Summary:
Extend overlapping store elision to handle overwrites of stores by
larger stores.
Nontemporal tests have been modified to add memory dependencies to
prevent store elision.
Reviewers: craig.topper, rnk, t.p.northover
Subscribers: javed.absar, hiraditya, llvm-commits
Differential Revision: https://reviews.llvm.org/D40969
llvm-svn: 321089
Search from AND nodes to find whether they can be propagated back to
loads, so that the AND and load can be combined into a narrow load.
We search through OR, XOR and other AND nodes and all bar one of the
leaves are required to be loads or constants. The exception node then
needs to be masked off meaning that the 'and' isn't removed, but the
loads(s) are narrowed still.
Differential Revision: https://reviews.llvm.org/D41177
llvm-svn: 320962
Recommitting rL319773, which was reverted due to a recursive issue
causing timeouts. This happened because I failed to check whether
the discovered loads could be narrowed further. In the case of a tree
with one or more narrow loads, that could not be further narrowed, as
well as a node that would need masking, an AND could be introduced
which could then be visited and recombined again with the same load.
This could again create the masking load, with would be combined
again... We now check that the load can be narrowed so that this
process stops.
Original commit message:
Search from AND nodes to find whether they can be propagated back to
loads, so that the AND and load can be combined into a narrow load.
We search through OR, XOR and other AND nodes and all bar one of the
leaves are required to be loads or constants. The exception node then
needs to be masked off meaning that the 'and' isn't removed, but the
loads(s) are narrowed still.
Differential Revision: https://reviews.llvm.org/D41177
llvm-svn: 320679
At first, I tried to thread the x86 needle and use a target hook (isVectorShiftByScalarCheap())
to disable the transform only for non-splat pow-of-2 constants, but not AVX2, but only some
element types, but...it's difficult.
Here we just avoid the loop with the x86 vector transform that conflicts with the general DAG
combine and preserve all of the existing behavior AFAICT otherwise.
Some tests that will probably fail if someone does try to restrict this in a more targeted way
for x86-only may be found in:
test/CodeGen/X86/combine-mul.ll
test/CodeGen/X86/vector-mul.ll
test/CodeGen/X86/widen_arith-5.ll
This should prevent the infinite looping seen with:
https://bugs.llvm.org/show_bug.cgi?id=35579
Differential Revision: https://reviews.llvm.org/D41040
llvm-svn: 320374
This commit is the first part of https://reviews.llvm.org/D40348.
In order to allow target combines to be performed on newly combined
indexed loads, add them back to the worklist. The remainder of the
above patch will be committed in subsequent revisions and will use
this. Test cases will be included with those follow-up commits.
llvm-svn: 320365
This is a preparatory step for D34515.
This change:
- makes nodes ISD::ADDCARRY and ISD::SUBCARRY legal for i32
- lowering is done by first converting the boolean value into the carry flag
using (_, C) ← (ARMISD::ADDC R, -1) and converted back to an integer value
using (R, _) ← (ARMISD::ADDE 0, 0, C). An ARMISD::ADDE between the two
operations does the actual addition.
- for subtraction, given that ISD::SUBCARRY second result is actually a
borrow, we need to invert the value of the second operand and result before
and after using ARMISD::SUBE. We need to invert the carry result of
ARMISD::SUBE to preserve the semantics.
- given that the generic combiner may lower ISD::ADDCARRY and
ISD::SUBCARRYinto ISD::UADDO and ISD::USUBO we need to update their lowering
as well otherwise i64 operations now would require branches. This implies
updating the corresponding test for unsigned.
- add new combiner to remove the redundant conversions from/to carry flags
to/from boolean values (ARMISD::ADDC (ARMISD::ADDE 0, 0, C), -1) → C
- fixes PR34045
- fixes PR34564
- fixes PR35103
Differential Revision: https://reviews.llvm.org/D35192
llvm-svn: 320355
I noticed this pattern in D38316 / D38388. We failed to combine a shuffle that is either
repeating a scalar insertion at the same position in a vector or translated to a different
element index.
Like the earlier patch, this could be an instcombine too, but since we opted to make this
a DAG transform earlier, I've made this one a DAG patch too.
We do not need any legality checking because the new insert is identical to the existing
insert except that it may have a different constant insertion operand.
The constant insertion test in test/CodeGen/X86/vector-shuffle-combining.ll was the
motivation for D38756.
Differential Revision: https://reviews.llvm.org/D40209
llvm-svn: 320050
Search from AND nodes to find whether they can be propagated back to
loads, so that the AND and load can be combined into a narrow load.
We search through OR, XOR and other AND nodes and all bar one of the
leaves are required to be loads or constants. The exception node then
needs to be masked off meaning that the 'and' isn't removed, but the
loads(s) are narrowed still.
Differential Revision: https://reviews.llvm.org/D39604
llvm-svn: 319773
Summary:
Found out, at code inspection, that there was a fault in
DAGCombiner::CombineConsecutiveLoads for big-endian targets.
A BUILD_PAIR is always having the least significant bits of
the composite value in element 0. So when we are doing the checks
for consecutive loads, for big endian targets, we should check
if the load to elt 1 is at the lower address and the load
to elt 0 is at the higher address.
Normally this bug only resulted in missed oppurtunities for
doing the load combine. I guess that in some rare situation it
could lead to faulty combines, but I've not seen that happen.
Note that this patch actually will trigger load combine for
some big endian regression tests.
One example is test/CodeGen/PowerPC/anon_aggr.ll where we now get
t76: i64,ch = load<LD8[FixedStack-9]
instead of
t37: i32,ch = load<LD4[FixedStack-10]>
t35: i32,ch = load<LD4[FixedStack-9]>
t41: i64 = build_pair t37, t35
before legalization. Then the legalization will split the LD8
into two loads, so the end result is the same. That should
verify that the transfomation is correct now.
Reviewers: niravd, hfinkel
Reviewed By: niravd
Subscribers: nemanjai, llvm-commits
Differential Revision: https://reviews.llvm.org/D40444
llvm-svn: 319771
Pull the checks upon the load out from ReduceLoadWidth into their own
function.
Differential Revision: https://reviews.llvm.org/D40833
llvm-svn: 319766
MatchRotate assumes the types of the types of LHS and RHS are equal,
which is always the case then they come from an OR node, but here
we're getting them from two different TRUNC nodes, so we have to check
the types.
llvm-svn: 319695
If the truncation has been pushed past the or-node, look through it and
truncate afterwards.
Differential revision: https://reviews.llvm.org/D40792
llvm-svn: 319692
Both LoadedVT and NarrowLoad are passed as references and neither
of them are used by any of its callers.
Differential Revision: https://reviews.llvm.org/D40713
llvm-svn: 319645
visitAND attempts to narrow the width of extending loads that are
then masked off. ReduceLoadWidth already exists for a similar purpose
and handles shifts, so I've moved the code to handle AND nodes there.
Differential Revision: https://reviews.llvm.org/D39595
llvm-svn: 319421
This is needed for cases when the memory access is not as big as the width of
the data type. For instance, storing i1 (1 bit) would be done in a byte (8
bits).
Using 'BitSize >> 3' (or '/ 8') would e.g. give the memory access of an i1 a
size of 0, which for instance makes alias analysis return NoAlias even when
it shouldn't.
There are no tests as this was done as a follow-up to the bugfix for the case
where this was discovered (r318824). This handles more similar cases.
Review: Björn Petterson
https://reviews.llvm.org/D40339
llvm-svn: 319173
Unoptimized IR can have linear sequences of stores to an array, where the
initial GEP for the first store is formed from the pointer to the array, and the
GEP for each store after the first is formed from the previous GEP with some
offset in an inductive fashion.
The (large) resulting DAG when analyzed by DAGCombine undergoes an excessive
number of combines as each store node is examined every time its' offset node
is combined with any child of the offset. One of the transformations is
findBetterNeighborChains which assists MergeConsecutiveStores. The former
relies on repeated chain walking to do its' work, however MergeConsecutiveStores
is disabled at O0 which makes the transformation redundant.
Any optimization level other than O0 would invoke InstCombine which would
resolve the chain of GEPs into flat base + offset GEP for each store which
does not exhibit the repeated examination of each store to the array.
Disabling this optimization fixes an excessive compile time issue (30~ minutes
for the test case provided) at O0.
Reviewers: niravd, craig.topper, t.p.northover
Differential Revision: https://reviews.llvm.org/D40193
llvm-svn: 319142
With AVX512 vXi1 types are legal so we shouldn't be extending them.
This change is similar to existing code in the zext(setcc) combine.
llvm-svn: 319120
Since i1 is a legal type, this:
NumBytes = Op1->getMemoryVT().getSizeInBits() >> 3;
is wrong and should be instead
NumBytes = Op0->getMemoryVT().getStoreSize();
There seems to be more places where this should be fixed outside DAGCombiner.
Review: Hal Finkel
https://bugs.llvm.org/show_bug.cgi?id=35366
llvm-svn: 318824
All these headers already depend on CodeGen headers so moving them into
CodeGen fixes the layering (since CodeGen depends on Target, not the
other way around).
llvm-svn: 318490
Change the calculation for the desired ValueType for non-sign
extending loads, as in those cases we don't care about the
higher bits. This creates a smaller ExtVT and allows for such
combinations as:
(srl (zextload i16, [addr]), 8) -> (zextload i8, [addr + 1])
Differential Revision: https://reviews.llvm.org/D40034
llvm-svn: 318390
In function DAGCombiner::visitSIGN_EXTEND_INREG, sext can be combined with extload even if sextload is not supported by target, then
if sext is the only user of extload, there is no big difference, no harm no benefit.
if extload has more than one user, the combined sextload may block extload from combining with other zext, causes extra zext instructions generated. As demonstrated by the attached test case.
This patch add the constraint that when sextload is not supported by target, sext can only be combined with extload if it is the only user of extload.
Differential Revision: https://reviews.llvm.org/D39108
llvm-svn: 316802
combineShuffleOfScalars is very conservative about shuffled BUILD_VECTORs that can be combined together.
This patch adds one additional case - if both BUILD_VECTORs represent splats of the same scalar value but with different UNDEF elements, then we should create a single splat BUILD_VECTOR, sharing only the UNDEF elements defined by the shuffle mask.
Differential Revision: https://reviews.llvm.org/D38696
llvm-svn: 316331
This is only currently used for mad/fma transforms.
This is the only case where it should be used for AMDGPU,
so add an opcode to be sure.
llvm-svn: 315740
Eg:
insert v4i32 V, (v2i16 X), 2 --> shuffle v8i16 V', X', {0,1,2,3,8,9,6,7}
This is a generalization of the IR fold in D38316 to handle insertion into a non-undef vector.
We may want to abandon that one if we can't find value in squashing the more specific pattern sooner.
We're using the existing legal shuffle target hook to avoid AVX512 horror with vXi1 shuffles.
There may be room for improvement in the shuffle lowering here, but that would be follow-up work.
Differential Revision: https://reviews.llvm.org/D38388
llvm-svn: 315460
Summary:
See https://llvm.org/PR33743 for more details
It seems that for non-power of 2 vector sizes, the algorithm can produce
non-matching sizes for input and result causing an assert.
This usually isn't a problem as the isAnyExtend check will weed these out, but
in some cases (most often with lots of undefined values for the mask indices) it
can pass this check for non power of 2 vectors.
Adding in an extra check that ensures that bit size will match for the result
and input (as required)
Subscribers: nhaehnle
Differential Revision: https://reviews.llvm.org/D35241
llvm-svn: 315307
Without this, we could end up trying to get the Nth (0-indexed) element
from a subvector of size N.
Differential Revision: https://reviews.llvm.org/D37880
llvm-svn: 314380
At least for the 64-bit and less case, we should be able to determine if we even have a mask without counting any bits. This also removes the need to explicitly check for 0 active bits, isMask will return false for 0.
llvm-svn: 313908
If we have an AssertZext of a truncated value that has already been AssertZext'ed,
we can assert on the wider source op to improve the zext-y knowledge:
assert (trunc (assert X, i8) to iN), i1 --> trunc (assert X, i1) to iN
This moves a fold from being Mips-specific to general combining, and x86 shows
improvements.
Differential Revision: https://reviews.llvm.org/D37017
llvm-svn: 313577
rL310710 allowed store merging to occur after legalization to catch stores that are created late,
but this exposes a logic hole seen in PR34217:
https://bugs.llvm.org/show_bug.cgi?id=34217
We will miss merging stores if the target lowers vector extracts into target-specific operations.
This patch allows store merging to occur both before and after legalization if the target chooses
to get maximum merging.
I don't think the potential regressions in the other tests are relevant. The tests are for
correctness of weird IR constructs rather than perf tests, and I think those are still correct.
Differential Revision: https://reviews.llvm.org/D37987
llvm-svn: 313564
We already have a combine for this pattern when the input to shl is add, so we just need to enable the transformation when the input is or.
Original patch by @tstellar
Differential Revision: https://reviews.llvm.org/D19325
llvm-svn: 313251
Fixes some combine issues for AMDGPU where we weren't
getting the many extract_vector_elt combines expected
in a future patch.
This should really be checking isOperationLegalOrCustom on
the extract. That improves a number of x86 lit tests, but
a few get stuck in an infinite loop from one place
where a similar looking extract is created. I have a
different workaround in the backend for that which
keeps many of those improvements, but also adds a few
regressions.
llvm-svn: 312730
The function combineShuffleToVectorExtend in DAGCombine might generate an illegal typed node after "legalize types" phase, causing assertion on non-simple type to fail afterwards.
Adding a type check in case the combine is running after the type legalize pass.
Differential Revision: https://reviews.llvm.org/D37330
llvm-svn: 312438
The loop dependence check looks for dependencies between store merge
candidates not captured by the chain sub-DAG doing a check of
predecessors which may be very large. Conservatively bound number of
nodes checked for compilation time. (Resolves PR34326).
Landing on behalf of Nirav Dave to unblock the 5.0.0 release.
Differential Revision: https://reviews.llvm.org/D37220
llvm-svn: 312022
Only do this before operations are legalized of BUILD_VECTOR is Legal for the target.
Differential Revision: https://reviews.llvm.org/D37186
llvm-svn: 311892
Summary:
If all the operands of a BUILD_VECTOR extract elements from same vector then split the
vector efficiently based on the maximum vector access index.
This will also fix PR 33784
Reviewers: zvi, delena, RKSimon, thakis
Reviewed By: RKSimon
Subscribers: chandlerc, eladcohen, llvm-commits
Differential Revision: https://reviews.llvm.org/D35788
llvm-svn: 311833
This goes back to a discussion about IR canonicalization. We'd like to preserve and convert
more IR to 'select' than we currently do because that's likely the best choice in IR:
http://lists.llvm.org/pipermail/llvm-dev/2016-September/105335.html
...but that's often not true for codegen, so we need to account for this pattern coming in
to the backend and transform it to better DAG ops.
Steps in this patch:
1. Add an EVT param to the existing convertSelectOfConstantsToMath() TLI hook to more finely
enable this transform. Other targets will probably want that anyway to distinguish scalars
from vectors. We're using that here to exclude AVX512 targets, but it may not be necessary.
2. Convert a vselect to ext+add. This eliminates a constant load/materialization, and the
vector ext is often free.
Implementing a more general fold using xor+and can be a follow-up for targets that don't have
a legal vselect. It's also possible that we can remove the TLI hook for the special case fold
implemented here because we're eliminating a constant, but it needs to be tested on other
targets.
Differential Revision: https://reviews.llvm.org/D36840
llvm-svn: 311731
When one operand is a user of another in a promoted binary operation
we may replace and delete the returned value before returning
triggering an assertion. Reorder node replacements to prevent this.
Fixes PR34137.
Landing on behalf of Nirav.
Differential Revision: https://reviews.llvm.org/D36581
llvm-svn: 311623
This partially reverts r311429 in favor of making ISD::isConstantSplatVector do something not confusing. Turns out the only other user of it was also having to deal with the weird property of it returning a smaller size.
So rather than continue to deal with this quirk everywhere, just make the interface do something sane.
Differential Revision: https://reviews.llvm.org/D37039
llvm-svn: 311510
Summary:
If all the operands of a BUILD_VECTOR extract elements from same vector then split the
vector efficiently based on the maximum vector access index.
Reviewers: zvi, delena, RKSimon, thakis
Reviewed By: RKSimon
Subscribers: chandlerc, eladcohen, llvm-commits
Differential Revision: https://reviews.llvm.org/D35788
llvm-svn: 311255
Summary: It is creating useless work as the commuted nodes is the same as the node we are working on in that case.
Reviewers: jyknight, nemanjai, mkuper, spatel, RKSimon, zvi, bkramer
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D33840
llvm-svn: 310832
Summary:
Without the SrcVT its hard to know what is really being asked for. For example if your target has 128, 256, and 512 bit vectors. Maybe extracting 128 from 256 is cheap, but maybe extracting 128 from 512 is not.
For x86 we do support extracting a quarter of a 512-bit register. But for i1 vectors we don't have isel patterns for extracting arbitrary pieces. So we need this to have a correct implementation of isExtractSubvectorCheap for mask vectors.
Reviewers: RKSimon, zvi, efriedma
Reviewed By: RKSimon
Subscribers: aemerson, javed.absar, kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D36649
llvm-svn: 310793
If all the operands of a BUILD_VECTOR extract elements from same vector then split the vector efficiently based on the maximum vector access index.
Reapplied with fix to only work with simple value types.
Committed on behalf of @jbhateja (Jatin Bhateja)
Differential Revision: https://reviews.llvm.org/D35788
llvm-svn: 310782
The previous rev (r310208) failed to account for overflow when subtracting the
constants to see if they're suitable for shift/lea. This version add a check
for that and more test were added in r310490.
We can convert any select-of-constants to math ops:
http://rise4fun.com/Alive/d7d
For this patch, I'm enhancing an existing x86 transform that uses fake multiplies
(they always become shl/lea) to avoid cmov or branching. The current code misses
cases where we have a negative constant and a positive constant, so this is just
trying to plug that hole.
The DAGCombiner diff prevents us from hitting a terrible inefficiency: we can start
with a select in IR, create a select DAG node, convert it into a sext, convert it
back into a select, and then lower it to sext machine code.
Some notes about the test diffs:
1. 2010-08-04-MaskedSignedCompare.ll - We were creating control flow that didn't exist in the IR.
2. memcmp.ll - Choose -1 or 1 is the case that got me looking at this again. We could avoid the
push/pop in some cases if we used 'movzbl %al' instead of an xor on a different reg? That's a
post-DAG problem though.
3. mul-constant-result.ll - The trade-off between sbb+not vs. setne+neg could be addressed if
that's a regression, but those would always be nearly equivalent.
4. pr22338.ll and sext-i1.ll - These tests have undef operands, so we don't actually care about these diffs.
5. sbb.ll - This shows a win for what is likely a common case: choose -1 or 0.
6. select.ll - There's another borderline case here: cmp+sbb+or vs. test+set+lea? Also, sbb+not vs. setae+neg shows up again.
7. select_const.ll - These are motivating cases for the enhancement; replace cmov with cheaper ops.
Assembly differences between movzbl and xor to avoid a partial reg stall are caused later by the X86 Fixup SetCC pass.
Differential Revision: https://reviews.llvm.org/D35340
llvm-svn: 310717
rL310372 enabled simplifyShuffleMask to support undef shuffle mask inputs, but its causing hangs.
Removing support until I can triage the problem
llvm-svn: 310699
Relanding after case to insert explicit truncation as necessary.
Allow SCALAR_TO_VECTOR of EXTRACT_VECTOR_ELT to reduce to
EXTRACT_SUBVECTOR of vector shuffle when output is smaller. Marginally
improves vector shuffle computations.
Reviewers: efriedma, RKSimon, spatel
Subscribers: javed.absar, llvm-commits
Differential Revision: https://reviews.llvm.org/D35566
llvm-svn: 310256
We can convert any select-of-constants to math ops:
http://rise4fun.com/Alive/d7d
For this patch, I'm enhancing an existing x86 transform that uses fake multiplies
(they always become shl/lea) to avoid cmov or branching. The current code misses
cases where we have a negative constant and a positive constant, so this is just
trying to plug that hole.
The DAGCombiner diff prevents us from hitting a terrible inefficiency: we can start
with a select in IR, create a select DAG node, convert it into a sext, convert it
back into a select, and then lower it to sext machine code.
Some notes about the test diffs:
1. 2010-08-04-MaskedSignedCompare.ll - We were creating control flow that didn't exist in the IR.
2. memcmp.ll - Choose -1 or 1 is the case that got me looking at this again. I
think we could avoid the push/pop in some cases if we used 'movzbl %al' instead of an xor on
a different reg? That's a post-DAG problem though.
3. mul-constant-result.ll - The trade-off between sbb+not vs. setne+neg could be addressed if
that's a regression, but I think those would always be nearly equivalent.
4. pr22338.ll and sext-i1.ll - These tests have undef operands, so I don't think we actually care about these diffs.
5. sbb.ll - This shows a win for what I think is a common case: choose -1 or 0.
6. select.ll - There's another borderline case here: cmp+sbb+or vs. test+set+lea? Also, sbb+not vs. setae+neg shows up again.
7. select_const.ll - These are motivating cases for the enhancement; replace cmov with cheaper ops.
Assembly differences between movzbl and xor to avoid a partial reg stall are caused later by the X86 Fixup SetCC pass.
Differential Revision: https://reviews.llvm.org/D35340
llvm-svn: 310208
If all the operands of a BUILD_VECTOR extract elements from same vector then split the vector efficiently based on the maximum vector access index.
Committed on behalf of @jbhateja (Jatin Bhateja)
Differential Revision: https://reviews.llvm.org/D35788
llvm-svn: 310058
During store merge we construct a sorted list of consecutive store
candidates and consider subsequences for merging into a single
store. For each subsequence we check if the stored value type is legal
the merged store would have valid and fast and if the constructed
value to be stored is valid. The only properties that affect this
check between subsequences is the size of the subsequence, the
alignment of the first store, the alignment of the stored load value
(when merging stores-of-loads), and whether the merged value is a
constant zero.
If we do not find a viable mergeable subsequence starting from the
first store of length N, we know that a subsequence starting at a
later store of length N will also fail unless the new store's
alignment, the new load's alignment (if we're merging store-of-loads),
or we've dropped stores of nonzero value and could construct a merged
stores of zero (for merging constants).
As a result if we fail to find a valid subsequence starting from the
first store we can safely skip considering subsequences that start
with subsequent stores unless one of the above properties is
true. This significantly (2x) improves compile time in some
pathological cases.
Reviewers: RKSimon, efriedma, zvi, spatel, waltl
Subscribers: grandinj, llvm-commits
Differential Revision: https://reviews.llvm.org/D35901
llvm-svn: 309830
This pattern shows up when lowering byval copies on AMDGPU.
The byval object access is split into 4-byte chunks, adding a
constant offset to the FixedStack base. When some of the offsets
turn into ors, this prevents combining the constant offsets.
This makes it not apparent that the object is there when matching
addressing modes, so it ends up using a scratch wave offset
relative access and the lengthy frame index expansion for that.
llvm-svn: 309775
This patch moves the DAGCombiner::GetDemandedBits function to SelectionDAG::GetDemandedBits as a first step towards making it easier for targets to get to the source of any demanded bits without the limitations of SimplifyDemandedBits.
Differential Revision: https://reviews.llvm.org/D35841
llvm-svn: 308983
Check the actual memory type stored and not the extended value size
when considering if truncated store merge is worthwhile.
Reviewers: efriedma, RKSimon, spatel, jyknight
Reviewed By: efriedma
Subscribers: llvm-commits, nhaehnle
Differential Revision: https://reviews.llvm.org/D35623
llvm-svn: 308833
Summary:
When pushing an extension of a constant bitwise operator on a load
into the load, change other uses of the load value if they exist to
prevent the old load from persisting.
Reviewers: spatel, RKSimon, efriedma
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D35030
llvm-svn: 308618
Most combines currently recognise scalar and splat-vector constants, but not non-uniform vector constants.
This patch introduces a matching mechanism that uses predicates to check against BUILD_VECTOR of ConstantSDNode, as well as scalar ConstantSDNode cases.
I've changed a couple of predicates to demonstrate - the combine-shl changes add currently unsupported cases, while the MatchRotate replaces an existing mechanism.
Differential Revision: https://reviews.llvm.org/D35492
llvm-svn: 308598
Re-recommiting after landing DAG extension-crash fix.
Recommiting after adding check to avoid miscomputing alias information
on addresses of the same base but different subindices.
Memory accesses offset from frame indices may alias, e.g., we
may merge write from function arguments passed on the stack when they
are contiguous. As a result, when checking aliasing, we consider the
underlying frame index's offset from the stack pointer.
Static allocs are realized as stack objects in SelectionDAG, but its
offset is not set until post-DAG causing DAGCombiner's alias check to
consider access to static allocas to frequently alias. Modify isAlias
to consider access between static allocas and access from other frame
objects to be considered aliasing.
Many test changes are included here. Most are fixes for tests which
indirectly relied on our aliasing ability and needed to be modified to
preserve their original intent.
The remaining tests have minor improvements due to relaxed
ordering. The exception is CodeGen/X86/2011-10-19-widen_vselect.ll
which has a minor degradation dispite though the pre-legalized DAG is
improved.
Reviewers: rnk, mkuper, jonpa, hfinkel, uweigand
Reviewed By: rnk
Subscribers: sdardis, nemanjai, javed.absar, llvm-commits
Differential Revision: https://reviews.llvm.org/D33345
llvm-svn: 308350
Reorder replacements to be user first in preparation for multi-level
folding to premptively avoid inadvertantly deleting later nodes from
sharing found from replacement.
llvm-svn: 308348
When replacing a node and it's operand, replacing the operand node may
cause the deletion of the original node leading to an assertion
failure. Case around these replacements to avoid this without relying
on inspecting the DELETED_NODE opcode in various extend
dagcombiner cases.
Fixes PR32515.
Reviewers: dbabokin, RKSimon, davide, chandlerc
Subscribers: chandlerc, llvm-commits
Differential Revision: https://reviews.llvm.org/D34095
llvm-svn: 308330
Recommiting after adding check to avoid miscomputing alias information
on addresses of the same base but different subindices.
Memory accesses offset from frame indices may alias, e.g., we
may merge write from function arguments passed on the stack when they
are contiguous. As a result, when checking aliasing, we consider the
underlying frame index's offset from the stack pointer.
Static allocs are realized as stack objects in SelectionDAG, but its
offset is not set until post-DAG causing DAGCombiner's alias check to
consider access to static allocas to frequently alias. Modify isAlias
to consider access between static allocas and access from other frame
objects to be considered aliasing.
Many test changes are included here. Most are fixes for tests which
indirectly relied on our aliasing ability and needed to be modified to
preserve their original intent.
The remaining tests have minor improvements due to relaxed
ordering. The exception is CodeGen/X86/2011-10-19-widen_vselect.ll
which has a minor degradation dispite though the pre-legalized DAG is
improved.
Reviewers: rnk, mkuper, jonpa, hfinkel, uweigand
Reviewed By: rnk
Subscribers: sdardis, nemanjai, javed.absar, llvm-commits
Differential Revision: https://reviews.llvm.org/D33345
llvm-svn: 308025
Reverting as it breaks tramp3d-v4 in the llvm test-suite. I added some
comments to https://reviews.llvm.org/D33345 about it.
This reverts commit r307546.
llvm-svn: 307589
Memory accesses offset from frame indices may alias, e.g., we
may merge write from function arguments passed on the stack when they
are contiguous. As a result, when checking aliasing, we consider the
underlying frame index's offset from the stack pointer.
Static allocs are realized as stack objects in SelectionDAG, but its
offset is not set until post-DAG causing DAGCombiner's alias check to
consider access to static allocas to frequently alias. Modify isAlias
to consider access between static allocas and access from other frame
objects to be considered aliasing.
Many test changes are included here. Most are fixes for tests which
indirectly relied on our aliasing ability and needed to be modified to
preserve their original intent.
The remaining tests have minor improvements due to relaxed
ordering. The exception is CodeGen/X86/2011-10-19-widen_vselect.ll
which has a minor degradation dispite though the pre-legalized DAG is
improved.
Reviewers: rnk, mkuper, jonpa, hfinkel, uweigand
Reviewed By: rnk
Subscribers: sdardis, nemanjai, javed.absar, llvm-commits
Differential Revision: https://reviews.llvm.org/D33345
llvm-svn: 307546
For two ROTR operations with shifts C1, C2; combined shift operand will be (C1 + C2) % bitsize.
Differential revision: https://reviews.llvm.org/D12833
llvm-svn: 307179
Summary:
Add a combine for creating a truncate to replace a build_vector composed of extracts with
indices that form a stride-2^N series.
Example:
v8i32 V = ...
v4i32 build_vector((extract_elt V, 0), (extract_elt V, 2), (extract_elt V, 4), (extract_elt V, 6))
-->
v4i32 truncate (bitcast V to v4i64)
Related discussion in llvm-dev about canonicalizing shuffles to
truncates in LLVM IR:
http://lists.llvm.org/pipermail/llvm-dev/2017-January/108936.html.
Reviewers: spatel, RKSimon, efriedma, igorb, craig.topper, wolfgangp, delena
Reviewed By: delena
Subscribers: guyblank, delena, javed.absar, llvm-commits
Differential Revision: https://reviews.llvm.org/D34077
llvm-svn: 307036
Relanding after restricting equalBaseIndex to not erroneuosly consider
a FrameIndices stemming from alloca from being comparable as its
offset is set post-selectionDAG.
Pull FrameIndex comparision reasoning from DAGCombiner::isAlias to
general BaseIndexOffset.
llvm-svn: 306688
Given no NaNs and no signed zeroes it folds:
(fmul X, (select (fcmp X > 0.0), -1.0, 1.0)) -> (fneg (fabs X))
(fmul X, (select (fcmp X > 0.0), 1.0, -1.0)) -> (fabs X)
Differential Revision: https://reviews.llvm.org/D34579
llvm-svn: 306592
That is pretty common for clang to produce code like
(shl %x, (and %amt, 31)). In this situation we can still perform
trunc (shl) into shl (trunc) conversion given the known value
range of shift amount.
Differential Revision: https://reviews.llvm.org/D34723
llvm-svn: 306499
When SelectionDAG merges consecutive stores and loads in MergeConsecutiveStores, it does not set dereferenceable flag for a created load instruction. This results in an assertion failure if SelectionDAG commonizes this load instruction with other load instructions, as well as it may miss optimization opportunities.
This patch sat dereferenceable flag for the newly created load instruction if all the load instructions to be merged are dereferenceable.
Differential Revision: https://reviews.llvm.org/D34679
llvm-svn: 306404
Move GlobalAddress Offset decomposition from initial match into
comparision check and removing the possibility of constructing a new
offseted global address when examining addresses.
llvm-svn: 305917
Add support for combining a build vector to a shuffle.
When the build vector is of extracted elements from 2 vectors (vec1, vec2) where vec2 is 2 times smaller than vec1.
llvm-svn: 305883
As all store merges checks are based on the memory operation
performed, allow use of truncated stores and extended loads as valid
input candidates for merging.
Relanding after fixing selection between truncated and normal store.
llvm-svn: 305701
In preparation for doing storemerge post-legalization, reorder
visitSTORE passes to move pre/post-index combining after store
merge. Reordered passes other than store merge are unaffected.
llvm-svn: 305473
As all store merges checks are based on the memory operation
performed, allow use of truncated stores and extended loads as valid
input candidates for merging.
llvm-svn: 305468
This step is just intended to reduce code duplication rather than change any functionality.
A follow-up would be to replace PPCTargetLowering::spliceIntoChain() usage with this new helper.
Differential Revision: https://reviews.llvm.org/D33649
llvm-svn: 305192
Summary: UADDO has 2 result, and one must check the result no before doing any kind of combine. Without it, the transform is invalid.
Reviewers: joerg
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D34088
llvm-svn: 305162
When considering merging stores values are the results of loads only
consider stores whose values come from loads from the same base.
This fixes much of the longer compile times in PR33330.
llvm-svn: 304934
Other calls to DAGCombiner::*PromoteOperand check the result, but here it could cause an assertion in getNode.
Falling back to any extend in this case instead of failing outright seems correct to me.
No test case because:
The failure was triggered by an out of tree backend. In order to trigger it, a backend would need to overload
TargetLowering::IsDesirableToPromoteOp to return true for a type for which ISD::SIGN_EXTEND_INREG is marked
illegal. In tree, only X86 overloads and sometimes returns true for MVT::i16 yet it marks
setOperationAction(ISD::SIGN_EXTEND_INREG, MVT::i16 , Legal);.
Patch by Jacob Young!
Differential Revision: https://reviews.llvm.org/D33633
llvm-svn: 304723
Summary:
This is a problem uncovered by stage2 testing. ADDCARRY end up being generated on target that do not support it.
The patch that introduced the problem has other patches layed on top of it, so we want to fix the issue rather than revert it to avoid creating a lor of churn.
A regression test will be added shortly, but this is committed as this in order to get the build back to green promptly.
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D33770
llvm-svn: 304409
Summary:
This is a continuation of the work started in D29872 . Passing the carry down as a value rather than as a glue allows for further optimizations. Introducing setcccarry makes the use of addc/subc unecessary and we can start the removal process.
This patch only introduce the optimization strictly required to get the same level of optimization as was available before nothing more.
Reviewers: jyknight, nemanjai, mkuper, spatel, RKSimon, zvi, bkramer
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D33374
llvm-svn: 304404
Summary: This pattern is no very useful per se, but it exposes optimization for toehr patterns that wouldn't kick in otherwize. It's very common and worth optimizing for.
Reviewers: jyknight, nemanjai, mkuper, spatel, RKSimon, zvi, bkramer
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D32756
llvm-svn: 304402
Correct references to alignment of store which may be deleted in a
previous iteration of merge. Instead use first store that would be
merged.
Corrects pr33172's use-after-poison caught by ASan.
Reviewers: spatel, hfinkel, RKSimon
Reviewed By: RKSimon
Subscribers: thegameg, javed.absar, llvm-commits
Differential Revision: https://reviews.llvm.org/D33686
llvm-svn: 304299
The extending load possibility was missed in:
https://reviews.llvm.org/rL304072
We might want to handle this cases as a follow-up, but bailing out for now
to avoid miscompiling.
llvm-svn: 304153
If we have (extract_subvector(load wide vector)) with no other users,
that can just be (load narrow vector). This is intentionally conservative.
Follow-ups may loosen the one-use constraint to account for the extract cost
or just remove the one-use check.
The memop chain updating is based on code that already exists multiple times
in x86 lowering, so that should be pulled into a helper function as a follow-up.
Background: this is a potential improvement noticed via regressions caused by
making x86's peekThroughBitcasts() not loop on consecutive bitcasts (see
comments in D33137).
Differential Revision: https://reviews.llvm.org/D33578
llvm-svn: 304072
In the best case:
extract (binop (concat X1, X2), (concat Y1, Y2)), N --> binop XN, YN
...we kill all of the extract/concat and just have narrow binops remaining.
If only one of the binop operands is amenable, this transform is still
worthwhile because we kill some of the extract/concat.
Optional bitcasting makes the code more complicated, but there doesn't
seem to be a way to avoid that.
The TODO about extending to more than bitwise logic is there because we really
will regress several x86 tests including madd, psad, and even a plain
integer-multiply-by-2 or shift-left-by-1. I don't think there's anything
fundamentally wrong with this patch that would cause those regressions; those
folds are just missing or brittle.
If we extend to more binops, I found that this patch will fire on at least one
non-x86 regression test. There's an ARM NEON test in
test/CodeGen/ARM/coalesce-subregs.ll with a pattern like:
t5: v2f32 = vector_shuffle<0,3> t2, t4
t6: v1i64 = bitcast t5
t8: v1i64 = BUILD_VECTOR Constant:i64<0>
t9: v2i64 = concat_vectors t6, t8
t10: v4f32 = bitcast t9
t12: v4f32 = fmul t11, t10
t13: v2i64 = bitcast t12
t16: v1i64 = extract_subvector t13, Constant:i32<0>
There was no functional change in the codegen from this transform from what I
could see though.
For the x86 test changes:
1. PR32790() is the closest call. We don't reduce the AVX1 instruction count in that case,
but we improve throughput. Also, on a core like Jaguar that double-pumps 256-bit ops,
there's an unseen win because two 128-bit ops have the same cost as the wider 256-bit op.
SSE/AVX2/AXV512 are not affected which is expected because only AVX1 has the extract/concat
ops to match the pattern.
2. do_not_use_256bit_op() is the best case. Everyone wins by avoiding the concat/extract.
Related bug for IR filed as: https://bugs.llvm.org/show_bug.cgi?id=33026
3. The SSE diffs in vector-trunc-math.ll are just scheduling/RA, so nothing real AFAICT.
4. The AVX1 diffs in vector-tzcnt-256.ll are all the same pattern: we reduced the instruction
count by one in each case by eliminating two insert/extract while adding one narrower logic op.
https://bugs.llvm.org/show_bug.cgi?id=32790
Differential Revision: https://reviews.llvm.org/D33137
llvm-svn: 303997
Summary:
While this makes some case better and some case worse - so it's unclear if it is a worthy combine just by itself - this is a useful canonicalisation.
As per discussion in D32756 .
Reviewers: jyknight, nemanjai, mkuper, spatel, RKSimon, zvi, bkramer
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D32916
llvm-svn: 303441
Summary:
In SelectionDAG, when a store is immediately chained to another store
to the same address, elide the first store as it has no observable
effects. This is causes small improvements dealing with intrinsics
lowered to stores.
Test notes:
* Many testcases overwrite store addresses multiple times and needed
minor changes, mainly making stores volatile to prevent the
optimization from optimizing the test away.
* Many X86 test cases optimized out instructions associated with
associated with va_start.
* Note that test_splat in CodeGen/AArch64/misched-stp.ll no longer has
dependencies to check and can probably be removed and potentially
replaced with another test.
Reviewers: rnk, john.brawn
Subscribers: aemerson, rengolin, qcolombet, jyknight, nemanjai, nhaehnle, javed.absar, llvm-commits
Differential Revision: https://reviews.llvm.org/D33206
llvm-svn: 303198
Summary:
Eli pointed out that it's unsafe to combine the shifts to ISD::SHL etc.,
because those are not defined for b > sizeof(a) * 8, even after some of
the combiners run.
However, PPCISD::SHL defines that behavior (as the instructions themselves).
Move the combination to the backend.
The tests in shift_mask.ll still pass.
Reviewers: echristo, hfinkel, efriedma, iteratee
Subscribers: nemanjai, llvm-commits
Differential Revision: https://reviews.llvm.org/D33076
llvm-svn: 302937
This reverts r302712.
The change fails with ASAN enabled:
ERROR: AddressSanitizer: use-after-poison on address ... at ...
READ of size 2 at ... thread T0
#0 ... in llvm::SDNode::getNumValues() const <snip>/include/llvm/CodeGen/SelectionDAGNodes.h:855:42
#1 ... in llvm::SDNode::hasAnyUseOfValue(unsigned int) const <snip>/lib/CodeGen/SelectionDAG/SelectionDAG.cpp:7270:3
#2 ... in llvm::SDValue::use_empty() const <snip> include/llvm/CodeGen/SelectionDAGNodes.h:1042:17
#3 ... in (anonymous namespace)::DAGCombiner::MergeConsecutiveStores(llvm::StoreSDNode*) <snip>/lib/CodeGen/SelectionDAG/DAGCombiner.cpp:12944:7
Reviewers: niravd
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D33081
llvm-svn: 302746
Summary:
Allow consecutive stores whose values come from consecutive loads to
merged in the presense of other uses of the loads. Previously this was
disallowed as in general the merged load cannot be shared with the
other uses. Merging N stores into 1 may cause as many as N redundant
loads. However in the context of caching this should have neglible
affect on memory pressure and reduce instruction count making it
almost always a win.
Fixes PR32086.
Reviewers: spatel, jyknight, andreadb, hfinkel, efriedma
Reviewed By: efriedma
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D30471
llvm-svn: 302712
Before r247167, the pass manager builder controlled which AA
implementations were used, exporting them all in the AliasAnalysis
analysis group.
Now, AAResultsWrapperPass always uses BasicAA, but still uses other AA
implementations if made available in the pass pipeline.
But regardless, SDAGISel is required at O0, and really doesn't need to
be doing fancy optimizations based on useful AA results.
Don't require AA at CodeGenOpt::None, and only use it otherwise.
This does have a functional impact (and one testcase is pessimized
because we can't reuse a load). But I think that's desirable no matter
what.
Note that this alone doesn't result in less DT computations: TwoAddress
was previously able to reuse the DT we computed for SDAG. That will be
fixed separately.
Differential Revision: https://reviews.llvm.org/D32766
llvm-svn: 302611
Remove an extra canonicalization step if ISD::ABS is going to be used anyway.
Updated x86 abs combine to check that we are lowering from both canonicalizations.
llvm-svn: 302337
Summary: Do the transform when the carry isn't used. It's a pattern exposed when legalizing large integers.
Reviewers: jyknight, nemanjai, mkuper, spatel, RKSimon, zvi, bkramer
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D32755
llvm-svn: 302047
Summary:
This is the corresponding llvm change to D28037 to ensure no performance
regression.
Reviewers: bogner, kbarton, hfinkel, iteratee, echristo
Subscribers: nemanjai, llvm-commits
Differential Revision: https://reviews.llvm.org/D28329
llvm-svn: 301990
Summary: This is a common pattern that arise when legalizing large integers operations. Only do it when Y + 1 cannot overflow as this would change the carry behavior of uaddo .
Reviewers: jyknight, nemanjai, mkuper, spatel, RKSimon, zvi, bkramer
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D32687
llvm-svn: 301922
Summary: Common pattern when legalizing large integers operations. Similar to D32687, when the carry isn't used.
Reviewers: jyknight, nemanjai, mkuper, spatel, RKSimon, zvi, bkramer
Differential Revision: https://reviews.llvm.org/D32738
llvm-svn: 301919
The existing code only looks at half of the tree when matching bswap + rol patterns ending in an OR tree (as opposed to a cascade).
Patch originally introduced by Jim Lewis.
Submitted on the behalf of Dinar Temirbulatov.
Differential Revision: https://reviews.llvm.org/D32039
llvm-svn: 301907
We discussed shrinking/widening of selects in IR in D26556, and I'll try to get back to that
patch eventually. But I'm hoping that this transform is less iffy in the DAG where we can check
legality of the select that we want to produce.
A few things to note:
1. We can't wait until after legalization and do this generically because (at least in the x86
tests from PR14657), we'll have PACKSS and bitcasts in the pattern.
2. This might benefit more of the SSE codegen if we lifted the legal-or-custom requirement, but
that requires a closer look to make sure we don't end up worse.
3. There's a 'vblendv' opportunity that we're missing that results in andn/and/or in some cases.
That should be fixed next.
4. I'm assuming that AVX1 offers the worst of all worlds wrt uneven ISA support with multiple
legal vector sizes, but if there are other targets like that, we should add more tests.
5. There's a codegen miracle in the multi-BB tests from PR14657 (the gcc auto-vectorization tests):
despite IR that is terrible for the target, this patch allows us to generate the optimal loop
code because something post-ISEL is hoisting the splat extends above the vector loops.
Differential Revision: https://reviews.llvm.org/D32620
llvm-svn: 301781
Summary: As per discution on how to get better codegen an large int legalization, it became clear that using a glue for the carry was preventing several desirable optimizations. Passing the carry down as a value allow for more flexibility.
Reviewers: jyknight, nemanjai, mkuper, spatel, RKSimon, zvi, bkramer
Subscribers: igorb, llvm-commits
Differential Revision: https://reviews.llvm.org/D29872
llvm-svn: 301775
This patch replaces the separate APInts for KnownZero/KnownOne with a single KnownBits struct. This is similar to what was done to ValueTracking's version recently.
This is largely a mechanical transformation from KnownZero to Known.Zero.
Differential Revision: https://reviews.llvm.org/D32569
llvm-svn: 301620
This patch uses various APInt methods to reduce the number of temporary APInts. These were all found while working through converting SelectionDAG's computeKnownBits to also use the KnownBits struct recently added to the ValueTracking version.
llvm-svn: 301618
Besides better codegen, the motivation is to be able to canonicalize this pattern
in IR (currently we don't) knowing that the backend is prepared for that.
This may also allow removing code for special constant cases in
DAGCombiner::foldSelectOfConstants() that was added in D30180.
Differential Revision: https://reviews.llvm.org/D31944
llvm-svn: 301457
While we use BaseIndexOffset in FindBetterNeighborChains to
appropriately realize they're almost the same address and should be
improved concurrently we do not use it in isAlias using the non-index
understanding FindBaseOffset instead. Adding a BaseIndexOffset check
in isAlias like should allow indexed stores to be merged.
FindBaseOffset to be excised in subsequent patch.
Reviewers: jyknight, aditya_nandakumar, bogner
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D31987
llvm-svn: 301187
This reverts commit r301105, 4, 3 and 1, as a follow up of the previous
revert, which broke even more bots.
For reference:
Revert "[APInt] Use operator<<= where possible. NFC"
Revert "[APInt] Use operator<<= instead of shl where possible. NFC"
Revert "[APInt] Use ashInPlace where possible."
PR32754.
llvm-svn: 301111
getSignBit is a static function that creates an APInt with only the sign bit set. getSignMask seems like a better name to convey its functionality. In fact several places use it and then store in an APInt named SignMask.
Differential Revision: https://reviews.llvm.org/D32108
llvm-svn: 300856
I've changed one of the tests to not fold away, but we didn't and still don't do the transform
that the comment claims we do (and I don't know why we'd want to do that).
Follow-up to:
https://reviews.llvm.org/rL300725https://reviews.llvm.org/rL300763
llvm-svn: 300772
This allows forming more 'not' ops, so we get improvements for ISAs that have and-not.
Follow-up to:
https://reviews.llvm.org/rL300725
llvm-svn: 300763
The patch itself is simple: stop discriminating against vectors in visitAnd() and again in
SimplifyDemandedBits().
Some notes for reference:
1. We're not consistent about calls to SimplifyDemandedBits in the various visitXXX functions.
Sometimes, we check if the RHS is a constant first. Other times (like here), we just dive in.
2. I'd like to break the vector shackles in steps for the sake of risk minimization, but we could
make similar simultaneous changes in other places if we think that would be better.
3. I don't know what the intent of the changed tests in this patch was supposed to be, but since
they wiggled in a positive way, I'm just going with that. :)
4. In the rotate tests, note that we can see through non-splat constants. This is a result of D24253.
5. My motivation for being here now is to make D31944 look better, so this is step 1 of N towards
improving the vector codegen in that patch without writing any actual new code.
Differential Revision: https://reviews.llvm.org/D32230
llvm-svn: 300725
This patch uses lshrInPlace to replace code where the object that lshr is called on is being overwritten with the result.
This adds an lshrInPlace(const APInt &) version as well.
Differential Revision: https://reviews.llvm.org/D32155
llvm-svn: 300566
Remove non-consecutive stores from store merge candidate search as
they cannot be merged and will prevent us from finding subsequent
mergeable store cases.
Reviewers: jyknight, bogner, javed.absar, spatel
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D32086
llvm-svn: 300561
This is a follow-on to r299096 which added support for fmadd.
Subtract does not have the case where with two multiply operands we commute in
order to fuse with the multiply with the fewer uses.
llvm-svn: 299572
This is a generic combine enabled via target hook to reduce icmp logic as discussed in:
https://bugs.llvm.org/show_bug.cgi?id=32401
It's likely that other targets will want to enable this hook for scalar transforms,
and there are probably other patterns that can use bitwise logic to reduce comparisons.
Note that we are missing an IR canonicalization for these patterns, and we will probably
prefer the pair-of-compares form in IR (shorter, more likely to fold).
Differential Revision: https://reviews.llvm.org/D31483
llvm-svn: 299542
This reverts commit r299047 which is incorrect because the
simplification may result in incorrect propogation of undefs to users of
the folded shuffle.
Thanks to Andrea Di Biagio for pointing this out.
llvm-svn: 299368
This moves the isMask and isShiftedMask functions to be class methods. They now use the MathExtras.h function for single word size and leading/trailing zeros/ones or countPopulation for the multiword size. The previous implementation made multiple temorary memory allocations to do the bitwise arithmetic operations to match the MathExtras.h implementation.
Differential Revision: https://reviews.llvm.org/D31565
llvm-svn: 299362
The code already allowed vector types in via "isInteger" (which might want
a more specific name), so use splat-friendly constant predicates to match
those types.
llvm-svn: 299304
This can only happen when we have a mix of zero and undef elements and the two vectors have a different arrangement of zeros/undefs. The shuffle should eventually be constant folded to all zeros.
Fixes PR32484.
llvm-svn: 299291
(and (setlt X, 0), (setlt Y, 0)) --> (setlt (and X, Y), 0)
We have 7 similar folds, but this one got away. The fact that the
x86 test with a branch didn't change is probably a separate bug. We
may also be missing this and the related folds in instcombine.
llvm-svn: 299252
Now alternatively to the TargetOption.AllowFPOpFusion global flag, FMUL->FADD
can also use the per operation FMF to allow fusion.
The idea here is not to port everything to the new scheme (e.g. fused
multiply-and-sub will be ported later) but that this work all the way from
clang.
The transformation is conditionalized on *both* the FADD and the FMUL having
the FMF contract flag.
Differential Revision: https://reviews.llvm.org/D31169
llvm-svn: 299096
Deal with case that initial node is deleted during dag-combine leading
to an assertional failure in promoteIntShiftOp.
Fixes PR32420.
Reviewers: spatel, RKSimon
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D31403
llvm-svn: 298931
Reorder work in PromoteIntBinOp to prevent stale (deleted) nodes from
being used.
Fixes PR32340 and PR32345.
Reviewers: hfinkel, dbabokin
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D31148
llvm-svn: 298923
Summary:
This class is a list of AttributeSetNodes corresponding the function
prototype of a call or function declaration. This class used to be
called ParamAttrListPtr, then AttrListPtr, then AttributeSet. It is
typically accessed by parameter and return value index, so
"AttributeList" seems like a more intuitive name.
Rename AttributeSetImpl to AttributeListImpl to follow suit.
It's useful to rename this class so that we can rename AttributeSetNode
to AttributeSet later. AttributeSet is the set of attributes that apply
to a single function, argument, or return value.
Reviewers: sanjoy, javed.absar, chandlerc, pete
Reviewed By: pete
Subscribers: pete, jholewinski, arsenm, dschuff, mehdi_amini, jfb, nhaehnle, sbc100, void, llvm-commits
Differential Revision: https://reviews.llvm.org/D31102
llvm-svn: 298393
Don't scalarize VSELECT->SETCC when operands/results needs to be widened,
or when the type of the SETCC operands are different from those of the VSELECT.
(VSELECT SETCC) and (VSELECT (AND/OR/XOR (SETCC,SETCC))) are handled.
The previous splitting of VSELECT->SETCC in DAGCombiner::visitVSELECT() is
no longer needed and has been removed.
Updated tests:
test/CodeGen/ARM/vuzp.ll
test/CodeGen/NVPTX/f16x2-instructions.ll
test/CodeGen/X86/2011-10-19-widen_vselect.ll
test/CodeGen/X86/2011-10-21-widen-cmp.ll
test/CodeGen/X86/psubus.ll
test/CodeGen/X86/vselect-pcmp.ll
Review: Eli Friedman, Simon Pilgrim
https://reviews.llvm.org/D29489
llvm-svn: 297930
Reduced version of D26357 - based on the discussion on llvm-dev about canonicalization of UMIN/UMAX/SMIN/SMAX as well as ABS I've reduced that patch to just the ABS ISD node (with x86/sse support) to improve basic combines and lowering.
ARM/AArch64, Hexagon, PowerPC and NVPTX all have similar instructions allowing us to make this a generic opcode and move away from the hard coded tablegen patterns which makes it tricky to match more complex patterns.
At the moment this patch doesn't attempt legalization as we only create an ABS node if its legal/custom.
Differential Revision: https://reviews.llvm.org/D29639
llvm-svn: 297780
Recommiting with compiler time improvements
Recommitting after fixup of 32-bit aliasing sign offset bug in DAGCombiner.
* Simplify Consecutive Merge Store Candidate Search
Now that address aliasing is much less conservative, push through
simplified store merging search and chain alias analysis which only
checks for parallel stores through the chain subgraph. This is cleaner
as the separation of non-interfering loads/stores from the
store-merging logic.
When merging stores search up the chain through a single load, and
finds all possible stores by looking down from through a load and a
TokenFactor to all stores visited.
This improves the quality of the output SelectionDAG and the output
Codegen (save perhaps for some ARM cases where we correctly constructs
wider loads, but then promotes them to float operations which appear
but requires more expensive constant generation).
Some minor peephole optimizations to deal with improved SubDAG shapes (listed below)
Additional Minor Changes:
1. Finishes removing unused AliasLoad code
2. Unifies the chain aggregation in the merged stores across code
paths
3. Re-add the Store node to the worklist after calling
SimplifyDemandedBits.
4. Increase GatherAllAliasesMaxDepth from 6 to 18. That number is
arbitrary, but seems sufficient to not cause regressions in
tests.
5. Remove Chain dependencies of Memory operations on CopyfromReg
nodes as these are captured by data dependence
6. Forward loads-store values through tokenfactors containing
{CopyToReg,CopyFromReg} Values.
7. Peephole to convert buildvector of extract_vector_elt to
extract_subvector if possible (see
CodeGen/AArch64/store-merge.ll)
8. Store merging for the ARM target is restricted to 32-bit as
some in some contexts invalid 64-bit operations are being
generated. This can be removed once appropriate checks are
added.
This finishes the change Matt Arsenault started in r246307 and
jyknight's original patch.
Many tests required some changes as memory operations are now
reorderable, improving load-store forwarding. One test in
particular is worth noting:
CodeGen/PowerPC/ppc64-align-long-double.ll - Improved load-store
forwarding converts a load-store pair into a parallel store and
a memory-realized bitcast of the same value. However, because we
lose the sharing of the explicit and implicit store values we
must create another local store. A similar transformation
happens before SelectionDAG as well.
Reviewers: arsenm, hfinkel, tstellarAMD, jyknight, nhaehnle
llvm-svn: 297695
Summary: This essentially does the same transform as for ADC.
Reviewers: jyknight, nemanjai, mkuper, spatel, RKSimon, zvi, bkramer
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D30417
llvm-svn: 297416
Summary: This essentially does the same transform as for SUBC.
Reviewers: jyknight, nemanjai, mkuper, spatel, RKSimon, zvi, bkramer
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D30437
llvm-svn: 297404
As discussed in the review thread for rL297026, this is actually 2 changes that
would independently fix all of the test cases in the patch:
1. Return undef in FoldConstantArithmetic for div/rem by 0.
2. Move basic undef simplifications for div/rem (simplifyDivRem()) before
foldBinopIntoSelect() as a matter of efficiency.
I will handle the case of vectors with any zero element as a follow-up. That change
is the DAG sibling for D30665 + adding a check of vector elements to FoldConstantVectorArithmetic().
I'm deleting the test for PR30693 because it does not test for the actual bug any more
(dangers of using bugpoint).
Differential Revision:
https://reviews.llvm.org/D30741
llvm-svn: 297384
This helps in cases involving bitfields where an AND is exposed by
legalization.
Differential Revision: https://reviews.llvm.org/D30472
llvm-svn: 297249
This is known incomplete and not called in the right order relative to
other folds, but that's the current behavior. I'm just trying to clean
this up before making actual functional changes to make the patch smaller.
The logic here should mimic the IR equivalents that are in InstSimplify's
simplifyDivRem().
llvm-svn: 297086
Refactoring of duplicated code and more fixes to follow.
This is motivated by the post-commit comments for r296699:
http://lists.llvm.org/pipermail/llvm-commits/Week-of-Mon-20170306/435182.html
Ie, we can crash if we're missing obvious simplifications like this that
exist in the IR simplifier or if these occur later than expected.
The x86 change for non-splat division shows a potential opportunity to improve
vector codegen: we assumed that since only one lane had meaningful results, we
should do the math in scalar. But that means moving back and forth from vector
registers.
llvm-svn: 297026
select Cond, C +/- 1, C --> add(ext Cond), C -- with a target hook.
This is part of the ongoing process to obsolete D24480. The motivation is to
canonicalize to select IR in InstCombine whenever possible, so we need to have a way to
undo that easily in codegen.
PowerPC is an obvious winner for this kind of transform because it has fast and complete
bit-twiddling abilities but generally lousy conditional execution perf (although this might
have changed in recent implementations).
x86 also sees some wins, but the effect is limited because these transforms already mostly
exist in its target-specific combineSelectOfTwoConstants(). The fact that we see any x86
changes just shows that that code is a mess of special-case holes. We may be able to remove
some of that logic now.
My guess is that other targets will want to enable this hook for most cases. The likely
follow-ups would be to add value type and/or the constants themselves as parameters for the
hook. As the tests in select_const.ll show, we can transform any select-of-constants to
math/logic, but the general transform for any 2 constants needs one more instruction
(multiply or 'and').
ARM is one target that I think may not want this for most cases. I see infinite loops there
because it wants to use selects to enable conditionally executed instructions.
Differential Revision: https://reviews.llvm.org/D30537
llvm-svn: 296977
This patch causes compile times for some patterns to explode. I have
a (large, unreduced) test case that slows down by more than 20x and
several test cases slow down by 2x. I'm sending some of the test cases
directly to Nirav and following up with more details in the review log,
but this should unblock anyone else hitting this.
llvm-svn: 296862
Summary:
Currently, when 't1: i1 = setcc t2, t3, cc' followed by 't4: i1 = xor t1, Constant:i1<-1>' is folded into 't5: i1 = setcc t2, t3 !cc', SDLoc of newly created SDValue 't5' follows SDLoc of 't4', not 't1'. However, as the opcode of newly created SDValue is 'setcc', it make more sense to take DebugLoc from 't1' than 't4'. For the code below
```
extern int bar();
extern int baz();
int foo(int x, int y) {
if (x != y)
return bar();
else
return baz();
}
```
, following is the bitcode representation of 'foo' at the end of llvm-ir level optimization:
```
define i32 @foo(i32 %x, i32 %y) !dbg !4 {
entry:
tail call void @llvm.dbg.value(metadata i32 %x, i64 0, metadata !9, metadata !11), !dbg !12
tail call void @llvm.dbg.value(metadata i32 %y, i64 0, metadata !10, metadata !11), !dbg !13
%cmp = icmp ne i32 %x, %y, !dbg !14
br i1 %cmp, label %if.then, label %if.else, !dbg !16
if.then: ; preds = %entry
%call = tail call i32 (...) @bar() #3, !dbg !17
br label %return, !dbg !18
if.else: ; preds = %entry
%call1 = tail call i32 (...) @baz() #3, !dbg !19
br label %return, !dbg !20
return: ; preds = %if.else, %if.then
%retval.0 = phi i32 [ %call, %if.then ], [ %call1, %if.else ]
ret i32 %retval.0, !dbg !21
}
!14 = !DILocation(line: 5, column: 9, scope: !15)
!16 = !DILocation(line: 5, column: 7, scope: !4)
```
As you can see, in 'entry' block, 'icmp' instruction and 'br' instruction have different debug locations. However, with current implementation, there's no distinction between debug locations of these two when they are lowered to asm instructions. This is because 'icmp' and 'br' become 'setcc' 'xor' and 'brcond' in SelectionDAG, where SDLoc of 'setcc' follows the debug location of 'icmp' but SDLOC of 'xor' and 'brcond' follows the debug location of 'br' instruction, and SDLoc of 'xor' overwrites SDLoc of 'setcc' when they are folded. This patch addresses this issue.
Reviewers: atrick, bogner, andreadb, craig.topper, aprantl
Reviewed By: andreadb
Subscribers: jlebar, mkuper, jholewinski, andreadb, llvm-commits
Differential Revision: https://reviews.llvm.org/D29813
llvm-svn: 296825
This bug was introduced with:
https://reviews.llvm.org/rL296699
There may be a way to loosen the restriction, but for now just bail out
on any opaque constant.
The tests show that opacity is target-specific. This goes back to cost
calculations in ConstantHoisting based on TTI->getIntImmCost().
llvm-svn: 296768
This is part of the ongoing attempt to improve select codegen for all targets and select
canonicalization in IR (see D24480 for more background). The transform is a subset of what
is done in InstCombine's FoldOpIntoSelect().
I first noticed a regression in the x86 avx512-insert-extract.ll tests with a patch that
hopes to convert more selects to basic math ops. This appears to be a general missing DAG
transform though, so I added tests for all standard binops in rL296621
(PowerPC was chosen semi-randomly; it has scripted FileCheck support, but so do ARM and x86).
The poor output for "sel_constants_shl_constant" is tracked with:
https://bugs.llvm.org/show_bug.cgi?id=32105
Differential Revision: https://reviews.llvm.org/D30502
llvm-svn: 296699
Add check that deleted nodes do not get added to worklist. This can
occur when a node's operand is simplified to an existing node.
This fixes PR32108.
Reviewers: jyknight, hfinkel, chandlerc
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D30506
llvm-svn: 296668
Resubmit r295336 after the bug with non-zero offset patterns on BE targets is fixed (r296336).
Support {a|s}ext, {a|z|s}ext load nodes as a part of load combine patters.
Reviewed By: filcab
Differential Revision: https://reviews.llvm.org/D29591
llvm-svn: 296651
Recommiting after fixup of 32-bit aliasing sign offset bug in DAGCombiner.
* Simplify Consecutive Merge Store Candidate Search
Now that address aliasing is much less conservative, push through
simplified store merging search and chain alias analysis which only
checks for parallel stores through the chain subgraph. This is cleaner
as the separation of non-interfering loads/stores from the
store-merging logic.
When merging stores search up the chain through a single load, and
finds all possible stores by looking down from through a load and a
TokenFactor to all stores visited.
This improves the quality of the output SelectionDAG and the output
Codegen (save perhaps for some ARM cases where we correctly constructs
wider loads, but then promotes them to float operations which appear
but requires more expensive constant generation).
Some minor peephole optimizations to deal with improved SubDAG shapes (listed below)
Additional Minor Changes:
1. Finishes removing unused AliasLoad code
2. Unifies the chain aggregation in the merged stores across code
paths
3. Re-add the Store node to the worklist after calling
SimplifyDemandedBits.
4. Increase GatherAllAliasesMaxDepth from 6 to 18. That number is
arbitrary, but seems sufficient to not cause regressions in
tests.
5. Remove Chain dependencies of Memory operations on CopyfromReg
nodes as these are captured by data dependence
6. Forward loads-store values through tokenfactors containing
{CopyToReg,CopyFromReg} Values.
7. Peephole to convert buildvector of extract_vector_elt to
extract_subvector if possible (see
CodeGen/AArch64/store-merge.ll)
8. Store merging for the ARM target is restricted to 32-bit as
some in some contexts invalid 64-bit operations are being
generated. This can be removed once appropriate checks are
added.
This finishes the change Matt Arsenault started in r246307 and
jyknight's original patch.
Many tests required some changes as memory operations are now
reorderable, improving load-store forwarding. One test in
particular is worth noting:
CodeGen/PowerPC/ppc64-align-long-double.ll - Improved load-store
forwarding converts a load-store pair into a parallel store and
a memory-realized bitcast of the same value. However, because we
lose the sharing of the explicit and implicit store values we
must create another local store. A similar transformation
happens before SelectionDAG as well.
Reviewers: arsenm, hfinkel, tstellarAMD, jyknight, nhaehnle
llvm-svn: 296476
DAGCombiner already supports peeking thorough shuffles to improve vector element extraction, but legalization often leaves us in situations where we need to extract vector elements after shuffles have already been lowered.
This patch adds support for VECTOR_EXTRACT_ELEMENT/PEXTRW/PEXTRB instructions to attempt to handle target shuffles as well. I've covered some basic scenarios including handling shuffle mask scaling and the implicit zero-extension of PEXTRW/PEXTRB, there is more that could be done here (that I've mentioned in TODOs) but I haven't found many cases where its worth it.
Differential Revision: https://reviews.llvm.org/D30176
llvm-svn: 296381
This pattern is essentially a i16 load from p+1 address:
%p1.i16 = bitcast i8* %p to i16*
%p2.i8 = getelementptr i8, i8* %p, i64 2
%v1 = load i16, i16* %p1.i16
%v2.i8 = load i8, i8* %p2.i8
%v2 = zext i8 %v2.i8 to i16
%v1.shl = shl i16 %v1, 8
%res = or i16 %v1.shl, %v2
Current implementation would identify %v1 load as the first byte load and would mistakenly emit a i16 load from %p1.i16 address. This patch adds a check that the first byte is loaded from a non-zero offset of the first load address. This way this address can be used as the base address for the combined value. Otherwise just give up combining.
llvm-svn: 296336
Recommiting after fixup of 32-bit aliasing sign offset bug in DAGCombiner.
* Simplify Consecutive Merge Store Candidate Search
Now that address aliasing is much less conservative, push through
simplified store merging search and chain alias analysis which only
checks for parallel stores through the chain subgraph. This is cleaner
as the separation of non-interfering loads/stores from the
store-merging logic.
When merging stores search up the chain through a single load, and
finds all possible stores by looking down from through a load and a
TokenFactor to all stores visited.
This improves the quality of the output SelectionDAG and the output
Codegen (save perhaps for some ARM cases where we correctly constructs
wider loads, but then promotes them to float operations which appear
but requires more expensive constant generation).
Some minor peephole optimizations to deal with improved SubDAG shapes (listed below)
Additional Minor Changes:
1. Finishes removing unused AliasLoad code
2. Unifies the chain aggregation in the merged stores across code
paths
3. Re-add the Store node to the worklist after calling
SimplifyDemandedBits.
4. Increase GatherAllAliasesMaxDepth from 6 to 18. That number is
arbitrary, but seems sufficient to not cause regressions in
tests.
5. Remove Chain dependencies of Memory operations on CopyfromReg
nodes as these are captured by data dependence
6. Forward loads-store values through tokenfactors containing
{CopyToReg,CopyFromReg} Values.
7. Peephole to convert buildvector of extract_vector_elt to
extract_subvector if possible (see
CodeGen/AArch64/store-merge.ll)
8. Store merging for the ARM target is restricted to 32-bit as
some in some contexts invalid 64-bit operations are being
generated. This can be removed once appropriate checks are
added.
This finishes the change Matt Arsenault started in r246307 and
jyknight's original patch.
Many tests required some changes as memory operations are now
reorderable, improving load-store forwarding. One test in
particular is worth noting:
CodeGen/PowerPC/ppc64-align-long-double.ll - Improved load-store
forwarding converts a load-store pair into a parallel store and
a memory-realized bitcast of the same value. However, because we
lose the sharing of the explicit and implicit store values we
must create another local store. A similar transformation
happens before SelectionDAG as well.
Reviewers: arsenm, hfinkel, tstellarAMD, jyknight, nhaehnle
llvm-svn: 296252
The motivation for filling out these select-of-constants cases goes back to D24480,
where we discussed removing an IR fold from add(zext) --> select. And that goes back to:
https://reviews.llvm.org/rL75531https://reviews.llvm.org/rL159230
The idea is that we should always canonicalize patterns like this to a select-of-constants
in IR because that's the smallest IR and the best for value tracking. Note that we currently
do the opposite in some cases (like the cases in *this* patch). Ie, the proposed folds in
this patch already exist in InstCombine today:
https://github.com/llvm-mirror/llvm/blob/master/lib/Transforms/InstCombine/InstCombineSelect.cpp#L1151
As this patch shows, most targets generate better machine code for simple ext/add/not ops
rather than a select of constants. So the follow-up steps to make this less of a patchwork
of special-case folds and missing IR canonicalization:
1. Have DAGCombiner convert any select of constants into ext/add/not ops.
2 Have InstCombine canonicalize in the other direction (create more selects).
Differential Revision: https://reviews.llvm.org/D30180
llvm-svn: 296137
During legalization we are often creating shuffles (via a build_vector scalarization stage) that are "any_extend_vector_inreg" style masks, and also other masks that are the equivalent of "truncate_vector_inreg" (if we had such a thing).
This patch is an attempt to match these cases to help undo the effects of just leaving shuffle lowering to handle it - which typically means we lose track of the undefined elements of the shuffles resulting in an unnecessary extension+truncation stage for widened illegal types.
The 2011-10-21-widen-cmp.ll regression will be fixed by making SIGN_EXTEND_VECTOR_IN_REG legal in SSE instead of lowering them to X86ISD::VSEXT (PR31712).
Differential Revision: https://reviews.llvm.org/D29454
llvm-svn: 295451
Resubmit -r295314 with PowerPC and AMDGPU tests updated.
Support {a|s}ext, {a|z|s}ext load nodes as a part of load combine patters.
Reviewed By: filcab
Differential Revision: https://reviews.llvm.org/D29591
llvm-svn: 295336
Support {a|s}ext, {a|z|s}ext load nodes as a part of load combine patters.
Reviewed By: filcab
Differential Revision: https://reviews.llvm.org/D29591
llvm-svn: 295314
We currently can't legalize those, but we should really not be creating
them in the first place, since legalization would probably look similar to the
way we legalize CONCAT_VECTORS - basically replace the INSERT with a BUILD.
This fixes PR311956.
Differential Revision: https://reviews.llvm.org/D29961
llvm-svn: 295213
This reverts commit r294186.
On an internal test, this triggers an out-of-memory error on PPC,
presumably because there is another dagcombine that does the exact
opposite triggering and endless loop consuming more and more memory.
Chandler has started at creating a reduced test case and we'll attach it
as soon as possible.
llvm-svn: 294288
Summary: This avoid the need to duplicate all pattern and actually end up exposing some opportunity to optimize existing pattern that did not exists in both directions on an existing test case.
Reviewers: mkuper, spatel, bkramer, RKSimon, zvi
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D29541
llvm-svn: 294125
Recommiting after fixing X86 inc/dec chain bug.
* Simplify Consecutive Merge Store Candidate Search
Now that address aliasing is much less conservative, push through
simplified store merging search and chain alias analysis which only
checks for parallel stores through the chain subgraph. This is cleaner
as the separation of non-interfering loads/stores from the
store-merging logic.
When merging stores search up the chain through a single load, and
finds all possible stores by looking down from through a load and a
TokenFactor to all stores visited.
This improves the quality of the output SelectionDAG and the output
Codegen (save perhaps for some ARM cases where we correctly constructs
wider loads, but then promotes them to float operations which appear
but requires more expensive constant generation).
Some minor peephole optimizations to deal with improved SubDAG shapes (listed below)
Additional Minor Changes:
1. Finishes removing unused AliasLoad code
2. Unifies the chain aggregation in the merged stores across code
paths
3. Re-add the Store node to the worklist after calling
SimplifyDemandedBits.
4. Increase GatherAllAliasesMaxDepth from 6 to 18. That number is
arbitrary, but seems sufficient to not cause regressions in
tests.
5. Remove Chain dependencies of Memory operations on CopyfromReg
nodes as these are captured by data dependence
6. Forward loads-store values through tokenfactors containing
{CopyToReg,CopyFromReg} Values.
7. Peephole to convert buildvector of extract_vector_elt to
extract_subvector if possible (see
CodeGen/AArch64/store-merge.ll)
8. Store merging for the ARM target is restricted to 32-bit as
some in some contexts invalid 64-bit operations are being
generated. This can be removed once appropriate checks are
added.
This finishes the change Matt Arsenault started in r246307 and
jyknight's original patch.
Many tests required some changes as memory operations are now
reorderable, improving load-store forwarding. One test in
particular is worth noting:
CodeGen/PowerPC/ppc64-align-long-double.ll - Improved load-store
forwarding converts a load-store pair into a parallel store and
a memory-realized bitcast of the same value. However, because we
lose the sharing of the explicit and implicit store values we
must create another local store. A similar transformation
happens before SelectionDAG as well.
Reviewers: arsenm, hfinkel, tstellarAMD, jyknight, nhaehnle
llvm-svn: 293893
Summary:
The affected transforms all implicitly use associativity of addition,
for which we usually require unsafe math to be enabled.
The "Aggressive" flag is only meant to convey information about the
performance of the fused ops relative to a fmul+fadd sequence.
Fixes Bug 31626.
Reviewers: spatel, hfinkel, mehdi_amini, arsenm, tstellarAMD
Subscribers: jholewinski, nemanjai, wdng, llvm-commits
Differential Revision: https://reviews.llvm.org/D28675
llvm-svn: 293635
The type system already requires that the number of vector elements must fit in 32-bits so an index should as well. Even if the type of the index were larger all we care about is that the constant index can fit in 64-bits so that we can call getZExtValue.
llvm-svn: 293413
* Simplify Consecutive Merge Store Candidate Search
Now that address aliasing is much less conservative, push through
simplified store merging search and chain alias analysis which only
checks for parallel stores through the chain subgraph. This is cleaner
as the separation of non-interfering loads/stores from the
store-merging logic.
When merging stores search up the chain through a single load, and
finds all possible stores by looking down from through a load and a
TokenFactor to all stores visited.
This improves the quality of the output SelectionDAG and the output
Codegen (save perhaps for some ARM cases where we correctly constructs
wider loads, but then promotes them to float operations which appear
but requires more expensive constant generation).
Some minor peephole optimizations to deal with improved SubDAG shapes (listed below)
Additional Minor Changes:
1. Finishes removing unused AliasLoad code
2. Unifies the chain aggregation in the merged stores across code
paths
3. Re-add the Store node to the worklist after calling
SimplifyDemandedBits.
4. Increase GatherAllAliasesMaxDepth from 6 to 18. That number is
arbitrary, but seems sufficient to not cause regressions in
tests.
5. Remove Chain dependencies of Memory operations on CopyfromReg
nodes as these are captured by data dependence
6. Forward loads-store values through tokenfactors containing
{CopyToReg,CopyFromReg} Values.
7. Peephole to convert buildvector of extract_vector_elt to
extract_subvector if possible (see
CodeGen/AArch64/store-merge.ll)
8. Store merging for the ARM target is restricted to 32-bit as
some in some contexts invalid 64-bit operations are being
generated. This can be removed once appropriate checks are
added.
This finishes the change Matt Arsenault started in r246307 and
jyknight's original patch.
Many tests required some changes as memory operations are now
reorderable, improving load-store forwarding. One test in
particular is worth noting:
CodeGen/PowerPC/ppc64-align-long-double.ll - Improved load-store
forwarding converts a load-store pair into a parallel store and
a memory-realized bitcast of the same value. However, because we
lose the sharing of the explicit and implicit store values we
must create another local store. A similar transformation
happens before SelectionDAG as well.
Reviewers: arsenm, hfinkel, tstellarAMD, jyknight, nhaehnle
llvm-svn: 293184
The previous patch (https://reviews.llvm.org/rL289538) got reverted because of a bug. Chandler also requested some changes to the algorithm.
http://lists.llvm.org/pipermail/llvm-commits/Week-of-Mon-20161212/413479.html
This is an updated patch. The key difference is that collectBitProviders (renamed to calculateByteProvider) now collects the origin of one byte, not the whole value. It simplifies the implementation and allows to stop the traversal earlier if we know that the result won't be used.
From the original commit:
Match a pattern where a wide type scalar value is loaded by several narrow loads and combined by shifts and ors. Fold it into a single load or a load and a bswap if the targets supports it.
Assuming little endian target:
i8 *a = ...
i32 val = a[0] | (a[1] << 8) | (a[2] << 16) | (a[3] << 24)
=>
i32 val = *((i32)a)
i8 *a = ...
i32 val = (a[0] << 24) | (a[1] << 16) | (a[2] << 8) | a[3]
=>
i32 val = BSWAP(*((i32)a))
This optimization was discussed on llvm-dev some time ago in "Load combine pass" thread. We came to the conclusion that we want to do this transformation late in the pipeline because in presence of atomic loads load widening is irreversible transformation and it might hinder other optimizations.
Eventually we'd like to support folding patterns like this where the offset has a variable and a constant part:
i32 val = a[i] | (a[i + 1] << 8) | (a[i + 2] << 16) | (a[i + 3] << 24)
Matching the pattern above is easier at SelectionDAG level since address reassociation has already happened and the fact that the loads are adjacent is clear. Understanding that these loads are adjacent at IR level would have involved looking through geps/zexts/adds while looking at the addresses.
The general scheme is to match OR expressions by recursively calculating the origin of individual bytes which constitute the resulting OR value. If all the OR bytes come from memory verify that they are adjacent and match with little or big endian encoding of a wider value. If so and the load of the wider type (and bswap if needed) is allowed by the target generate a load and a bswap if needed.
Reviewed By: RKSimon, filcab, chandlerc
Differential Revision: https://reviews.llvm.org/D27861
llvm-svn: 293036
clang already emits this with -cl-no-signed-zeros, but codegen
doesn't do anything with it. Treat it like the other fast math
attributes, and change one place to use it.
llvm-svn: 293024
Even with aggressive fusion enabled, this requires duplicating
the fmul, or increases an fadd to another fma which is not an
improvement.
llvm-svn: 291642
Summary:
Instead of matching:
(a + i) + 1 -> (a + i, undef, 1)
Now it matches:
(a + i) + 1 -> (a, i, 1)
Reviewers: rengolin
Differential Revision: http://reviews.llvm.org/D26367
From: Evgeny Stupachenko <evstupac@gmail.com>
llvm-svn: 291012
1.Fix pessimized case in FIXME.
2.Add tests for it.
3.The canonicalisation on shifts results in different sequence for
tests of machine-licm.Correct some check lines.
Differential Revision: https://reviews.llvm.org/D27916
llvm-svn: 290410
This is for splitMergedValStore in DAG Combine to share the target query interface
with similar logic in CodeGenPrepare.
Differential Revision: https://reviews.llvm.org/D24707
llvm-svn: 290363
idiom.
r289538: Match load by bytes idiom and fold it into a single load
r289540: Fix a buildbot failure introduced by r289538
r289545: Use more detailed assertion messages in the code ...
r289646: Add a couple of assertions to the load combine code ...
This DAG combine has a bad crash in it that is quite hard to trigger
sadly -- it relies on sneaking code with UB through the SDAG build and
into this particular combine. I've responded to the original commit with
a test case that reproduces it.
However, the code also has other problems that will require substantial
changes to address and so I'm going ahead and reverting it for now. This
should unblock us and perhaps others that are hitting the crash in the
wild and will let a fresh patch with updated approach come in cleanly
afterward.
Sorry for any trouble or disruption!
llvm-svn: 289916
The original motivation for this patch comes from wanting to canonicalize
more IR to selects and also canonicalizing min/max.
If we're going to do that, we need more backend fixups to undo select codegen
when simpler ops will do. I chose AArch64 for the tests because that shows the
difference in the simplest way. This should fix:
https://llvm.org/bugs/show_bug.cgi?id=31175
Differential Revision: https://reviews.llvm.org/D27489
llvm-svn: 289738
Retrying after fixing after removing load-store factoring through
token factors in favor of improved token factor operand pruning
Simplify Consecutive Merge Store Candidate Search
Now that address aliasing is much less conservative, push through
simplified store merging search which only checks for parallel stores
through the chain subgraph. This is cleaner as the separation of
non-interfering loads/stores from the store-merging logic.
Whem merging stores, search up the chain through a single load, and
finds all possible stores by looking down from through a load and a
TokenFactor to all stores visited. This improves the quality of the
output SelectionDAG and generally the output CodeGen (with some
exceptions).
Additional Minor Changes:
1. Finishes removing unused AliasLoad code
2. Unifies the the chain aggregation in the merged stores across
code paths
3. Re-add the Store node to the worklist after calling
SimplifyDemandedBits.
4. Increase GatherAllAliasesMaxDepth from 6 to 18. That number is
arbitrary, but seemed sufficient to not cause regressions in
tests.
This finishes the change Matt Arsenault started in r246307 and
jyknight's original patch.
Many tests required some changes as memory operations are now
reorderable. Some tests relying on the order were changed to use
volatile memory operations
Noteworthy tests:
CodeGen/AArch64/argument-blocks.ll -
It's not entirely clear what the test_varargs_stackalign test is
supposed to be asserting, but the new code looks right.
CodeGen/AArch64/arm64-memset-inline.lli -
CodeGen/AArch64/arm64-stur.ll -
CodeGen/ARM/memset-inline.ll -
The backend now generates *worse* code due to store merging
succeeding, as we do do a 16-byte constant-zero store efficiently.
CodeGen/AArch64/merge-store.ll -
Improved, but there still seems to be an extraneous vector insert
from an element to itself?
CodeGen/PowerPC/ppc64-align-long-double.ll -
Worse code emitted in this case, due to the improved store->load
forwarding.
CodeGen/X86/dag-merge-fast-accesses.ll -
CodeGen/X86/MergeConsecutiveStores.ll -
CodeGen/X86/stores-merging.ll -
CodeGen/Mips/load-store-left-right.ll -
Restored correct merging of non-aligned stores
CodeGen/AMDGPU/promote-alloca-stored-pointer-value.ll -
Improved. Correctly merges buffer_store_dword calls
CodeGen/AMDGPU/si-triv-disjoint-mem-access.ll -
Improved. Sidesteps loading a stored value and
merges two stores
CodeGen/X86/pr18023.ll -
This test has been removed, as it was asserting incorrect
behavior. Non-volatile stores *CAN* be moved past volatile loads,
and now are.
CodeGen/X86/vector-idiv.ll -
CodeGen/X86/vector-lzcnt-128.ll -
It's basically impossible to tell what these tests are actually
testing. But, looks like the code got better due to the memory
operations being recognized as non-aliasing.
CodeGen/X86/win32-eh.ll -
Both loads of the securitycookie are now merged.
Reviewers: arsenm, hfinkel, tstellarAMD, jyknight, nhaehnle
Subscribers: wdng, nhaehnle, nemanjai, arsenm, weimingz, niravd, RKSimon, aemerson, qcolombet, dsanders, resistor, tstellarAMD, t.p.northover, spatel
Differential Revision: https://reviews.llvm.org/D14834
llvm-svn: 289659
Generalize sdiv/udiv/srem/urem combines using APInt::isPowerOf2, which only works for const/splat-const values, to call SelectionDAG::isKnownToBeAPowerOfTwo instead which recognises many more cases.
Added a DAGCombiner::BuildLogBase2 helper since PowerOf2 combines often involve taking the log2 of such a value.
Differential Revision: https://reviews.llvm.org/D27714
llvm-svn: 289654
Match a pattern where a wide type scalar value is loaded by several narrow loads and combined by shifts and ors. Fold it into a single load or a load and a bswap if the targets supports it.
Assuming little endian target:
i8 *a = ...
i32 val = a[0] | (a[1] << 8) | (a[2] << 16) | (a[3] << 24)
=>
i32 val = *((i32)a)
i8 *a = ...
i32 val = (a[0] << 24) | (a[1] << 16) | (a[2] << 8) | a[3]
=>
i32 val = BSWAP(*((i32)a))
This optimization was discussed on llvm-dev some time ago in "Load combine pass" thread. We came to the conclusion that we want to do this transformation late in the pipeline because in presence of atomic loads load widening is irreversible transformation and it might hinder other optimizations.
Eventually we'd like to support folding patterns like this where the offset has a variable and a constant part:
i32 val = a[i] | (a[i + 1] << 8) | (a[i + 2] << 16) | (a[i + 3] << 24)
Matching the pattern above is easier at SelectionDAG level since address reassociation has already happened and the fact that the loads are adjacent is clear. Understanding that these loads are adjacent at IR level would have involved looking through geps/zexts/adds while looking at the addresses.
The general scheme is to match OR expressions by recursively calculating the origin of individual bits which constitute the resulting OR value. If all the OR bits come from memory verify that they are adjacent and match with little or big endian encoding of a wider value. If so and the load of the wider type (and bswap if needed) is allowed by the target generate a load and a bswap if needed.
Reviewed By: hfinkel, RKSimon, filcab
Differential Revision: https://reviews.llvm.org/D26149
llvm-svn: 289538
Retrying after fixing overly aggressive load-store forwarding optimization.
Simplify Consecutive Merge Store Candidate Search
Now that address aliasing is much less conservative, push through
simplified store merging search which only checks for parallel stores
through the chain subgraph. This is cleaner as the separation of
non-interfering loads/stores from the store-merging logic.
Whem merging stores, search up the chain through a single load, and
finds all possible stores by looking down from through a load and a
TokenFactor to all stores visited. This improves the quality of the
output SelectionDAG and generally the output CodeGen (with some
exceptions).
Additional Minor Changes:
1. Finishes removing unused AliasLoad code
2. Unifies the the chain aggregation in the merged stores across
code paths
3. Re-add the Store node to the worklist after calling
SimplifyDemandedBits.
4. Increase GatherAllAliasesMaxDepth from 6 to 18. That number is
arbitrary, but seemed sufficient to not cause regressions in
tests.
This finishes the change Matt Arsenault started in r246307 and
jyknight's original patch.
Many tests required some changes as memory operations are now
reorderable. Some tests relying on the order were changed to use
volatile memory operations
Noteworthy tests:
CodeGen/AArch64/argument-blocks.ll -
It's not entirely clear what the test_varargs_stackalign test is
supposed to be asserting, but the new code looks right.
CodeGen/AArch64/arm64-memset-inline.lli -
CodeGen/AArch64/arm64-stur.ll -
CodeGen/ARM/memset-inline.ll -
The backend now generates *worse* code due to store merging
succeeding, as we do do a 16-byte constant-zero store efficiently.
CodeGen/AArch64/merge-store.ll -
Improved, but there still seems to be an extraneous vector insert
from an element to itself?
CodeGen/PowerPC/ppc64-align-long-double.ll -
Worse code emitted in this case, due to the improved store->load
forwarding.
CodeGen/X86/dag-merge-fast-accesses.ll -
CodeGen/X86/MergeConsecutiveStores.ll -
CodeGen/X86/stores-merging.ll -
CodeGen/Mips/load-store-left-right.ll -
Restored correct merging of non-aligned stores
CodeGen/AMDGPU/promote-alloca-stored-pointer-value.ll -
Improved. Correctly merges buffer_store_dword calls
CodeGen/AMDGPU/si-triv-disjoint-mem-access.ll -
Improved. Sidesteps loading a stored value and
merges two stores
CodeGen/X86/pr18023.ll -
This test has been removed, as it was asserting incorrect
behavior. Non-volatile stores *CAN* be moved past volatile loads,
and now are.
CodeGen/X86/vector-idiv.ll -
CodeGen/X86/vector-lzcnt-128.ll -
It's basically impossible to tell what these tests are actually
testing. But, looks like the code got better due to the memory
operations being recognized as non-aliasing.
CodeGen/X86/win32-eh.ll -
Both loads of the securitycookie are now merged.
Reviewers: arsenm, hfinkel, tstellarAMD, jyknight, nhaehnle
Subscribers: wdng, nhaehnle, nemanjai, arsenm, weimingz, niravd, RKSimon, aemerson, qcolombet, dsanders, resistor, tstellarAMD, t.p.northover, spatel
Differential Revision: https://reviews.llvm.org/D14834
llvm-svn: 289221
Handle the case where a sign extension has ended up being split into separate stages (typically to get around vector legal ops) and a zext + sext_in_reg gets inserted.
Differential Revision: https://reviews.llvm.org/D27461
llvm-svn: 288842
Summary:
When X = 0 and Y = inf, the original code produces inf, but the transformed
code produces nan. So this transform (and its relatives) should only be
used when the no-infs-fp-math flag is explicitly enabled.
Also disable the transform using fmad (intermediate rounding) when unsafe-math
is not enabled, since it can reduce the precision of the result; consider this
example with binary floating point numbers with two bits of mantissa:
x = 1.01
y = 111
x * (y + 1) = 1.01 * 1000 = 1010 (this is the exact result; no rounding occurs at any step)
x * y + x = 1000.11 + 1.01 =r 1000 + 1.01 = 1001.01 =r 1000 (with rounding towards zero)
The example relies on rounding towards zero at least in the second step.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=98578
Reviewers: RKSimon, tstellarAMD, spatel, arsenm
Subscribers: wdng, llvm-commits
Differential Revision: https://reviews.llvm.org/D26602
llvm-svn: 288506
Forward store values to matching loads down through token
factors. Factored from D14834.
Reviewers: jyknight, hfinkel
Subscribers: hfinkel, nemanjai, llvm-commits
Differential Revision: https://reviews.llvm.org/D26080
llvm-svn: 287773
We have the following DAGCombiner transformations:
(mul (shl X, c1), c2) -> (mul X, c2 << c1)
(mul (shl X, C), Y) -> (shl (mul X, Y), C)
(shl (mul x, c1), c2) -> (mul x, c1 << c2)
Usually the constant shift is optimised by SelectionDAG::getNode when it is
constructed, by SelectionDAG::FoldConstantArithmetic, but when we're dealing
with vectors and one of those vector constants contains an undef element
FoldConstantArithmetic does not fold and we enter an infinite loop.
Fix this by making FoldConstantArithmetic use getNode to decide how to fold each
vector element, the same as FoldConstantVectorArithmetic does, and rather than
adding the constant shift to the work list instead only apply the transformation
if it's already been folded into a constant, as if it's not we're going to loop
endlessly. Additionally add missing NoOpaques to one of those transformations,
which I noticed when writing the tests for this.
Differential Revision: https://reviews.llvm.org/D26605
llvm-svn: 287766
Implemented widening (v2f32) and splitting (v16f64).
On splitting, I use "popcnt" to calculate memory increment.
More type legalization work will come in the next patches.
llvm-svn: 287761
The generic infrastructure to compute the Newton series for reciprocal and
reciprocal square root was conceived to allow a target to compute the series
itself. However, the original code did not properly consider this condition
if returned by a target. This patch addresses the issues to allow a target
to compute the series on its own.
Differential revision: https://reviews.llvm.org/D22975
llvm-svn: 286523
We were failing to extract a constant splat shift value if the shifted value was being masked.
The (shl (and (setcc) N01CV) N1CV) -> (and (setcc) N01CV<<N1CV) combine was unnecessarily preventing this.
llvm-svn: 286454
Summary:
Have MergeConsecutiveStores explicitly return information about the stores
that were merged, so that we can safely determine whether the starting
node has been freed.
Reviewers: chandlerc, bogner, niravd
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D25601
llvm-svn: 285916
2 new intrinsics covering AVX-512 compress/expand functionality.
This implementation includes syntax, DAG builder, operation lowering and tests.
Does not include: handling of illegal data types, codegen prepare pass and the cost model.
llvm-svn: 285876
Instead of asserting that the shift count is != 0 we just bail out
as it's not profitable trying to optimize a node which will be
removed anyway.
Differential Revision: https://reviews.llvm.org/D26098
llvm-svn: 285480
Summary:
Do *not* perform combines such as:
vector_shuffle<4,1,2,3>(build_vector(Ud, C0, C1 C2), scalar_to_vector(X))
->
build_vector(X, C0, C1, C2)
Keeping the shuffle allows lowering the constant build_vector to a materialized
constant vector (such as a vector-load from the constant-pool or some other idiom).
Reviewers: delena, igorb, spatel, mkuper, andreadb, RKSimon
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D25524
llvm-svn: 285063
0 - X --> 0, if the sub is NUW
0 - X --> 0, if X is 0 or the minimum signed value and the sub is NSW
0 - X --> X, if X is 0 or the minimum signed value
This is the DAG equivalent of:
https://reviews.llvm.org/rL284649
plus the fold for the NUW case which already existed in InstSimplify.
Note that we miss a vector fold because of a deficiency in the DAG version of
computeKnownBits().
llvm-svn: 284844
As discussed in D24815, let's start the process of killing off the broken fast-math global
state housed in TargetOptions and eliminate the need for function-level fast-math attributes.
Here we enable two similar folds that are possible when we don't care about signed-zero:
fadd nsz x, 0 --> x
fsub nsz 0, x --> -x
Note that although the test cases include a 'sin' function call, I'm side-stepping the
FMF-on-calls question (and lack of support in the DAG) for now. It's not needed for these
tests - isNegatibleForFree/GetNegatedExpression just look through a ISD::FSIN node.
Also, when we create an FNEG node and propagate the Flags of the FSUB to it, this doesn't
actually do anything today because Flags are silently dropped for any node that is not a
binary operator.
Differential Revision: https://reviews.llvm.org/D25297
llvm-svn: 284824
This is a retry of r284495 which was reverted at r284513 due to use-after-scope bugs
caused by faulty usage of StringRef.
This version also renames a pair of functions:
getRecipEstimateDivEnabled()
getRecipEstimateSqrtEnabled()
as suggested by Eric Christopher.
original commit msg:
[Target] remove TargetRecip class; move reciprocal estimate isel functionality to TargetLowering
This is a follow-up to https://reviews.llvm.org/D24816 - where we changed reciprocal estimates to be function attributes
rather than TargetOptions.
This patch is intended to be a structural, but not functional change. By moving all of the
TargetRecip functionality into TargetLowering, we can remove all of the reciprocal estimate
state, shield the callers from the string format implementation, and simplify/localize the
logic needed for a target to enable this.
If a function has a "reciprocal-estimates" attribute, those settings may override the target's
default reciprocal preferences for whatever operation and data type we're trying to optimize.
If there's no attribute string or specific setting for the op/type pair, just use the target
default settings.
As noted earlier, a better solution would be to move the reciprocal estimate settings to IR
instructions and SDNodes rather than function attributes, but that's a multi-step job that
requires infrastructure improvements. I intend to work on that, but it's not clear how long
it will take to get all the pieces in place.
Differential Revision: https://reviews.llvm.org/D25440
llvm-svn: 284746
This will get the same ConstantSDNode scalar or vector splat value as the current separate dyn_cast<ConstantSDNode> / isVector() approach.
llvm-svn: 284578
This is a follow-up to D24816 - where we changed reciprocal estimates to be function attributes
rather than TargetOptions.
This patch is intended to be a structural, but not functional change. By moving all of the
TargetRecip functionality into TargetLowering, we can remove all of the reciprocal estimate
state, shield the callers from the string format implementation, and simplify/localize the
logic needed for a target to enable this.
If a function has a "reciprocal-estimates" attribute, those settings may override the target's
default reciprocal preferences for whatever operation and data type we're trying to optimize.
If there's no attribute string or specific setting for the op/type pair, just use the target
default settings.
As noted earlier, a better solution would be to move the reciprocal estimate settings to IR
instructions and SDNodes rather than function attributes, but that's a multi-step job that
requires infrastructure improvements. I intend to work on that, but it's not clear how long
it will take to get all the pieces in place.
Differential Revision: https://reviews.llvm.org/D25440
llvm-svn: 284495
