don't do the same for SREM/UREM.
This patch makes general vector SREM/UREM x20 as costly as scalar, the same approach as we do for SDIV/UDIV. The patch also extends the existing SDIV/UDIV constant costs for SREM/UREM - at the moment this means the additional cost of a MUL+SUB (see D48975).
Differential Revision: https://reviews.llvm.org/D48980
llvm-svn: 336486
This patch changes order of transform in InstCombineCompares to avoid
performing transforms based on ranges which produce complex bit arithmetics
before more simple things (like folding with constants) are done. See PR37636
for the motivating example.
Differential Revision: https://reviews.llvm.org/D48584
Reviewed By: spatel, lebedev.ri
llvm-svn: 336172
Summary:
Comment on Transforms/LoopVersioning/incorrect-phi.ll: With the change
SCEV is able to prove that the loop doesn't wrap-self (due to zext i16
to i64), disabling the entire loop versioning pass. Removed the zext and
just use i64.
Reviewers: sanjoy
Subscribers: jlebar, hiraditya, javed.absar, bixia, llvm-commits
Differential Revision: https://reviews.llvm.org/D48409
llvm-svn: 336140
Summary:
This patch introduce new intrinsic -
strip.invariant.group that was described in the
RFC: Devirtualization v2
Reviewers: rsmith, hfinkel, nlopes, sanjoy, amharc, kuhar
Subscribers: arsenm, nhaehnle, JDevlieghere, hiraditya, xbolva00, llvm-commits
Differential Revision: https://reviews.llvm.org/D47103
Co-authored-by: Krzysztof Pszeniczny <krzysztof.pszeniczny@gmail.com>
llvm-svn: 336073
We can have AddRec with loops having many predecessors.
This changes an assert to an early return.
Differential Revision: https://reviews.llvm.org/D48766
llvm-svn: 335965
This pass is being added in order to make the information available to BasicAA,
which can't do caching of this information itself, but possibly this information
may be useful for other passes.
Incorporates code based on Daniel Berlin's implementation of Tarjan's algorithm.
Differential Revision: https://reviews.llvm.org/D47893
llvm-svn: 335857
This patch adds a custom trunc store lowering for v4i8 vector types.
Since there is not v.4b register, the v4i8 is promoted to v4i16 (v.4h)
and default action for v4i8 is to extract each element and issue 4
byte stores.
A better strategy would be to extended the promoted v4i16 to v8i16
(with undef elements) and extract and store the word lane which
represents the v4i8 subvectores. The construction:
define void @foo(<4 x i16> %x, i8* nocapture %p) {
%0 = trunc <4 x i16> %x to <4 x i8>
%1 = bitcast i8* %p to <4 x i8>*
store <4 x i8> %0, <4 x i8>* %1, align 4, !tbaa !2
ret void
}
Can be optimized from:
umov w8, v0.h[3]
umov w9, v0.h[2]
umov w10, v0.h[1]
umov w11, v0.h[0]
strb w8, [x0, #3]
strb w9, [x0, #2]
strb w10, [x0, #1]
strb w11, [x0]
ret
To:
xtn v0.8b, v0.8h
str s0, [x0]
ret
The patch also adjust the memory cost for autovectorization, so the C
code:
void foo (const int *src, int width, unsigned char *dst)
{
for (int i = 0; i < width; i++)
*dst++ = *src++;
}
can be vectorized to:
.LBB0_4: // %vector.body
// =>This Inner Loop Header: Depth=1
ldr q0, [x0], #16
subs x12, x12, #4 // =4
xtn v0.4h, v0.4s
xtn v0.8b, v0.8h
st1 { v0.s }[0], [x2], #4
b.ne .LBB0_4
Instead of byte operations.
llvm-svn: 335735
We can prove that some delinearized subscripts do not wrap around to become
negative by the fact that they are from inbound geps of load/store locations.
This helps improve the delinearisation in cases where we can't prove that they
are non-negative from SCEV alone.
Differential Revision: https://reviews.llvm.org/D48481
llvm-svn: 335481
AArch64 was only setting costs for SK_Transpose, which meant that many of the simpler shuffles (e.g. SK_Select and SK_PermuteSingleSrc for larger vector elements) was being severely overestimated by the default shuffle expansion.
This patch adds costs to help improve SLP performance and avoid a regression in reductions introduced by D48174.
I'm not very knowledgeable about AArch64 shuffle lowering so I've kept the extra costs to a minimum - someone who knows this code can add extra costs which should improve vectorization a lot more.
Differential Revision: https://reviews.llvm.org/D48172
llvm-svn: 335329
Summary:
This initiates a discussion on changing Polly accordingly while re-applying r335197 (D48338).
I have never worked on Polly. The proposed change to param_div_div_div_2.ll is not educated, but just patterns that match the output.
All LLVM files are already reviewed in D48338.
Reviewers: jdoerfert, bollu, efriedma
Subscribers: jlebar, sanjoy, hiraditya, llvm-commits, bixia
Differential Revision: https://reviews.llvm.org/D48453
llvm-svn: 335292
Summary:
Also explicitly port over some tests in llvm.amdgcn.image.* that were
missing. Some tests are removed because they no longer apply (i.e.
explicitly testing building an address vector via insertelement).
This is in preparation for the eventual removal of the old-style
intrinsics.
Some additional notes:
- constant-address-space-32bit.ll: change some GCN-NEXT to GCN because
the instruction schedule was subtly altered
- insert_vector_elt.ll: the old test didn't actually test anything,
because %tmp1 was not used; remove the load, because it doesn't work
(Because of the amdgpu_ps calling convention? In any case, it's
orthogonal to what the test claims to be testing.)
Change-Id: Idfa99b6512ad139e755e82b8b89548ab08f0afcf
Reviewers: arsenm, rampitec
Subscribers: MatzeB, qcolombet, kzhuravl, wdng, yaxunl, dstuttard, tpr, t-tye, javed.absar, llvm-commits
Differential Revision: https://reviews.llvm.org/D48018
llvm-svn: 335229
This enables da-delinearize in Dependence Analysis for delinearizing array
accesses into multiple dimensions. This can help to increase the power of
Dependence analysis on multi-dimensional arrays and prevent having to fall
back to the slower and less accurate MIV tests. It adds static checks on the
bounds of the arrays to ensure that one dimension doesn't overflow into
another, and brings our code in line with our tests.
Differential Revision: https://reviews.llvm.org/D45872
llvm-svn: 335217
These were being over cautious for costs for one/two op general shuffles - VSHUFPD doesn't have to replicate the same shuffle in both lanes like VSHUFPS does.
llvm-svn: 335216
Summary:
Try to match udiv and urem patterns, and sink zext down to the leaves.
I'm not entirely sure why some unrelated tests change, but the added <nsw>s seem right.
Reviewers: sanjoy
Subscribers: jlebar, hiraditya, bixia, llvm-commits
Differential Revision: https://reviews.llvm.org/D48338
llvm-svn: 335197
Summary:
Related to https://bugs.llvm.org/show_bug.cgi?id=37793, https://reviews.llvm.org/D46760#1127287
We'd like to do this canonicalization https://rise4fun.com/Alive/Gmc
But it is currently restricted by rL155136 / rL155362, which says:
```
// This is a constant shift of a constant shift. Be careful about hiding
// shl instructions behind bit masks. They are used to represent multiplies
// by a constant, and it is important that simple arithmetic expressions
// are still recognizable by scalar evolution.
//
// The transforms applied to shl are very similar to the transforms applied
// to mul by constant. We can be more aggressive about optimizing right
// shifts.
//
// Combinations of right and left shifts will still be optimized in
// DAGCombine where scalar evolution no longer applies.
```
I think these tests show that for *constants*, SCEV has no issues with that canonicalization.
Reviewers: mkazantsev, spatel, efriedma, sanjoy
Reviewed By: mkazantsev
Subscribers: sanjoy, javed.absar, llvm-commits, stoklund, bixia
Differential Revision: https://reviews.llvm.org/D48229
llvm-svn: 335101
This reverts r334428. It incorrectly marks some multiplications as nuw. Tim
Shen is working on a proper fix.
Original commit message:
[SCEV] Add nuw/nsw to mul ops in StrengthenNoWrapFlags where safe.
Summary:
Previously we would add them for adds, but not multiplies.
llvm-svn: 335016
In particular, when asked to print a MemoryAccess, we'll now print where
defs are optimized to, and we'll print optimized access types.
This patch also introduces an operator<< to make printing AliasResults
easier.
Patch by Juneyoung Lee!
Differential Revision: https://reviews.llvm.org/D47860
llvm-svn: 334760
Summary:
Specifically, we transform
zext(2^K * (trunc X to iN)) to iM ->
2^K * (zext(trunc X to i{N-K}) to iM)<nuw>
This is helpful because pulling the 2^K out of the zext allows further
optimizations.
Reviewers: sanjoy
Subscribers: hiraditya, llvm-commits, timshen
Differential Revision: https://reviews.llvm.org/D48158
llvm-svn: 334737
Summary:
Previously we would do this simplification only if it did not introduce
any new truncs (excepting new truncs which replace other cast ops).
This change weakens this condition: If the number of truncs stays the
same, but we're able to transform trunc(X + Y) to X + trunc(Y), that's
still simpler, and it may open up additional transformations.
While we're here, also clean up some duplicated code.
Reviewers: sanjoy
Subscribers: hiraditya, llvm-commits
Differential Revision: https://reviews.llvm.org/D48160
llvm-svn: 334736
IndVarSimplify sometimes makes transforms basing on users that are trivially dead. In particular,
if DCE wasn't run before it, there may be a dead `sext/zext` in loop that will trigger widening
transforms, however it makes no sense to do it.
This patch teaches IndVarsSimplify ignore the mist trivial cases of that.
Differential Revision: https://reviews.llvm.org/D47974
Reviewed By: sanjoy
llvm-svn: 334567
As discussed on PR33744, this patch relaxes ShuffleKind::SK_Alternate which requires shuffle masks to only match an alternating pattern from its 2 sources:
e.g. v4f32: <0,5,2,7> or <4,1,6,3>
This seems far too restrictive as most SIMD hardware which will implement it using a general blend/bit-select instruction, so replaces it with SK_Select, permitting elements from either source as long as they are inline:
e.g. v4f32: <0,5,2,7>, <4,1,6,3>, <0,1,6,7>, <4,1,2,3> etc.
This initial patch just updates the name and cost model shuffle mask analysis, later patch reviews will update SLP to better utilise this - it still limits itself to SK_Alternate style patterns.
Differential Revision: https://reviews.llvm.org/D47985
llvm-svn: 334513
As discussed on D47985, identity shuffle masks should probably be free.
I've limited this to the case where the input and output types all match - but we could probably accept all cases.
Differential Revision: https://reviews.llvm.org/D47986
llvm-svn: 334506