We used to only allow SCEVAddRecExpr for pointer expressions in order to
be able to compute the bounds. However this is also trivially possible
for loop-invariant addresses (scUnknown) since then the bounds are the
address itself.
Interestingly, we used allow this for the special case when the
loop-invariant address happens to also be an SCEVAddRecExpr (in an outer
loop).
There are a couple more loops that are vectorized in SPEC after this.
My guess is that the main reason we don't see more because for example a
loop-invariant load is vectorized into a splat vector with several
vector-inserts. This is likely to make the vectorization unprofitable.
I.e. we don't notice that a later LICM will move all of this out of the
loop so the cost estimate should really be 0.
llvm-svn: 264243
Summary:
These intrinsics expose the BUFFER_ATOMIC_* instructions and will be used
by Mesa to implement atomics with buffer semantics. The intrinsic interface
matches that of buffer.load.format and buffer.store.format, except that the
GLC bit is not exposed (it is automatically deduced based on whether the
return value is used).
The change of hasSideEffects is required for TableGen to accept the pattern
that matches the intrinsic.
Reviewers: tstellarAMD, arsenm
Subscribers: arsenm, rivanvx, llvm-commits
Differential Revision: http://reviews.llvm.org/D18151
llvm-svn: 263791
Summary:
As explained by the comment, threads will typically see different values
returned by atomic instructions even if the arguments are equal.
Reviewers: arsenm, tstellarAMD
Subscribers: arsenm, llvm-commits
Differential Revision: http://reviews.llvm.org/D18156
llvm-svn: 263719
Summary:
When multiple threads perform an atomic op with the same arguments, they
will usually see different return values.
Reviewers: arsenm, tstellarAMD
Subscribers: arsenm, llvm-commits
Differential Revision: http://reviews.llvm.org/D18101
llvm-svn: 263440
This patch fixes the problem which occurs when loop-vectorize tries to use @llvm.masked.load/store intrinsic for a non-default addrspace pointer. It fails with "Calling a function with a bad signature!" assertion in CallInst constructor because it tries to pass a non-default addrspace pointer to the pointer argument which has default addrspace.
The fix is to add pointer type as another overloaded type to @llvm.masked.load/store intrinsics.
Reviewed By: reames
Differential Revision: http://reviews.llvm.org/D17270
llvm-svn: 263158
actually finish wiring up the old call graph.
There were bugs in the old call graph that hadn't been caught because it
wasn't being tested. It wasn't being tested because it wasn't in the
pipeline system and we didn't have a printing pass to run in tests. This
fixes all of that.
As for why I'm still keeping the old call graph alive its so that I can
port GlobalsAA to the new pass manager with out forking it to work with
the lazy call graph. That's clearly the right eventual design, but it
seems pragmatic to defer that until its necessary. The old call graph
works just fine for GlobalsAA.
llvm-svn: 263104
As part of r251146 InstCombine was extended to call computeKnownBits on
every value in the function to determine whether it happens to be
constant. This increases typical compiletime by 1-3% (5% in irgen+opt
time) in my measurements. On the other hand this case did not trigger
once in the whole llvm-testsuite.
This patch introduces the notion of ExpensiveCombines which are only
enabled for OptLevel > 2. I removed the check in InstructionSimplify as
that is called from various places where the OptLevel is not known but
given the rarity of the situation I think a check in InstCombine is
enough.
Differential Revision: http://reviews.llvm.org/D16835
llvm-svn: 263047
Building on the previous change, this generalizes
ScalarEvolution::getRangeViaFactoring to work with
{Ext(C?A:B)+k0,+,Ext(C?A:B)+k1} where Ext can be a zero extend, sign
extend or truncate operation, and k0 and k1 are constants.
llvm-svn: 262979
This change generalizes ScalarEvolution::getRangeViaFactoring to work
with {Ext(C?A:B),+,Ext(C?A:B)} where Ext can be a zero extend, sign
extend or truncate operation.
llvm-svn: 262978
Summary:
This testcase had me confused. It made me believe that you can use
alias scopes and alias scopes list interchangeably with alias.scope and
noalias. Both langref and the other testcase use scope lists so I went
looking.
Turns out using scope directly only happens to work by chance. When
ScopedNoAliasAAResult::mayAliasInScopes traverses this as a scope list:
!1 = !{!1, !0, !"some scope"}
, the first entry is in fact a scope but only because the scope is
happened to be defined self-referentially to make it unique globally.
The remaining elements in the tuple (!0, !"some scope") are considered
as scopes but AliasScopeNode::getDomain will just bail on those without
any error.
This change avoids this ambiguity in the test but I've also been
wondering if we should issue some sort of a diagnostics.
Reviewers: dexonsmith, hfinkel
Subscribers: mssimpso, llvm-commits
Differential Revision: http://reviews.llvm.org/D16670
llvm-svn: 262841
This experiment was originally about trying to use facts implied dominating conditions to infer more precise known bits. While the compile time was found to be acceptable on several large code bases, we never found sufficiently profitable examples to justify turning on the code by default. Given this, it's time to abandon the experiment.
Several folks have commented that they've found this useful for experimentation, but nothing has come of those experiments. Given how easy the patch is to apply, there's no reason to leave the code in tree.
For anyone interested in further investigation in this area, I recommend finding the summary email I sent on one of the original review threads. In particular, I now believe the use-list based approach is strictly worse than the dom-tree-walking approach.
llvm-svn: 262646
After r262438 we can have provably positive NSW SCEV expressions whose
zero extensions cannot be simplified (since r262438 makes SCEV better at
computing constant ranges). This means demoting sexts of positive add
recurrences eagerly can result in an unsimplified zero extension where
we could have had a simplified sign extension. This change fixes the
issue by teaching SCEV to demote sext of a positive SCEV expression to a
zext only if the sext could not be simplified.
llvm-svn: 262638
Have ScalarEvolution::getRange re-consider cases like "{C?A:B,+,C?P:Q}"
by factoring out "C" and computing RangeOf{A,+,P} union RangeOf({B,+,Q})
instead.
The latter can be easier to compute precisely in cases like
"{C?0:N,+,C?1:-1}" N is the backedge taken count of the loop; since in
such cases the latter form simplifies to [0,N+1) union [0,N+1).
llvm-svn: 262438
it to actually test the new pass manager AA wiring.
This patch was extracted from the (somewhat too large) D12357 and
rebosed on top of the slightly different design of the new pass manager
AA wiring that I just landed. With this we can start testing the AA in
a thorough way with the new pass manager.
Some minor cleanups to the code in the pass was necessitated here, but
otherwise it is a very minimal change.
Differential Revision: http://reviews.llvm.org/D17372
llvm-svn: 261403
reference-edge SCCs.
This essentially builds a more normal call graph as a subgraph of the
"reference graph" that was the old model. This allows both to exist and
the different use cases to use the aspect which addresses their needs.
Specifically, the pass manager and other *ordering* constrained logic
can use the reference graph to achieve conservative order of visit,
while analyses reasoning about attributes and other properties derived
from reachability can reason about the direct call graph.
Note that this isn't necessarily complete: it doesn't model edges to
declarations or indirect calls. Those can be found by scanning the
instructions of the function if desirable, and in fact every user
currently does this in order to handle things like calls to instrinsics.
If useful, we could consider caching this information in the call graph
to save the instruction scans, but currently that doesn't seem to be
important.
An important realization for why the representation chosen here works is
that the call graph is a formal subset of the reference graph and thus
both can live within the same data structure. All SCCs of the call graph
are necessarily contained within an SCC of the reference graph, etc.
The design is to build 'RefSCC's to model SCCs of the reference graph,
and then within them more literal SCCs for the call graph.
The formation of actual call edge SCCs is not done lazily, unlike
reference edge 'RefSCC's. Instead, once a reference SCC is formed, it
directly builds the call SCCs within it and stores them in a post-order
sequence. This is used to provide a consistent platform for mutation and
update of the graph. The post-order also allows for very efficient
updates in common cases by bounding the number of nodes (and thus edges)
considered.
There is considerable common code that I'm still looking for the best
way to factor out between the various DFS implementations here. So far,
my attempts have made the code harder to read and understand despite
reducing the duplication, which seems a poor tradeoff. I've not given up
on figuring out the right way to do this, but I wanted to wait until
I at least had the system working and tested to continue attempting to
factor it differently.
This also requires introducing several new algorithms in order to handle
all of the incremental update scenarios for the more complex structure
involving two edge colorings. I've tried to comment the algorithms
sufficiently to make it clear how this is expected to work, but they may
still need more extensive documentation.
I know that there are some changes which are not strictly necessarily
coupled here. The process of developing this started out with a very
focused set of changes for the new structure of the graph and
algorithms, but subsequent changes to bring the APIs and code into
consistent and understandable patterns also ended up touching on other
aspects. There was no good way to separate these out without causing
*massive* merge conflicts. Ultimately, to a large degree this is
a rewrite of most of the core algorithms in the LCG class and so I don't
think it really matters much.
Many thanks to the careful review by Sanjoy Das!
Differential Revision: http://reviews.llvm.org/D16802
llvm-svn: 261040
than the SCC object, and have it scan the instruction stream directly
rather than relying on call records.
This makes the behavior of this routine consistent between libc routines
and LLVM intrinsics for libc routines. We can go and start teaching it
about those being norecurse, but we should behave the same for the
intrinsic and the libc routine rather than differently. I chatted with
James Molloy and the inconsistency doesn't seem intentional and likely
is due to intrinsic calls not being modelled in the call graph analyses.
This also fixes a bug where we would deduce norecurse on optnone
functions, when generally we try to handle optnone functions as-if they
were replaceable and thus unanalyzable.
llvm-svn: 260813
Summary:
Passes that call `getAnalysisIfAvailable<T>` also need to call
`addUsedIfAvailable<T>` in `getAnalysisUsage` to indicate to the
legacy pass manager that it uses `T`. This contract was being
violated by passes that used `createLegacyPMAAResults`. This change
fixes this by exposing a helper in AliasAnalysis.h,
`addUsedAAAnalyses`, that is complementary to createLegacyPMAAResults
and does the right thing when called from `getAnalysisUsage`.
Reviewers: chandlerc
Subscribers: mcrosier, llvm-commits
Differential Revision: http://reviews.llvm.org/D17010
llvm-svn: 260183
IndVarSimplify assumes scAddRecExpr to be expanded in literal form instead of
canonical form by calling disableCanonicalMode after it creates SCEVExpander.
When CanonicalMode is disabled, SCEVExpander::expand should always return PHI
node for scAddRecExpr. r259736 broke the assumption.
The fix is to let SCEVExpander::expand skip the reuse Value logic if
CanonicalMode is false.
In addition, Besides IndVarSimplify, LSR pass also calls disableCanonicalMode
before doing rewrite. We can remove the original check of LSRMode in reuse
Value logic and use CanonicalMode instead.
llvm-svn: 260174
sanitizer issue. The PredicatedScalarEvolution's copy constructor
wasn't copying the Generation value, and was leaving it un-initialized.
Original commit message:
[SCEV][LAA] Add no wrap SCEV predicates and use use them to improve strided pointer detection
Summary:
This change adds no wrap SCEV predicates with:
- support for runtime checking
- support for expression rewriting:
(sext ({x,+,y}) -> {sext(x),+,sext(y)}
(zext ({x,+,y}) -> {zext(x),+,sext(y)}
Note that we are sign extending the increment of the SCEV, even for
the zext case. This is needed to cover the fairly common case where y would
be a (small) negative integer. In order to do this, this change adds two new
flags: nusw and nssw that are applicable to AddRecExprs and permit the
transformations above.
We also change isStridedPtr in LAA to be able to make use of
these predicates. With this feature we should now always be able to
work around overflow issues in the dependence analysis.
Reviewers: mzolotukhin, sanjoy, anemet
Subscribers: mzolotukhin, sanjoy, llvm-commits, rengolin, jmolloy, hfinkel
Differential Revision: http://reviews.llvm.org/D15412
llvm-svn: 260112
We shouldn't assert when there are no memchecks, since we
can have SCEV checks. There is already an assert covering
the case where there are no SCEV checks or memchecks.
This also changes the LAA pointer wrapping versioning test
to use the loop versioning pass (this was how I managed to
trigger the assert in the loop versioning pass).
llvm-svn: 260086
Summary:
This change adds no wrap SCEV predicates with:
- support for runtime checking
- support for expression rewriting:
(sext ({x,+,y}) -> {sext(x),+,sext(y)}
(zext ({x,+,y}) -> {zext(x),+,sext(y)}
Note that we are sign extending the increment of the SCEV, even for
the zext case. This is needed to cover the fairly common case where y would
be a (small) negative integer. In order to do this, this change adds two new
flags: nusw and nssw that are applicable to AddRecExprs and permit the
transformations above.
We also change isStridedPtr in LAA to be able to make use of
these predicates. With this feature we should now always be able to
work around overflow issues in the dependence analysis.
Reviewers: mzolotukhin, sanjoy, anemet
Subscribers: mzolotukhin, sanjoy, llvm-commits, rengolin, jmolloy, hfinkel
Differential Revision: http://reviews.llvm.org/D15412
llvm-svn: 260085
Current SCEV expansion will expand SCEV as a sequence of operations
and doesn't utilize the value already existed. This will introduce
redundent computation which may not be cleaned up throughly by
following optimizations.
This patch introduces an ExprValueMap which is a map from SCEV to the
set of equal values with the same SCEV. When a SCEV is expanded, the
set of values is checked and reused whenever possible before generating
a sequence of operations.
The original commit triggered regressions in Polly tests. The regressions
exposed two problems which have been fixed in current version.
1. Polly will generate a new function based on the old one. To generate an
instruction for the new function, it builds SCEV for the old instruction,
applies some tranformation on the SCEV generated, then expands the transformed
SCEV and insert the expanded value into new function. Because SCEV expansion
may reuse value cached in ExprValueMap, the value in old function may be
inserted into new function, which is wrong.
In SCEVExpander::expand, there is a logic to check the cached value to
be used should dominate the insertion point. However, for the above
case, the check always passes. That is because the insertion point is
in a new function, which is unreachable from the old function. However
for unreachable node, DominatorTreeBase::dominates thinks it will be
dominated by any other node.
The fix is to simply add a check that the cached value to be used in
expansion should be in the same function as the insertion point instruction.
2. When the SCEV is of scConstant type, expanding it directly is cheaper than
reusing a normal value cached. Although in the cached value set in ExprValueMap,
there is a Constant type value, but it is not easy to find it out -- the cached
Value set is not sorted according to the potential cost. Existing reuse logic
in SCEVExpander::expand simply chooses the first legal element from the cached
value set.
The fix is that when the SCEV is of scConstant type, don't try the reuse
logic. simply expand it.
Differential Revision: http://reviews.llvm.org/D12090
llvm-svn: 259736
Current SCEV expansion will expand SCEV as a sequence of operations
and doesn't utilize the value already existed. This will introduce
redundent computation which may not be cleaned up throughly by
following optimizations.
This patch introduces an ExprValueMap which is a map from SCEV to the
set of equal values with the same SCEV. When a SCEV is expanded, the
set of values is checked and reused whenever possible before generating
a sequence of operations.
Differential Revision: http://reviews.llvm.org/D12090
llvm-svn: 259662
This regresses a test in LoopVectorize, so I'll need to go away and think about how to solve this in a way that isn't broken.
From the writeup in PR26071:
What's happening is that ComputeKnownZeroes is telling us that all bits except the LSB are zero. We're then deciding that only the LSB needs to be demanded from the icmp's inputs.
This is where we're wrong - we're assuming that after simplification the bits that were known zero will continue to be known zero. But they're not - during trivialization the upper bits get changed (because an XOR isn't shrunk), so the icmp fails.
The fault is in demandedbits - its contract does clearly state that a non-demanded bit may either be zero or one.
llvm-svn: 259649
differentiate between indirect references to functions an direct calls.
This doesn't do a whole lot yet other than change the print out produced
by the analysis, but it lays the groundwork for a very major change I'm
working on next: teaching the call graph to actually be a call graph,
modeling *both* the indirect reference graph and the call graph
simultaneously. More details on that in the next patch though.
The rest of this is essentially a bunch of over-engineering that won't
be interesting until the next patch. But this also isolates essentially
all of the churn necessary to introduce the edge abstraction from the
very important behavior change necessary in order to separately model
the two graphs. So it should make review of the subsequent patch a bit
easier at the cost of making this patch seem poorly motivated. ;]
Differential Revision: http://reviews.llvm.org/D16038
llvm-svn: 259463
The computation of ICmp demanded bits is independent of the individual operand being evaluated. We simply return a mask consisting of the minimum leading zeroes of both operands.
We were incorrectly passing "I" to ComputeKnownBits - this should be "UserI->getOperand(0)". In cases where we were evaluating the 1th operand, we were taking the minimum leading zeroes of it and itself.
This should fix PR26266.
llvm-svn: 258690
Some patterns of select+compare allow us to know exactly the value of the uppermost bits in the select result. For example:
%b = icmp ugt i32 %a, 5
%c = select i1 %b, i32 2, i32 %a
Here we know that %c is bounded by 5, and therefore KnownZero = ~APInt(5).getActiveBits() = ~7.
There are several such patterns, and this patch attempts to understand a reasonable subset of them - namely when the base values are the same (as above), and when they are related by a simple (add nsw), for example (add nsw %a, 4) and %a.
llvm-svn: 257769
Summary:
Since globals may escape as function arguments (even when they have been
found to be non-escaping, because of optimizations such as memcpyoptimizer
that replaces stores with memcpy), all arguments to a function are checked
during query to make sure they are identifiable. At that time, also ensure
we return a conservative result only if the arguments don't alias to our global.
Reviewers: hfinkel, jmolloy
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D16140
llvm-svn: 257750
The early return seems to be missed. This causes a radical and wrong loop
optimization on powerpc. It isn't reproducible on x86_64, because
"UseInterleaved" is false.
Patch by Tim Shen.
llvm-svn: 257134
See PR25822 for a more full summary, but we were conflating the concepts of "capture" and "escape". We were proving nocapture and using that proof to infer noescape, which is not true. Escaped-ness is a function-local property - as soon as a value is used in a call argument it escapes. Capturedness is a related but distinct property. It implies a *temporally limited* escape. Consider:
static int a;
int b;
int g(int * nocapture arg);
int f() {
a = 2; // Even though a escapes to g, it is not captured so can be treated as non-escaping here.
g(&a); // But here it must be treated as escaping.
g(&b); // Now that g(&a) has returned we know it was not captured so we can treat it as non-escaping again.
}
The original commit did not sufficiently understand this nuance and so caused PR25822 and PR26046.
r248576 included both a performance improvement (which has been backed out) and a related conformance fix (which has been kept along with its testcase).
llvm-svn: 257058
Summary:
This reverts commit 5a9e526f29cf8510ab5c3d566fbdcf47ac24e1e9.
As per discussion in D15665
This also add a test case so that regression introduced by that diff are not reintroduced.
Reviewers: vaivaswatha, jmolloy, hfinkel, reames
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D15919
llvm-svn: 256932
Most of the properties of memset_pattern16 can be now covered by the generic attributes and inferred by InferFunctionAttrs. The only exceptions are:
- We don't yet have a writeonly attribute for the first argument.
- We don't have an attribute for modeling the access size facts encoded in MemoryLocation.cpp.
Differential Revision: http://reviews.llvm.org/D15879
llvm-svn: 256911
The cost is calculated for all X86 targets. When gather/scatter instruction
is not supported we calculate the cost of scalar sequence.
Differential revision: http://reviews.llvm.org/D15677
llvm-svn: 256519
Summary: This patch changes gc.statepoint intrinsic's return type to token type instead of i32 type. Using token types could prevent LLVM to merge different gc.statepoint nodes into PHI nodes and cause further problems with gc relocations. The patch also changes the way on how gc.relocate and gc.result look for their corresponding gc.statepoint on unwind path. The current implementation uses the selector value extracted from a { i8*, i32 } landingpad as a hook to find the gc.statepoint, while the patch directly uses a token type landingpad (http://reviews.llvm.org/D15405) to find the gc.statepoint.
Reviewers: sanjoy, JosephTremoulet, pgavlin, igor-laevsky, mjacob
Subscribers: reames, mjacob, sanjoy, llvm-commits
Differential Revision: http://reviews.llvm.org/D15662
llvm-svn: 256443
This patch removes all weight-related interfaces from BPI and replace
them by probability versions. With this patch, we won't use edge weight
anymore in either IR or MC passes. Edge probabilitiy is a better
representation in terms of CFG update and validation.
Differential revision: http://reviews.llvm.org/D15519
llvm-svn: 256263
This is recommit of r256028 with minor fixes in unittests:
CodeGen/Mips/eh.ll
CodeGen/Mips/insn-zero-size-bb.ll
Original commit message:
When identifying blocks post-dominated by an unreachable-terminated block
in BranchProbabilityInfo, consider only the edge to the normal destination
block if the terminator is InvokeInst and let calcInvokeHeuristics() decide
edge weights for the InvokeInst.
llvm-svn: 256202
This patch transforms truncation between vectors of integers into
X86ISD::PACKUS/PACKSS operations during DAG combine. We don't do it in
lowering phase because after type legalization, the original truncation
will be turned into a BUILD_VECTOR with each element that is extracted
from a vector and then truncated, and from them it is difficult to do
this optimization. This greatly improves the performance of truncations
on some specific types.
Cost table is updated accordingly.
Differential revision: http://reviews.llvm.org/D14588
llvm-svn: 256194
Summary:
The analysis of shader inputs was completely wrong. We were passing the
wrong index to AttributeSet::hasAttribute() and the logic for which
inputs where in SGPRs was wrong too.
Reviewers: arsenm
Subscribers: arsenm, llvm-commits
Differential Revision: http://reviews.llvm.org/D15608
llvm-svn: 256082
This patch removes all getEdgeWeight() interfaces from CodeGen directory. As
getEdgeProbability() is a little more expensive than getEdgeWeight(), I will
compose a patch soon in which BPI only stores probabilities instead of edge
weights so that getEdgeProbability() will have O(1) time.
Differential revision: http://reviews.llvm.org/D15489
llvm-svn: 256039
When identifying blocks post-dominated by an unreachable-terminated block
in BranchProbabilityInfo, consider only the edge to the normal destination
block if the terminator is InvokeInst and let calcInvokeHeuristics() decide
edge weights for the InvokeInst.
llvm-svn: 256028
Summary:
1. Modify AnalyzeCallGraph() to retain function info for external functions
if the function has [InaccessibleMemOr]ArgMemOnly flags.
2. When analyzing the use of a global is function parameter at a call site,
mark the callee also as modifying the global appropriately.
3. Add additional test cases.
Depends on D15499
Reviewers: hfinkel, jmolloy
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D15605
llvm-svn: 255994
Previously in the conversion cost table there are no entries for integer-integer
conversions on SSE2. This will result in imprecise costs for certain vectorized
operations. This patch adds those entries for SSE2 and SSE4.1. The cost numbers
are counted from the result of running llc on the new test case in this patch.
Differential revision: http://reviews.llvm.org/D15132
llvm-svn: 255315
Currently in LLVM's cost model, a vectorized arithmetic instruction will have
high cost if its type is split into multiple registers. However, this
punishment is too heavy and unnecessary. The overhead of the split should not
be on arithmetic instructions but instructions that implement the split. Note
that during vectorization we have calculated the register pressure, and we
only choose proper interleaving factor (and also vectorization factor) so
that we don't use more registers than the maximum number.
Here is a very simple example: if a vadd has the cost 1, and if we double VF
so that we need two registers to perform it, then its cost will become 4 with
the current implementation, which will prevent us to use larger VF.
Differential revision: http://reviews.llvm.org/D15159
llvm-svn: 254671
I checked and updated the cost of AVX-512 conversion operations. Added cost of conversion operations in DQ mode.
Conversion of illegal types that requires vector split is not calculated right now (like for other X86 targets).
Differential Revision: http://reviews.llvm.org/D15074
llvm-svn: 254494
The cost for scalarized operations is computed as N * (scalar operation
cost + 1 extractelement + 1 insertelement). This partially fixes
inflating the cost of scalarized operations since every operation is
scalarized and free. I don't think we want any cost asociated with
scalarization, but for now insertelement is still counted. I'm not sure
if we should pretend that insertelement is also free, or add a way
to compute a custom scalarization cost.
llvm-svn: 254438
Cost calculation for vector GEP failed with due to invalid cast to GEP index operand.
The bug is fixed, added a test.
http://reviews.llvm.org/D14976
llvm-svn: 254408
Note, this was reviewed (and more details are in) http://lists.llvm.org/pipermail/llvm-commits/Week-of-Mon-20151109/312083.html
These intrinsics currently have an explicit alignment argument which is
required to be a constant integer. It represents the alignment of the
source and dest, and so must be the minimum of those.
This change allows source and dest to each have their own alignments
by using the alignment attribute on their arguments. The alignment
argument itself is removed.
There are a few places in the code for which the code needs to be
checked by an expert as to whether using only src/dest alignment is
safe. For those places, they currently take the minimum of src/dest
alignments which matches the current behaviour.
For example, code which used to read:
call void @llvm.memcpy.p0i8.p0i8.i32(i8* %dest, i8* %src, i32 500, i32 8, i1 false)
will now read:
call void @llvm.memcpy.p0i8.p0i8.i32(i8* align 8 %dest, i8* align 8 %src, i32 500, i1 false)
For out of tree owners, I was able to strip alignment from calls using sed by replacing:
(call.*llvm\.memset.*)i32\ [0-9]*\,\ i1 false\)
with:
$1i1 false)
and similarly for memmove and memcpy.
I then added back in alignment to test cases which needed it.
A similar commit will be made to clang which actually has many differences in alignment as now
IRBuilder can generate different source/dest alignments on calls.
In IRBuilder itself, a new argument was added. Instead of calling:
CreateMemCpy(Dst, Src, getInt64(Size), DstAlign, /* isVolatile */ false)
you now call
CreateMemCpy(Dst, Src, getInt64(Size), DstAlign, SrcAlign, /* isVolatile */ false)
There is a temporary class (IntegerAlignment) which takes the source alignment and rejects
implicit conversion from bool. This is to prevent isVolatile here from passing its default
parameter to the source alignment.
Note, changes in future can now be made to codegen. I didn't change anything here, but this
change should enable better memcpy code sequences.
Reviewed by Hal Finkel.
llvm-svn: 253511
The underlying issues surrounding codegen for 32-bit vselects have been resolved. The pessimistic costs for 64-bit vselects remain due to the bad
scalarization that is still happening there.
I tested this on A57 in T32, A32 and A64 modes. I saw no regressions, and some improvements.
From my benchmarks, I saw these improvements in A57 (T32)
spec.cpu2000.ref.177_mesa 5.95%
lnt.SingleSource/Benchmarks/Shootout/strcat 12.93%
lnt.MultiSource/Benchmarks/MiBench/telecomm-CRC32/telecomm-CRC32 11.89%
I also measured A57 A32, A53 T32 and A9 T32 and found no performance regressions. I see much bigger wins in third-party benchmarks with this change
Differential Revision: http://reviews.llvm.org/D14743
llvm-svn: 253349
A function can be marked as norecurse if:
* The SCC to which it belongs has cardinality 1; and either
a) It does not call any non-norecurse function. This includes self-recursion; or
b) It only has one callsite and the function that callsite is within is marked norecurse.
a) is best propagated bottom-up and b) is best propagated top-down.
We build up the norecurse attributes bottom-up using the existing SCC pass, and mark functions with no obvious recursion (but not provably norecurse) to sweep later, top-down.
llvm-svn: 252862
This is fix for PR24059.
When we are hoisting instruction above some condition it may turn out
that metadata on this instruction was control dependant on the condition.
This metadata becomes invalid and we need to drop it.
This patch should cover most obvious places of speculative execution (which
I have found by greping isSafeToSpeculativelyExecute). I think there are more
cases but at least this change covers the severe ones.
Differential Revision: http://reviews.llvm.org/D14398
llvm-svn: 252604
Summary:
GetUnderlyingObjects() can return "null" among its list of objects,
we don't want to deduce that two pointers can point to the same
memory in this case, so filter it out.
Reviewers: anemet
Subscribers: dexonsmith, llvm-commits
From: Mehdi Amini <mehdi.amini@apple.com>
llvm-svn: 252149
Summary:
We now collect all types of dependences including lexically forward
deps not just "interesting" ones.
Reviewers: hfinkel
Subscribers: rengolin, llvm-commits
Differential Revision: http://reviews.llvm.org/D13256
llvm-svn: 251985
Summary:
When the dependence distance in zero then we have a loop-independent
dependence from the earlier to the later access.
No current client of LAA uses forward dependences so other than
potentially hitting the MaxDependences threshold earlier, this change
shouldn't affect anything right now.
This and the previous patch were tested together for compile-time
regression. None found in LNT/SPEC.
Reviewers: hfinkel
Subscribers: rengolin, llvm-commits
Differential Revision: http://reviews.llvm.org/D13255
llvm-svn: 251973
Summary:
Before this change, we didn't use to collect forward dependences since
none of the current clients (LV, LDist) required them.
The motivation to also collect forward dependences is a new pass
LoopLoadElimination (LLE) which discovers store-to-load forwarding
opportunities across the loop's backedge. The pass uses both lexically
forward or backward loop-carried dependences to detect these
opportunities.
The new pass also analyzes loop-independent (forward) dependences since
they can conflict with the loop-carried dependences in terms of how the
data flows through memory.
The newly added test only covers loop-carried forward dependences
because loop-independent ones are currently categorized as NoDep. The
next patch will fix this.
The two patches were tested together for compile-time regression. None
found in LNT/SPEC.
Note that with this change LAA provides all dependences rather than just
"interesting" ones. A subsequent NFC patch will remove the now trivial
isInterestingDependence and rename the APIs.
Reviewers: hfinkel
Subscribers: jmolloy, rengolin, llvm-commits
Differential Revision: http://reviews.llvm.org/D13254
llvm-svn: 251972
Have `getConstantEvolutionLoopExitValue` work correctly with multiple
entry loops.
As far as I can tell, `getConstantEvolutionLoopExitValue` never did the
right thing for multiple entry loops; and before r249712 it would
silently return an incorrect answer. r249712 changed SCEV to fail an
assert on a multiple entry loop, and this change fixes the underlying
issue.
llvm-svn: 251770
Prevent `createNodeFromSelectLikePHI` from creating SCEV expressions
that break LCSSA.
A better fix for the same issue is to teach SCEVExpander to not break
LCSSA by inserting PHI nodes at appropriate places. That's planned for
the future.
Fixes PR25360.
llvm-svn: 251756
Summary:
When forming expressions for phi nodes having an incoming value from
outside the loop A and a value coming from the previous iteration B
we were forming an AddRec if:
- B was an AddRec
- the value A was equal to the value for B at iteration -1 (or equal
to the value of B shifted by one iteration, at iteration 0)
In this case, we were computing the expression to be the expression of
B, shifted by one iteration.
This changes generalizes the logic above by removing the restriction that
B needs to be an AddRec. For this we introduce two expression rewriters
that allow us to
- shift an expression by one iteration
- get the value of an expression at iteration 0
This allows us to get SCEV expressions for PHI nodes when these expressions
are not AddRecExprs.
Reviewers: sanjoy
Subscribers: llvm-commits, sanjoy
Differential Revision: http://reviews.llvm.org/D14175
llvm-svn: 251700
This teaches SCEV to compute //max// backedge taken counts for loops
like
for (int i = k; i != 0; i >>>= 1)
whatever();
SCEV yet cannot represent the exact backedge count for these loops, and
this patch does not change that. This is really geared towards teaching
SCEV that loops like the above are *not* infinite.
llvm-svn: 251558
In getArgModRefInfo we consider all arguments as having MRI_ModRef.
However for arguments marked with readonly attribute we can return
more precise answer - MRI_Ref.
Differential Revision: http://reviews.llvm.org/D13992
llvm-svn: 251525
When checking if an indirect global (a global with pointer type) is only assigned by allocation functions, first check if the global is itself initialized. If it is, it's not only assigned by allocation functions.
This fixes PR25309. Thanks to David Majnemer for reducing the test case!
llvm-svn: 251508
Even though we may not know the value of the shifter operand, it's possible we know the shifter operand is non-zero. This can allow us to infer more known bits - for example:
%1 = load %p !range {1, 5}
%2 = shl %q, %1
We don't know %1, but we do know that it is nonzero so %2[0] is known zero, and importantly %2 is known non-zero.
Calling isKnownNonZero is nontrivially expensive so use an Optional to run it lazily and cache its result.
llvm-svn: 251294
If the loaded type sizes don't match the element type of the sequential type, all bets are off and the addresses may, indeed, overlap.
Surprisingly, this just got caught in one test, on one builder, out of the 30+ builders testing this change. Congratulations go to http://lab.llvm.org:8011/builders/clang-aarch64-lnt/builds/5205.
llvm-svn: 251112
Summary:
This uses `ScalarEvolution::getRange` and not potentially control
dependent `nsw` and `nuw` bits on the arithmetic instruction.
Reviewers: atrick, hfinkel, nlewycky
Subscribers: llvm-commits, sanjoy
Differential Revision: http://reviews.llvm.org/D13613
llvm-svn: 251048
Instead of bailing out when we see loads, analyze them. If we can prove that the loaded-from address must escape, then we can conclude that a load from that address must escape too and therefore cannot alias a non-addr-taken global.
When checking if a Value can alias a non-addr-taken global, if the Value is a LoadInst of a non-global, recurse instead of bailing.
If we can follow a trail of loads up to some base that is captured, we know by inference that all the loads we followed are also captured.
llvm-svn: 251017
If the final indices of two GEPs can be proven to not be equal, and
the GEP is of a SequentialType (not a StructType), then the two GEPs
do not alias.
llvm-svn: 251016
isKnownNonEqual(A, B) returns true if it can be determined that A != B.
At the moment it only knows two facts, that a non-wrapping add of nonzero to a value cannot be that value:
A + B != A [where B != 0, addition is nsw or nuw]
and that contradictory known bits imply two values are not equal.
This patch also hooks this up to InstSimplify; InstSimplify had a peephole for the first fact but not the second so this teaches InstSimplify a new trick too (alas no measured performance impact!)
llvm-svn: 251012
Weak linkage and friends allow a symbol to be overriden outside the
code generator's model, so GlobalsAA shouldn't assume that anything it
can compute about such a symbol is valid.
llvm-svn: 250156
This patch also allows the -delinearize pass to delinearize expressions that do
not have an outermost SCEVAddRec expression. The SCEV::delinearize
infrastructure allowed this since r240952, but the -delinearize pass was not
updated yet.
llvm-svn: 250018
The new implementation works at least as well as the old implementation
did.
Also delete the associated preparation tests. They don't exercise
interesting corner cases of the new implementation. All the codegen
tests of the EH tables have already been ported.
llvm-svn: 249918
Instead of bailing out when we see an icmp, we can instead at least
say that if the upper bits of both operands are known zero, they are
not demanded. This doesn't help with signed comparisons, but it's at
least better than bailing out.
llvm-svn: 249687
Like adds and subtracts, muls ripple only to the left so we can use
the same logic.
While we're here, add a print method to DemandedBits so it can be used
with -analyze, which we'll use in the testcase.
llvm-svn: 249686
This reverts commit r249528 and reapply r249431. The fix for the
fallout has been commited in r249575.
From: Mehdi Amini <mehdi.amini@apple.com>
llvm-svn: 249581
With this patch, clang -O3 optimizes correctly providing > 1000x speedup on this artificial benchmark):
for (a=0; a<n; a++)
for (b=0; b<n; b++)
for (c=0; c<n; c++)
for (d=0; d<n; d++)
for (e=0; e<n; e++)
for (f=0; f<n; f++)
x++;
From test-suite/SingleSource/Benchmarks/Shootout/nestedloop.c
Reviewers: sanjoyd
Differential Revision: http://reviews.llvm.org/D13390
From: Mehdi Amini <mehdi.amini@apple.com>
llvm-svn: 249431
This commit changes the interface of the vld[1234], vld[234]lane, and vst[1234],
vst[234]lane ARM neon intrinsics and associates an address space with the
pointer that these intrinsics take. This changes, e.g.,
<2 x i32> @llvm.arm.neon.vld1.v2i32(i8*, i32)
to
<2 x i32> @llvm.arm.neon.vld1.v2i32.p0i8(i8*, i32)
This change ensures that address spaces are fully taken into account in the ARM
target during lowering of interleaved loads and stores.
Differential Revision: http://reviews.llvm.org/D12985
llvm-svn: 248887
The XOP shifts just have logical/arithmetic versions and the left/right shifts are controlled by whether the value is positive/negative. Because of this I've added new X86ISD nodes instead of trying to force them to use the existing shift nodes.
Additionally Excavator cores (bdver4) support XOP and AVX2 - meaning that it should use the AVX2 shifts when it can and fall back to XOP in other cases.
Differential Revision: http://reviews.llvm.org/D8690
llvm-svn: 248878
If a PHI starts at a non-negative constant, monotonically increases
(only adds of a constant are supported at the moment) and that add
does not wrap, then the PHI is known never to be zero.
llvm-svn: 248796
Summary:
If the trip count of a specific backedge is `N`, then we know that
backedge is effectively guarded by the condition `{0,+,1} u< N`. This
change teaches SCEV to use this condition to prove things in
`isLoopBackedgeGuardedByCond`.
Depends on D12948
Depends on D12949
The original checkin, r248608 had to be backed out due to an issue with
a ObjCXX unit test. That issue is now fixed, so re-landing.
Reviewers: atrick, reames, majnemer, hfinkel
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D12950
llvm-svn: 248638
BranchProbability now is represented by its numerator and denominator in uint32_t type. This patch changes this representation into a fixed point that is represented by the numerator in uint32_t type and a constant denominator 1<<31. This is quite similar to the representation of BlockMass in BlockFrequencyInfoImpl.h. There are several pros and cons of this change:
Pros:
1. It uses only a half space of the current one.
2. Some operations are much faster like plus, subtraction, comparison, and scaling by an integer.
Cons:
1. Constructing a probability using arbitrary numerator and denominator needs additional calculations.
2. It is a little less precise than before as we use a fixed denominator. For example, 1 - 1/3 may not be exactly identical to 1 / 3 (this will lead to many BranchProbability unit test failures). This should not matter when we only use it for branch probability. If we use it like a rational value for some precise calculations we may need another construct like ValueRatio.
One important reason for this change is that we propose to store branch probabilities instead of edge weights in MachineBasicBlock. We also want clients to use probability instead of weight when adding successors to a MBB. The current BranchProbability has more space which may be a concern.
Differential revision: http://reviews.llvm.org/D12603
llvm-svn: 248633
Summary:
If the trip count of a specific backedge is `N`, then we know that
backedge is effectively guarded by the condition `{0,+,1} u< N`. This
change teaches SCEV to use this condition to prove things in
`isLoopBackedgeGuardedByCond`.
Depends on D12948
Depends on D12949
Reviewers: atrick, reames, majnemer, hfinkel
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D12950
llvm-svn: 248608
Arguments to function calls marked "nocapture" can be marked as
non-escaping. However, nocapture is defined in terms of the lifetime
of the callee, and if the callee can directly or indirectly recurse to
the caller, the semantics of nocapture are invalid.
Therefore, we eagerly discover which SCC each function belongs to,
and later can check if callee and caller of a callsite belong to
the same SCC, in which case there could be recursion.
This means that we can't be so optimistic in
getModRefInfo(ImmutableCallsite) - previously we assumed all call
arguments never aliased with an escaping global. Now we need to check,
because a global could now be passed as an argument but still not
escape.
This also solves a related conformance problem: MemCpyOptimizer can
turn non-escaping stores of globals into calls to intrinsics like
llvm.memcpy/llvm/memset. This confuses GlobalsAA, which knows the
global can't escape and so returns NoModRef when queried, when
obviously a memcpy/memset call does indeed reference and modify its
arguments.
This fixes PR24800, PR24801, and PR24802.
llvm-svn: 248576
If the shifter operand is a constant, and all of the bits shifted out
are known to be zero, then if X is known non-zero at least one
non-zero bit must remain.
llvm-svn: 248508
Turns out that not every basic block is guaranteed to have a node within the DominatorTree. This is really hard to trigger, but the test case from the PR managed to do so. There's active discussion continuing about what documentation and/or invariants needed cleaned up.
llvm-svn: 248216
This patch addresses the issue of SCEV division asserting on some
input expressions (e.g., non-affine expressions) and quietly giving
up on others. When giving up, we set the quotient to be equal to
zero and the remainder to be equal to the numerator. With this
patch, we always quietly give up when we cannot perform the
division.
This patch also adds a test case for DependenceAnalysis that
previously caused an assertion.
Differential Revision: http://reviews.llvm.org/D11725
llvm-svn: 247314
Summary:
PR24757 was caused by some incorect math in
`ScalarEvolution::HowFarToZero` -- the smallest unsigned solution for X
in
2^N * A = 2^N * X
is not necessarily A.
Reviewers: atrick, majnemer, meheff
Subscribers: llvm-commits, sanjoy
Differential Revision: http://reviews.llvm.org/D12721
llvm-svn: 247242
with the new pass manager, and no longer relying on analysis groups.
This builds essentially a ground-up new AA infrastructure stack for
LLVM. The core ideas are the same that are used throughout the new pass
manager: type erased polymorphism and direct composition. The design is
as follows:
- FunctionAAResults is a type-erasing alias analysis results aggregation
interface to walk a single query across a range of results from
different alias analyses. Currently this is function-specific as we
always assume that aliasing queries are *within* a function.
- AAResultBase is a CRTP utility providing stub implementations of
various parts of the alias analysis result concept, notably in several
cases in terms of other more general parts of the interface. This can
be used to implement only a narrow part of the interface rather than
the entire interface. This isn't really ideal, this logic should be
hoisted into FunctionAAResults as currently it will cause
a significant amount of redundant work, but it faithfully models the
behavior of the prior infrastructure.
- All the alias analysis passes are ported to be wrapper passes for the
legacy PM and new-style analysis passes for the new PM with a shared
result object. In some cases (most notably CFL), this is an extremely
naive approach that we should revisit when we can specialize for the
new pass manager.
- BasicAA has been restructured to reflect that it is much more
fundamentally a function analysis because it uses dominator trees and
loop info that need to be constructed for each function.
All of the references to getting alias analysis results have been
updated to use the new aggregation interface. All the preservation and
other pass management code has been updated accordingly.
The way the FunctionAAResultsWrapperPass works is to detect the
available alias analyses when run, and add them to the results object.
This means that we should be able to continue to respect when various
passes are added to the pipeline, for example adding CFL or adding TBAA
passes should just cause their results to be available and to get folded
into this. The exception to this rule is BasicAA which really needs to
be a function pass due to using dominator trees and loop info. As
a consequence, the FunctionAAResultsWrapperPass directly depends on
BasicAA and always includes it in the aggregation.
This has significant implications for preserving analyses. Generally,
most passes shouldn't bother preserving FunctionAAResultsWrapperPass
because rebuilding the results just updates the set of known AA passes.
The exception to this rule are LoopPass instances which need to preserve
all the function analyses that the loop pass manager will end up
needing. This means preserving both BasicAAWrapperPass and the
aggregating FunctionAAResultsWrapperPass.
Now, when preserving an alias analysis, you do so by directly preserving
that analysis. This is only necessary for non-immutable-pass-provided
alias analyses though, and there are only three of interest: BasicAA,
GlobalsAA (formerly GlobalsModRef), and SCEVAA. Usually BasicAA is
preserved when needed because it (like DominatorTree and LoopInfo) is
marked as a CFG-only pass. I've expanded GlobalsAA into the preserved
set everywhere we previously were preserving all of AliasAnalysis, and
I've added SCEVAA in the intersection of that with where we preserve
SCEV itself.
One significant challenge to all of this is that the CGSCC passes were
actually using the alias analysis implementations by taking advantage of
a pretty amazing set of loop holes in the old pass manager's analysis
management code which allowed analysis groups to slide through in many
cases. Moving away from analysis groups makes this problem much more
obvious. To fix it, I've leveraged the flexibility the design of the new
PM components provides to just directly construct the relevant alias
analyses for the relevant functions in the IPO passes that need them.
This is a bit hacky, but should go away with the new pass manager, and
is already in many ways cleaner than the prior state.
Another significant challenge is that various facilities of the old
alias analysis infrastructure just don't fit any more. The most
significant of these is the alias analysis 'counter' pass. That pass
relied on the ability to snoop on AA queries at different points in the
analysis group chain. Instead, I'm planning to build printing
functionality directly into the aggregation layer. I've not included
that in this patch merely to keep it smaller.
Note that all of this needs a nearly complete rewrite of the AA
documentation. I'm planning to do that, but I'd like to make sure the
new design settles, and to flesh out a bit more of what it looks like in
the new pass manager first.
Differential Revision: http://reviews.llvm.org/D12080
llvm-svn: 247167
Summary:
We are not scalarizing the wide selects in codegen for i16 and i32 and
therefore we can remove the amortization factor. We still have issues
with i64 vectors in codegen though.
Reviewers: mcrosier
Subscribers: mcrosier, aemerson, llvm-commits, rengolin
Differential Revision: http://reviews.llvm.org/D12724
llvm-svn: 247156
This corner case happens when we have an irreducible SCC that is
deeply nested. As we work down the tree, the backedge masses start
getting smaller and smaller until we reach one that is down to 0.
Since we distribute the incoming mass using the backedge masses as
weight, the distributor does not allow zero weights. So, we simply
ignore them (which will just use the weights of the non-zero nodes).
llvm-svn: 247050
Pre-P8, when we generate code for unaligned vector loads (for Altivec and QPX
types), even when accounting for the combining that takes place for multiple
consecutive such loads, there is at least one load instructions and one
permutation for each load. Make sure the cost reported reflects the cost of the
permutes as well.
llvm-svn: 246807
I'm adding a regression test to better cover code generation for unaligned
vector loads and stores, but there's no functional change to the code
generation here. There is an improvement to the cost model for unaligned vector
loads and stores, mostly for QPX (for which we were not previously accounting
for the permutation-based loads), and the cost model implementation is cleaner.
llvm-svn: 246712
Hopefully this will end the GEPs saga!
This commit reverts r245394, i.e., it reapplies r221876 while incorporating the
fixes from D11847.
r221876 was not reapplied alone because it was not safe and D11847 was not
applied alone because it needs r221876 to produce correct results.
This should fix PR24596.
Original commit message for r221876:
Let's try this again...
This reverts r219432, plus a bug fix.
Description of the bug in r219432 (by Nick):
The bug was using AllPositive to break out of the loop; if the loop break
condition i != e is changed to i != e && AllPositive then the
test_modulo_analysis_with_global test I've added will fail as the Modulo will
be calculated incorrectly (as the last loop iteration is skipped, so Modulo
isn't updated with its Scale).
Nick also adds this comment:
ComputeSignBit is safe to use in loops as it takes into account phi nodes, and
the == EK_ZeroEx check is safe in loops as, no matter how the variable changes
between iterations, zero-extensions will always guarantee a zero sign bit. The
isValueEqualInPotentialCycles check is therefore definitely not needed as all
the variable analysis holds no matter how the variables change between loop
iterations.
And this patch also adds another enhancement to GetLinearExpression - basically
to convert ConstantInts to Offsets (see test_const_eval and
test_const_eval_scaled for the situations this improves).
Original commit message:
This reverts r218944, which reverted r218714, plus a bug fix.
Description of the bug in r218714 (by Nick):
The original patch forgot to check if the Scale in VariableGEPIndex flipped the
sign of the variable. The BasicAA pass iterates over the instructions in the
order they appear in the function, and so BasicAliasAnalysis::aliasGEP is
called with the variable it first comes across as parameter GEP1. Adding a
%reorder label puts the definition of %a after %b so aliasGEP is called with %b
as the first parameter and %a as the second. aliasGEP later calculates that %a
== %b + 1 - %idxprom where %idxprom >= 0 (if %a was passed as the first
parameter it would calculate %b == %a - 1 + %idxprom where %idxprom >= 0) -
ignoring that %idxprom is scaled by -1 here lead the patch to incorrectly
conclude that %a > %b.
Revised patch by Nick White, thanks! Thanks to Lang to isolating the bug.
Slightly modified by me to add an early exit from the loop and avoid
unnecessary, but expensive, function calls.
Original commit message:
Two related things:
1. Fixes a bug when calculating the offset in GetLinearExpression. The code
previously used zext to extend the offset, so negative offsets were converted
to large positive ones.
2. Enhance aliasGEP to deduce that, if the difference between two GEP
allocations is positive and all the variables that govern the offset are also
positive (i.e. the offset is strictly after the higher base pointer), then
locations that fit in the gap between the two base pointers are NoAlias.
Patch by Nick White!
Message from D11847:
Un-revert of r241981 and fix for PR23626. The 'Or' case of GetLinearExpression
delegates to 'Add' if possible, and if not it returns an Opaque value.
Unfortunately the Scale and Offsets weren't being set (and so defaulted to 0) -
and a scale of zero effectively removes the variable from the GEP instruction.
This meant that BasicAA would return MustAliases when it should have been
returning PartialAliases (and PR23626 was an example of the GVN pass using an
incorrect MustAlias to merge loads from what should have been different
pointers).
Differential Revision: http://reviews.llvm.org/D11847
Patch by Nick White <n.j.white@gmail.com>!
llvm-svn: 246502
Prior to this patch, we hadn't been marking StratifiedSets with the
appropriate StratifiedAttrs when handling the result of no-args call
instructions. This caused us to report NoAlias when handed, for
example, an escaped alloca and a result from an opaque function. Now we
properly mark the return value of said functions.
Thanks again to Chandler, Richard, and Nick for pinging me about this.
Differential review: http://reviews.llvm.org/D12408
llvm-svn: 246240
Fix how DependenceAnalysis calls delinearization, mirroring what is done in
Delinearization.cpp (mostly by making sure to call getSCEVAtScope before
delinearizing, and by removing the unnecessary 'Pairs == 1' check).
Patch by Vaivaswatha Nagaraj!
llvm-svn: 245408
Summary:
This change limits the minimum cost of an insert/extract
element operation to 2 in cases where this would result
in mixing of NEON and VFP code.
Reviewers: rengolin
Subscribers: mssimpso, aemerson, llvm-commits, rengolin
Differential Revision: http://reviews.llvm.org/D12030
llvm-svn: 245225
This change makes ScalarEvolution a stand-alone object and just produces
one from a pass as needed. Making this work well requires making the
object movable, using references instead of overwritten pointers in
a number of places, and other refactorings.
I've also wired it up to the new pass manager and added a RUN line to
a test to exercise it under the new pass manager. This includes basic
printing support much like with other analyses.
But there is a big and somewhat scary change here. Prior to this patch
ScalarEvolution was never *actually* invalidated!!! Re-running the pass
just re-wired up the various other analyses and didn't remove any of the
existing entries in the SCEV caches or clear out anything at all. This
might seem OK as everything in SCEV that can uses ValueHandles to track
updates to the values that serve as SCEV keys. However, this still means
that as we ran SCEV over each function in the module, we kept
accumulating more and more SCEVs into the cache. At the end, we would
have a SCEV cache with every value that we ever needed a SCEV for in the
entire module!!! Yowzers. The releaseMemory routine would dump all of
this, but that isn't realy called during normal runs of the pipeline as
far as I can see.
To make matters worse, there *is* actually a key that we don't update
with value handles -- there is a map keyed off of Loop*s. Because
LoopInfo *does* release its memory from run to run, it is entirely
possible to run SCEV over one function, then over another function, and
then lookup a Loop* from the second function but find an entry inserted
for the first function! Ouch.
To make matters still worse, there are plenty of updates that *don't*
trip a value handle. It seems incredibly unlikely that today GVN or
another pass that invalidates SCEV can update values in *just* such
a way that a subsequent run of SCEV will incorrectly find lookups in
a cache, but it is theoretically possible and would be a nightmare to
debug.
With this refactoring, I've fixed all this by actually destroying and
recreating the ScalarEvolution object from run to run. Technically, this
could increase the amount of malloc traffic we see, but then again it is
also technically correct. ;] I don't actually think we're suffering from
tons of malloc traffic from SCEV because if we were, the fact that we
never clear the memory would seem more likely to have come up as an
actual problem before now. So, I've made the simple fix here. If in fact
there are serious issues with too much allocation and deallocation,
I can work on a clever fix that preserves the allocations (while
clearing the data) between each run, but I'd prefer to do that kind of
optimization with a test case / benchmark that shows why we need such
cleverness (and that can test that we actually make it faster). It's
possible that this will make some things faster by making the SCEV
caches have higher locality (due to being significantly smaller) so
until there is a clear benchmark, I think the simple change is best.
Differential Revision: http://reviews.llvm.org/D12063
llvm-svn: 245193
Summary:
http://reviews.llvm.org/D11212 made Scalar Evolution able to propagate NSW and NUW flags from instructions to SCEVs for add instructions. This patch expands that to sub, mul and shl instructions.
This change makes LSR able to generate pointer induction variables for loops like these, where the index is 32 bit and the pointer is 64 bit:
for (int i = 0; i < numIterations; ++i)
sum += ptr[i - offset];
for (int i = 0; i < numIterations; ++i)
sum += ptr[i * stride];
for (int i = 0; i < numIterations; ++i)
sum += ptr[3 * (i << 7)];
Reviewers: atrick, sanjoy
Subscribers: sanjoy, majnemer, hfinkel, llvm-commits, meheff, jingyue, eliben
Differential Revision: http://reviews.llvm.org/D11860
llvm-svn: 245118
I looked into adding a warning / error for this to FileCheck, but there doesn't
seem to be a good way to avoid it triggering on the instances of it in RUN lines.
llvm-svn: 244481
In PR24288 it was pointed out that the easy case of a non-escaping
global and something that *obviously* required an escape sometimes is
hidden behind PHIs (or selects in theory). Because we have this binary
test, we can easily just check that all possible input values satisfy
the requirement. This is done with a (very small) recursion through PHIs
and selects. With this, the specific example from the PR is correctly
folded by GVN.
Differential Revision: http://reviews.llvm.org/D11707
llvm-svn: 244078
This patch vectorizes the v2i64/v4i64 ASHR shift operations - the last remaining integer vector shifts that are still being transferred to/from the scalar unit to be completed.
Differential Revision: http://reviews.llvm.org/D11439
llvm-svn: 243569
Summary:
Make Scalar Evolution able to propagate NSW and NUW flags from instructions to SCEVs in some cases. This is based on reasoning about when poison from instructions with these flags would trigger undefined behavior. This gives a 13% speed-up on some Eigen3-based Google-internal microbenchmarks for NVPTX.
There does not seem to be clear agreement about when poison should be considered to propagate through instructions. In this analysis, poison propagates only in cases where that should be uncontroversial.
This change makes LSR able to create induction variables for expressions like &ptr[i + offset] for loops like this:
for (int i = 0; i < limit; ++i) {
sum += ptr[i + offset];
}
Here ptr is a 64 bit pointer and offset is a 32 bit integer. For NVPTX, LSR currently creates an induction variable for i + offset instead, which is not as fast. Improving this situation is what brings the 13% speed-up on some Eigen3-based Google-internal microbenchmarks for NVPTX.
There are more details in this discussion on llvmdev.
June: http://lists.cs.uiuc.edu/pipermail/llvmdev/2015-June/thread.html#87234
July: http://lists.cs.uiuc.edu/pipermail/llvmdev/2015-July/thread.html#87392
Patch by Bjarke Roune
Reviewers: eliben, atrick, sanjoy
Subscribers: majnemer, hfinkel, jingyue, meheff, llvm-commits
Differential Revision: http://reviews.llvm.org/D11212
llvm-svn: 243460
no-alias with non-addr-taken globals: they cannot alias a captured
pointer.
If the non-global underlying object would have been a capture were it to
alias the global, we can firmly conclude no-alias. It isn't reasonable
for a transformation to introduce a capture in a way observable by an
alias analysis. Consider, even if it were to temporarily capture one
globals address into another global and then restore the other global
afterward, there would be no way for the load in the alias query to
observe that capture event correctly. If it observes it then the
temporary capturing would have changed the meaning of the program,
making it an invalid transformation. Even instrumentation passes or
a pass which is synthesizing stores to global variables to expose race
conditions in programs could not trigger this unless it queried the
alias analysis infrastructure mid-transform, in which case it seems
reasonable to return results from before the transform started.
See the comments in the change for a more detailed outlining of the
theory here.
This should address the primary performance regression found when the
non-conservatively-correct path of the alias query was disabled.
Differential Revision: http://reviews.llvm.org/D11410
llvm-svn: 243405
This is effectively an NFC but we can no longer print the index of the
pointer group so instead I print its address. This still lets us
cross-check the section that list the checks against the section that
list the groups (see how I modified the test).
E.g. before we printed this:
Run-time memory checks:
Check 0:
Comparing group 0:
%arrayidxC = getelementptr inbounds i16, i16* %c, i64 %store_ind
%arrayidxC1 = getelementptr inbounds i16, i16* %c, i64 %store_ind_inc
Against group 1:
%arrayidxA = getelementptr i16, i16* %a, i64 %ind
%arrayidxA1 = getelementptr i16, i16* %a, i64 %add
...
Grouped accesses:
Group 0:
(Low: %c High: (78 + %c))
Member: {%c,+,4}<%for.body>
Member: {(2 + %c),+,4}<%for.body>
Now we print this (changes are underlined):
Run-time memory checks:
Check 0:
Comparing group (0x7f9c6040c320):
~~~~~~~~~~~~~~
%arrayidxC1 = getelementptr inbounds i16, i16* %c, i64 %store_ind_inc
%arrayidxC = getelementptr inbounds i16, i16* %c, i64 %store_ind
Against group (0x7f9c6040c358):
~~~~~~~~~~~~~~
%arrayidxA1 = getelementptr i16, i16* %a, i64 %add
%arrayidxA = getelementptr i16, i16* %a, i64 %ind
...
Grouped accesses:
Group 0x7f9c6040c320:
~~~~~~~~~~~~~~
(Low: %c High: (78 + %c))
Member: {(2 + %c),+,4}<%for.body>
Member: {%c,+,4}<%for.body>
llvm-svn: 243354
r243250 appeared to break clang/test/Analysis/dead-store.c on one of the build
slaves, but I couldn't reproduce this failure locally. Probably a false
positive as I saw this test was broken by r243246 or r243247 too but passed
later without people fixing anything.
llvm-svn: 243253
Summary:
This patch updates TargetTransformInfoImplCRTPBase::getGEPCost to consider
addressing modes. It now returns TCC_Free when the GEP can be completely folded
to an addresing mode.
I started this patch as I refactored SLSR. Function isGEPFoldable looks common
and is indeed used by some WIP of mine. So I extracted that logic to getGEPCost.
Furthermore, I noticed getGEPCost wasn't directly tested anywhere. The best
testing bed seems CostModel, but its getInstructionCost method invokes
getAddressComputationCost for GEPs which provides very coarse estimation. So
this patch also makes getInstructionCost call the updated getGEPCost for GEPs.
This change inevitably breaks some tests because the cost model changes, but
nothing looks seriously wrong -- if we believe the new cost model is the right
way to go, these tests should be updated.
This patch is not perfect yet -- the comments in some tests need to be updated.
I want to know whether this is a right approach before fixing those details.
Reviewers: chandlerc, hfinkel
Subscribers: aschwaighofer, llvm-commits, aemerson
Differential Revision: http://reviews.llvm.org/D9819
llvm-svn: 243250
Summary:
In the benchmark (https://github.com/vetter/shoc) we are researching,
the duplicated load is not eliminated because MemoryDependenceAnalysis
hit the BlockScanLimit. This patch change it into a command line option
instead of a hardcoded value.
Patched by Xuetian Weng.
Test Plan: test/Analysis/MemoryDependenceAnalysis/memdep-block-scan-limit.ll
Reviewers: jingyue, reames
Subscribers: reames, llvm-commits
Differential Revision: http://reviews.llvm.org/D11366
llvm-svn: 242842
basic changes to the IR such as folding pointers through PHIs, Selects,
integer casts, store/load pairs, or outlining.
This leaves the feature available behind a flag. This flag's default
could be flipped if necessary, but the real-world performance impact of
this particular feature of GMR may not be sufficiently significant for
many folks to want to run the risk.
Currently, the risk here is somewhat mitigated by half-hearted attempts
to update GlobalsModRef when the rest of the optimizer changes
something. However, I am currently trying to remove that update
mechanism as it makes migrating the AA infrastructure to a form that can
be readily shared between new and old pass managers very challenging.
Without this update mechanism, it is possible that this still unlikely
failure mode will start to trip people, and so I wanted to try to
proactively avoid that.
There is a lengthy discussion on the mailing list about why the core
approach here is flawed, and likely would need to look totally different
to be both reasonably effective and resilient to basic IR changes
occuring. This patch is essentially the first of two which will enact
the result of that discussion. The next patch will remove the current
update mechanism.
Thanks to lots of folks that helped look at this from different angles.
Especial thanks to Michael Zolotukhin for doing some very prelimanary
benchmarking of LTO without GlobalsModRef to get a rough idea of the
impact we could be facing here. So far, it looks very small, but there
are some concerns lingering from other benchmarking. The default here
may get flipped if performance results end up pointing at this as a more
significant issue.
Also thanks to Pete and Gerolf for reviewing!
Differential Revision: http://reviews.llvm.org/D11213
llvm-svn: 242512
Summary:
The checking pointer grouping algorithm assumes that the
starts/ends of the pointers are well formed (start <= end).
The runtime memory checking algorithm also assumes this by doing:
start0 < end1 && start1 < end0
to detect conflicts. This check only works if start0 <= end0 and
start1 <= end1.
This change correctly orders the interval ends by either checking
the stride (if it is constant) or by using min/max SCEV expressions.
Reviewers: anemet, rengolin
Subscribers: rengolin, llvm-commits
Differential Revision: http://reviews.llvm.org/D11149
llvm-svn: 242400
Summary:
This patch allows phi nodes like
%x = phi [ %incptr, ... ] [ %var, ... ]
%incptr = getelementptr %x, 1
to be analyzed by BasicAliasAnalysis.
In aliasPHI, we can detect incoming values that are recursive GEPs with a
constant offset. Instead of trying to analyze a recursive GEP (and failing),
we now ignore it and instead set the size of the memory referenced by
the PHINode to UnknownSize. This represents all the possible memory
locations the pointer represented by the PHINode could be advanced to
by the GEP.
For now, this new behavior is turned off by default to allow debugging of
performance degradations seen with SPEC/x86 and Hexagon benchmarks.
The flag -basicaa-recphi turns it on.
Reviewers: hfinkel, sanjoy
Subscribers: tobiasvk_caf, sanjoy, llvm-commits
Differential Revision: http://reviews.llvm.org/D10368
llvm-svn: 242320
Summary:
The iteration order within a member of DepCands is deterministic
and therefore we don't have to sort the accesses within a member.
We also don't have to copy the indices of the pointers into a
vector, since we can iterate over the members of the class.
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D11145
llvm-svn: 242033
While the v4i32 shl operation is already vectorized using a cvttps2dq/pmulld pattern, the lshr/ashr opeations are still scalarized.
This patch adds vectorization support for non-uniform v4i32 shift operations - it splats constant shift amounts to allow them to use the immediate sse shift instructions, or extracts/zero-extends non-constant shift amounts. The individual results are then blended together.
Differential Revision: http://reviews.llvm.org/D11063
llvm-svn: 241989
r236894 caused PR23626 (Clang miscompiles webkit's base64 decoder), and was
reverted in r237984. This reapplies the patch with an additional test case for
PR23626 and the associated fix (both scales and offsets in the
BasicAliasAnalysis::constantOffsetHeuristic should initially be zero).
Patch by Nick White, thanks!
llvm-svn: 241981
This change adds new attribute called "argmemonly". Function marked with this attribute can only access memory through it's argument pointers. This attribute directly corresponds to the "OnlyAccessesArgumentPointees" ModRef behaviour in alias analysis.
Differential Revision: http://reviews.llvm.org/D10398
llvm-svn: 241979
Summary:
We were missing a corner case where DepCands was not available,
but we were using DepCands to compute the checking pointer
groups.
This adds a test for that regression.
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D11068
llvm-svn: 241818
Summary:
The checking pointer group construction algorithm relied on the iteration on DepCands.
We would need the same leaders across runs and the same iteration order over the underlying std::set for determinism.
This changes the algorithm to process the pointers in the order in which they were added to the runtime check, which is deterministic.
We need to update the tests, since the order in which pointers appear has changed.
No new tests were added, since it is impossible to test for non-determinism.
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D11064
llvm-svn: 241809
This commit ([LAA] Fix estimation of number of memchecks) regressed the
logic a bit. We shouldn't quit the analysis if we encounter a pointer
without known bounds *unless* we actually need to emit a memcheck for
it.
The original code was using NumComparisons which is now computed
differently. Instead I compute NeedRTCheck from NumReadPtrChecks and
NumWritePtrChecks.
As side note, I find the separation of NeedRTCheck and CanDoRT
confusing, so I will try to merge them in a follow-up patch.
llvm-svn: 241756
Summary:
Often filter-like loops will do memory accesses that are
separated by constant offsets. In these cases it is
common that we will exceed the threshold for the
allowable number of checks.
However, it should be possible to merge such checks,
sice a check of any interval againt two other intervals separated
by a constant offset (a,b), (a+c, b+c) will be equivalent with
a check againt (a, b+c), as long as (a,b) and (a+c, b+c) overlap.
Assuming the loop will be executed for a sufficient number of
iterations, this will be true. If not true, checking against
(a, b+c) is still safe (although not equivalent).
As long as there are no dependencies between two accesses,
we can merge their checks into a single one. We use this
technique to construct groups of accesses, and then check
the intervals associated with the groups instead of
checking the accesses directly.
Reviewers: anemet
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D10386
llvm-svn: 241673
This patch adds vectorization support for uniform constant i64 arithmetic shift right operators.
Differential Revision: http://reviews.llvm.org/D9645
llvm-svn: 241514
Summary:
r240039 adds a test case to check that CallGraph does the right thing
with respect to non-leaf intrinsics like statepoint and patchpoint.
This ports the same test case to LazyCallGraph. LazyCallGraph already
does the right thing with respect to escaping function pointers so there
is no need to change any code.
Reviewers: chandlerc
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D10582
llvm-svn: 241226
Summary:
Scalar evolution does not propagate the non-wrapping flags to values
that are derived from a non-wrapping induction variable because
the non-wrapping property could be flow-sensitive.
This change is a first attempt to establish the non-wrapping property in
some simple cases. The main idea is to look through the operations
defining the pointer. As long as we arrive to a non-wrapping AddRec via
a small chain of non-wrapping instruction, the pointer should not wrap
either.
I believe that this essentially is what Andy described in
http://article.gmane.org/gmane.comp.compilers.llvm.cvs/220731 as the way
forward.
Reviewers: aschwaighofer, nadav, sanjoy, atrick
Reviewed By: atrick
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D10472
llvm-svn: 240798
Merged separate (but equivalent) SSE2/AVX512F tests.
Removed codegen tests since these are already done better in test/CodeGen/X86.
The actual cost values still need to be updated to match recent codegen improvements.
llvm-svn: 240219
Summary:
Since FunctionMap has llvm::Function pointers as keys, the order in
which the traversal happens can differ from run to run, causing spurious
FileCheck failures. Have CallGraph::print sort the CallGraphNodes by
name before printing them.
Reviewers: bogner, chandlerc
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D10575
llvm-svn: 240191
Summary:
Currently intrinsics don't affect the creation of the call graph.
This is not accurate with respect to statepoint and patchpoint
intrinsics -- these do call (or invoke) LLVM level functions.
This change fixes this inconsistency by adding a call to the external
node for call sites that call these non-leaf intrinsics. This coupled
with the fact that these intrinsics also escape the function pointer
they call gives us a conservatively correct call graph.
Reviewers: reames, chandlerc, atrick, pgavlin
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D10526
llvm-svn: 240039
The personality routine currently lives in the LandingPadInst.
This isn't desirable because:
- All LandingPadInsts in the same function must have the same
personality routine. This means that each LandingPadInst beyond the
first has an operand which produces no additional information.
- There is ongoing work to introduce EH IR constructs other than
LandingPadInst. Moving the personality routine off of any one
particular Instruction and onto the parent function seems a lot better
than have N different places a personality function can sneak onto an
exceptional function.
Differential Revision: http://reviews.llvm.org/D10429
llvm-svn: 239940
Summary:
When propagating mass through irregular loops, the mass flowing through
each loop header may not be equal. This was causing wrong frequencies
to be computed for irregular loop headers.
Fixed by keeping track of masses flowing through each of the headers in
an irregular loop. To do this, we now keep track of per-header backedge
weights. After the loop mass is distributed through the loop, the
backedge weights are used to re-distribute the loop mass to the loop
headers.
Since each backedge will have a mass proportional to the different
branch weights, the loop headers will end up with a more approximate
weight distribution (as opposed to the current distribution that assumes
that every loop header is the same).
Reviewers: dexonsmith
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D10348
llvm-svn: 239843
Summary:
ValueTracking used to overwrite the analysis results computed from
assumes and dominating conditions. This patch fixes this issue.
Test Plan: test/Analysis/ValueTracking/assume.ll
Reviewers: hfinkel, majnemer
Reviewed By: majnemer
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D10283
llvm-svn: 239718
This patch ensures that SHL/SRL/SRA shifts for i8 and i16 vectors avoid scalarization. It builds on the existing i8 SHL vectorized implementation of moving the shift bits up to the sign bit position and separating the 4, 2 & 1 bit shifts with several improvements:
1 - SSE41 targets can use (v)pblendvb directly with the sign bit instead of performing a comparison to feed into a VSELECT node.
2 - pre-SSE41 targets were masking + comparing with an 0x80 constant - we avoid this by using the fact that a set sign bit means a negative integer which can be compared against zero to then feed into VSELECT, avoiding the need for a constant mask (zero generation is much cheaper).
3 - SRA i8 needs to be unpacked to the upper byte of a i16 so that the i16 psraw instruction can be correctly used for sign extension - we have to do more work than for SHL/SRL but perf tests indicate that this is still beneficial.
The i16 implementation is similar but simpler than for i8 - we have to do 8, 4, 2 & 1 bit shifts but less shift masking is involved. SSE41 use of (v)pblendvb requires that the i16 shift amount is splatted to both bytes however.
Tested on SSE2, SSE41 and AVX machines.
Differential Revision: http://reviews.llvm.org/D9474
llvm-svn: 239509
For GEP instructions isDereferenceablePointer checks that all indices are constant and within bounds. Replace this index calculation logic to a call to accumulateConstantOffset. Separated from the http://reviews.llvm.org/D9791
Reviewed By: sanjoy
Differential Revision: http://reviews.llvm.org/D9874
llvm-svn: 239299
Summary:
We need to add a runtime memcheck for pair of accesses (x,y) where at least one of x and y
are writes.
Assuming we have w writes and r reads, currently this number is estimated as being
w* (w+r-1). This estimation will count (write,write) pairs twice and will overestimate
the number of checks required.
This change adds a getNumberOfChecks method to RuntimePointerCheck, which
will count the number of runtime checks needed (similar in implementation to
needsAnyChecking) and uses it to produce the correct number of runtime checks.
Test Plan:
llvm test suite
spec2k
spec2k6
Performance results: no changes observed (not surprising since the formula for 1 writer is basically the same, which would covers most cases - at least with the current check limit).
Reviewers: anemet
Reviewed By: anemet
Subscribers: mzolotukhin, llvm-commits
Differential Revision: http://reviews.llvm.org/D10217
llvm-svn: 239295
Summary:
In continuation to an earlier commit to DependenceAnalysis.cpp by jingyue (r222100), the type for all subscripts in a coupled group need to be the same since constraints from one subscript may be propagated to another during testing. During testing, new SCEVs may be created and the operands for these need to be the same.
This patch extends unifySubscriptType() to work on lists of subscript pairs, ensuring a common extended type for all of them.
Test Plan:
Added a test case to NonCanonicalizedSubscript.ll which causes dependence analysis to crash without this fix.
All regression tests pass.
Reviewers: spop, sebpop, jingyue
Reviewed By: jingyue
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D9698
llvm-svn: 238573
