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
One of the testcases introduced by D9365 had incorrect !dereferenceable metadata on load. It must fail but it doesn't due to incorrect order of CHECK/CHECK-NOT commands in test. Fixed both.
Reviewed By: sanjoy
Differential Revision: http://reviews.llvm.org/D9877
llvm-svn: 237897
Summary:
Introduce dereferenceable, dereferenceable_or_null metadata for loads
with the same semantic as corresponding attributes.
This patch depends on http://reviews.llvm.org/D9253
Patch by Artur Pilipenko!
Reviewers: hfinkel, sanjoy, reames
Reviewed By: sanjoy, reames
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D9365
llvm-svn: 237720
When dependence analysis encounters a non-constant distance between
memory accesses it aborts the analysis and falls back to run-time checks
only. In this case we weren't resetting the array of dependences.
llvm-svn: 237574
"Store to invariant address..." is moved as the last line. This is not
the prime result of the analysis. Plus it simplifies some of the tests.
llvm-svn: 237573
collectUpperBound hits an assertion when the back edge count is wider then the desired type.
If that happens, truncate the backedge count.
Patch by Philip Pfaffe!
llvm-svn: 237439
Summary:
This change adds two new parameters to the statepoint intrinsic, `i64 id`
and `i32 num_patch_bytes`. `id` gets propagated to the ID field
in the generated StackMap section. If the `num_patch_bytes` is
non-zero then the statepoint is lowered to `num_patch_bytes` bytes of
nops instead of a call (the spill and reload code remains unchanged).
A non-zero `num_patch_bytes` is useful in situations where a language
runtime requires complete control over how a call is lowered.
This change brings statepoints one step closer to patchpoints. With
some additional work (that is not part of this patch) it should be
possible to get rid of `TargetOpcode::STATEPOINT` altogether.
PlaceSafepoints generates `statepoint` wrappers with `id` set to
`0xABCDEF00` (the old default value for the ID reported in the stackmap)
and `num_patch_bytes` set to `0`. This can be made more sophisticated
later.
Reviewers: reames, pgavlin, swaroop.sridhar, AndyAyers
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D9546
llvm-svn: 237214
Summary:
Now it's much easier to follow what's happening in this test.
Also removed some unused metadata entries.
Reviewers: hfinkel
Reviewed By: hfinkel
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D9601
llvm-svn: 236981
Summary:
There are several unhandled edge cases in BasicAA's GetLinearExpression
method. This changes fixes outstanding issues, including zext / sext of
a constant with the sign bit set, and the refusal to decompose zexts or
sexts of wrapping arithmetic.
Test Plan: Unit tests added in //q.ext.ll//.
Patch by Nick White.
Reviewers: hfinkel, sanjoy
Reviewed By: hfinkel, sanjoy
Subscribers: sanjoy, llvm-commits, hfinkel
Differential Revision: http://reviews.llvm.org/D6682
llvm-svn: 236894
This changes the shape of the statepoint intrinsic from:
@llvm.experimental.gc.statepoint(anyptr target, i32 # call args, i32 unused, ...call args, i32 # deopt args, ...deopt args, ...gc args)
to:
@llvm.experimental.gc.statepoint(anyptr target, i32 # call args, i32 flags, ...call args, i32 # transition args, ...transition args, i32 # deopt args, ...deopt args, ...gc args)
This extension offers the backend the opportunity to insert (somewhat) arbitrary code to manage the transition from GC-aware code to code that is not GC-aware and back.
In order to support the injection of transition code, this extension wraps the STATEPOINT ISD node generated by the usual lowering lowering with two additional nodes: GC_TRANSITION_START and GC_TRANSITION_END. The transition arguments that were passed passed to the intrinsic (if any) are lowered and provided as operands to these nodes and may be used by the backend during code generation.
Eventually, the lowering of the GC_TRANSITION_{START,END} nodes should be informed by the GC strategy in use for the function containing the intrinsic call; for now, these nodes are instead replaced with no-ops.
Differential Revision: http://reviews.llvm.org/D9501
llvm-svn: 236888
Summary:
This addresses PR 22718. When branch weights are too large, they were
being clamped to the range [1, MaxWeightForBB]. But this clamping is
only applied to edges that go outside the range, so it distorts the
relative branch probabilities.
This patch changes the weight calculation to scale every branch so the
relative probabilities are preserved. The scaling is done differently
now. First, all the branch weights are added up, and if the sum exceeds
32 bits, it computes an integer scale to bring all the weights within
the range.
The patch fixes an existing test that had slightly wrong branch
probabilities due to the previous clamping. It now gets branch weights
scaled accordingly.
Reviewers: dexonsmith
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D9442
llvm-svn: 236750
Specifically, if a pointer accesses different underlying objects in each
iteration, don't look through the phi node defining the pointer.
The motivating case is the underlyling-objects-2.ll testcase. Consider
the loop nest:
int **A;
for (i)
for (j)
A[i][j] = A[i-1][j] * B[j]
This loop is transformed by Load-PRE to stash away A[i] for the next
iteration of the outer loop:
Curr = A[0]; // Prev_0
for (i: 1..N) {
Prev = Curr; // Prev = PHI (Prev_0, Curr)
Curr = A[i];
for (j: 0..N)
Curr[j] = Prev[j] * B[j]
}
Since A[i] and A[i-1] are likely to be independent pointers,
getUnderlyingObjects should not assume that Curr and Prev share the same
underlying object in the inner loop.
If it did we would try to dependence-analyze Curr and Prev and the
analysis of the corresponding SCEVs would fail with non-constant
distance.
To fix this, the getUnderlyingObjects API is extended with an optional
LoopInfo parameter. This is effectively what controls whether we want
the above behavior or the original. Currently, I only changed to use
this approach for LoopAccessAnalysis.
The other testcase is to guard the opposite case where we do want to
look through the loop PHI. If we step through an array by incrementing
a pointer, the underlying object is the incoming value of the phi as the
loop is entered.
Fixes rdar://problem/19566729
llvm-svn: 235634
An assert was triggered when attempting to create a new SCEV
with operands of different types in the visitAddRecExpr. In this
test case, the operand types of the numerator and denominator
are different. The SCEV division code should generate a
conservative answer when this happens.
Differential Revision: http://reviews.llvm.org/D9021
llvm-svn: 235511
n/1 generates a quotient equal to n and a remainder of 0.
If this case is not recognized, then the SCEV divide() function
can return a remainder that is greater than or equal to the
denominator, which means the delinearized subscripts for the
test case will be incorrect.
Differential Revision: http://reviews.llvm.org/D9003
llvm-svn: 235311
See r230786 and r230794 for similar changes to gep and load
respectively.
Call is a bit different because it often doesn't have a single explicit
type - usually the type is deduced from the arguments, and just the
return type is explicit. In those cases there's no need to change the
IR.
When that's not the case, the IR usually contains the pointer type of
the first operand - but since typed pointers are going away, that
representation is insufficient so I'm just stripping the "pointerness"
of the explicit type away.
This does make the IR a bit weird - it /sort of/ reads like the type of
the first operand: "call void () %x(" but %x is actually of type "void
()*" and will eventually be just of type "ptr". But this seems not too
bad and I don't think it would benefit from repeating the type
("void (), void () * %x(" and then eventually "void (), ptr %x(") as has
been done with gep and load.
This also has a side benefit: since the explicit type is no longer a
pointer, there's no ambiguity between an explicit type and a function
that returns a function pointer. Previously this case needed an explicit
type (eg: a function returning a void() function was written as
"call void () () * @x(" rather than "call void () * @x(" because of the
ambiguity between a function returning a pointer to a void() function
and a function returning void).
No ambiguity means even function pointer return types can just be
written alone, without writing the whole function's type.
This leaves /only/ the varargs case where the explicit type is required.
Given the special type syntax in call instructions, the regex-fu used
for migration was a bit more involved in its own unique way (as every
one of these is) so here it is. Use it in conjunction with the apply.sh
script and associated find/xargs commands I've provided in rr230786 to
migrate your out of tree tests. Do let me know if any of this doesn't
cover your cases & we can iterate on a more general script/regexes to
help others with out of tree tests.
About 9 test cases couldn't be automatically migrated - half of those
were functions returning function pointers, where I just had to manually
delete the function argument types now that we didn't need an explicit
function type there. The other half were typedefs of function types used
in calls - just had to manually drop the * from those.
import fileinput
import sys
import re
pat = re.compile(r'((?:=|:|^|\s)call\s(?:[^@]*?))(\s*$|\s*(?:(?:\[\[[a-zA-Z0-9_]+\]\]|[@%](?:(")?[\\\?@a-zA-Z0-9_.]*?(?(3)"|)|{{.*}}))(?:\(|$)|undef|inttoptr|bitcast|null|asm).*$)')
addrspace_end = re.compile(r"addrspace\(\d+\)\s*\*$")
func_end = re.compile("(?:void.*|\)\s*)\*$")
def conv(match, line):
if not match or re.search(addrspace_end, match.group(1)) or not re.search(func_end, match.group(1)):
return line
return line[:match.start()] + match.group(1)[:match.group(1).rfind('*')].rstrip() + match.group(2) + line[match.end():]
for line in sys.stdin:
sys.stdout.write(conv(re.search(pat, line), line))
llvm-svn: 235145
This commit makes LLVM not estimate branch probabilities when doing a
single bit bitmask tests.
The code that originally made me discover this is:
if ((a & 0x1) == 0x1) {
..
}
In this case we don't actually have any branch probability information
and should not assume to have any. LLVM transforms this into:
%and = and i32 %a, 1
%tobool = icmp eq i32 %and, 0
So, in this case, the result of a bitwise and is compared against 0,
but nevertheless, we should not assume to have probability
information.
CodeGen/ARM/2013-10-11-select-stalls.ll started failing because the
changed probabilities changed the results of
ARMBaseInstrInfo::isProfitableToIfCvt() and led to an Ifcvt of the
diamond in the test. AFAICT, the test was never meant to test this and
thus changing the test input slightly to not change the probabilities
seems like the best way to preserve the meaning of the test.
llvm-svn: 234979
if ((a & 0x1) == 0x1) {
..
}
In this case we don't actually have any branch probability information and
should not assume to have any. LLVM transforms this into:
%and = and i32 %a, 1
%tobool = icmp eq i32 %and, 0
So, in this case, the result of a bitwise and is compared against 0,
but nevertheless, we should not assume to have probability
information.
llvm-svn: 234898
Fix oversight in -analyze output. PtrRtCheck contains the pointers that
need to be checked against each other and not whether memchecks are
necessary.
For instance in the testcase PtrRtCheck has four elements but all
no-alias so no checking is necessary.
llvm-svn: 234833
Summary:
Some optimizations such as jump threading and loop unswitching can negatively
affect performance when applied to divergent branches. The divergence analysis
added in this patch conservatively estimates which branches in a GPU program
can diverge. This information can then help LLVM to run certain optimizations
selectively.
Test Plan: test/Analysis/DivergenceAnalysis/NVPTX/diverge.ll
Reviewers: resistor, hfinkel, eliben, meheff, jholewinski
Subscribers: broune, bjarke.roune, madhur13490, tstellarAMD, dberlin, echristo, jholewinski, llvm-commits
Differential Revision: http://reviews.llvm.org/D8576
llvm-svn: 234567
(Re-apply r234361 with a fix and a testcase for PR23157)
Both run-time pointer checking and the dependence analysis are capable
of dealing with uniform addresses. I.e. it's really just an orthogonal
property of the loop that the analysis computes.
Run-time pointer checking will only try to reason about SCEVAddRec
pointers or else gives up. If the uniform pointer turns out the be a
SCEVAddRec in an outer loop, the run-time checks generated will be
correct (start and end bounds would be equal).
In case of the dependence analysis, we work again with SCEVs. When
compared against a loop-dependent address of the same underlying object,
the difference of the two SCEVs won't be constant. This will result in
returning an Unknown dependence for the pair.
When compared against another uniform access, the difference would be
constant and we should return the right type of dependence
(forward/backward/etc).
The changes also adds support to query this property of the loop and
modify the vectorizer to use this.
Patch by Ashutosh Nema!
llvm-svn: 234424
Both run-time pointer checking and the dependence analysis are capable
of dealing with uniform addresses. I.e. it's really just an orthogonal
property of the loop that the analysis computes.
Run-time pointer checking will only try to reason about SCEVAddRec
pointers or else gives up. If the uniform pointer turns out the be a
SCEVAddRec in an outer loop, the run-time checks generated will be
correct (start and end bounds would be equal).
In case of the dependence analysis, we work again with SCEVs. When
compared against a loop-dependent address of the same underlying object,
the difference of the two SCEVs won't be constant. This will result in
returning an Unknown dependence for the pair.
When compared against another uniform access, the difference would be
constant and we should return the right type of dependence
(forward/backward/etc).
The changes also adds support to query this property of the loop and
modify the vectorizer to use this.
Patch by Ashutosh Nema!
llvm-svn: 234361
Summary:
This change teaches ScalarEvolution::isLoopBackedgeGuardedByCond to look
at edges within the loop body that dominate the latch. We don't do an
exhaustive search for all possible edges, but only a quick walk up the
dom tree.
This re-lands r233447. r233447 was reverted because it caused massive
compile-time regressions. This change has a fix for the same issue.
llvm-svn: 233829
Summary:
This is part 1 of fixes to address the problems described in
https://llvm.org/bugs/show_bug.cgi?id=22719.
The restriction to limit loop scales to 4,096 does not really prevent
overflows anymore, as the underlying algorithm has changed and does
not seem to suffer from this problem.
Additionally, artificially restricting loop scales to such a low number
skews frequency information, making loops of equal hotness appear to
have very different hotness properties.
The only loops that are artificially restricted to a scale of 4096 are
infinite loops (those loops with an exit mass of 0). This prevents
infinite loops from skewing the frequencies of other regions in the CFG.
At the end of propagation, frequencies are scaled to values that take no
more than 64 bits to represent. When the range of frequencies to be
represented fits within 61 bits, it pushes up the scaling factor to a
minimum of 8 to better distinguish small frequency values. Otherwise,
small frequency values are all saturated down at 1.
Tested on x86_64.
Reviewers: dexonsmith
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D8718
llvm-svn: 233826
Summary:
This change teaches ScalarEvolution::isLoopBackedgeGuardedByCond to look
at edges within the loop body that dominate the latch. We don't do an
exhaustive search for all possible edges, but only a quick walk up the
dom tree.
Reviewers: atrick, hfinkel
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D8627
llvm-svn: 233447
This was discussed a while back and I left it optional for migration. Since it's been far more than the 'week or two' that was discussed, time to actually make this manditory.
llvm-svn: 233357
Summary:
With the introduction of MarkPendingLoopPredicates in r157092, I don't
think the bailout is needed anymore.
Reviewers: atrick, nicholas
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D8624
llvm-svn: 233296
Summary:
`ComputeNumSignBits` returns incorrect results for `srem` instructions.
This change fixes the issue and adds a test case.
Reviewers: nadav, nicholas, atrick
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D8600
llvm-svn: 233225
Summary:
This change teaches isImpliedCond to infer things like "X sgt 0" => "X -
1 sgt -1". The `ConstantRange` class has the logic to do the heavy
lifting, this change simply gets ScalarEvolution to exploit that when
reasonable.
Depends on D8345
Reviewers: atrick
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D8346
llvm-svn: 232576
Summary:
ScalarEvolutionExpander assumes that the header block of a loop is a
legal place to have a use for a phi node. This is true only for phis
that are either in the header or dominate the header block, but it is
not true for phi nodes that are strictly internal to the loop body.
This change teaches ScalarEvolutionExpander to place uses of PHI nodes
in the basic block the PHI nodes belong to. This is always legal, and
`hoistIVInc` ensures that the said position dominates `IsomorphicInc`.
Reviewers: atrick
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D8311
llvm-svn: 232189
Similar to gep (r230786) and load (r230794) changes.
Similar migration script can be used to update test cases, which
successfully migrated all of LLVM and Polly, but about 4 test cases
needed manually changes in Clang.
(this script will read the contents of stdin and massage it into stdout
- wrap it in the 'apply.sh' script shown in previous commits + xargs to
apply it over a large set of test cases)
import fileinput
import sys
import re
rep = re.compile(r"(getelementptr(?:\s+inbounds)?\s*\()((<\d*\s+x\s+)?([^@]*?)(|\s*addrspace\(\d+\))\s*\*(?(3)>)\s*)(?=$|%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|zeroinitializer|<|\[\[[a-zA-Z]|\{\{)", re.MULTILINE | re.DOTALL)
def conv(match):
line = match.group(1)
line += match.group(4)
line += ", "
line += match.group(2)
return line
line = sys.stdin.read()
off = 0
for match in re.finditer(rep, line):
sys.stdout.write(line[off:match.start()])
sys.stdout.write(conv(match))
off = match.end()
sys.stdout.write(line[off:])
llvm-svn: 232184
There's a missed optimization opportunity where we could look at the full chain of computation and take the intersection of the flags instead of only looking one instruction deep.
llvm-svn: 232134
The dependences are now expose through the new getInterestingDependences
API so we can use that with -analyze too and fix the FIXME.
This lets us remove the test that relied on -debug to check the
dependences.
llvm-svn: 231807
This crash occurs due to memory corruption when trying to update dependency
direction based on Constraints.
This crash was observed during lnt regression of Polybench benchmark test case dynprog.
Review: http://reviews.llvm.org/D8059
llvm-svn: 231788
This crash in Dependency analysis is because we assume here that in case of UsefulGEP
both source and destination have the same number of operands which may not be true.
This incorrect assumption results in crash while populating Pairs. Fix the same.
This crash was observed during lnt regression for code such as-
struct s{
int A[10][10];
int C[10][10][10];
} S;
void dep_constraint_crash_test(int k,int N) {
for( int i=0;i<N;i++)
for( int j=0;j<N;j++)
S.A[0][0] = S.C[0][0][k];
}
Review: http://reviews.llvm.org/D8162
llvm-svn: 231784
CFLAA didn't know how to properly handle ConstantExprs; it would silently
ignore them. This was a problem if the ConstantExpr is, say, a GEP of a global,
because CFLAA wouldn't realize that there's a global there. :)
llvm-svn: 231743
We now treat pointers given to ptrtoint and pointers retrieved from
inttoptr as similar to arguments or globals (can alias anything, etc.)
This solves some of the problems we were having with giving incorrect
results.
llvm-svn: 231741
Summary:
This removes some duplicated code, and also helps optimization: e.g. in
the test case added, `%idx ULT 128` in `@x` is not currently optimized
to `true` by `-indvars` but will be, after this change.
The only functional change in ths commit is that for add recurrences,
ScalarEvolution::getRange will be more aggressive -- computing the
unsigned (resp. signed) range for a SCEVAddRecExpr will now look at the
NSW (resp. NUW) bits and check for signed (resp. unsigned) overflow.
This can be a strict improvement in some cases (such as the attached
test case), and should be no worse in other cases.
Reviewers: atrick, nlewycky
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D8142
llvm-svn: 231709
Summary:
Unused in this commit, but will be used in a subsequent change (D8142)
by a FileCheck test.
Reviewers: atrick
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D8143
llvm-svn: 231708
Summary:
Teach SCEV to prove no overflow for an add recurrence by proving
something about the range of another add recurrence a loop-invariant
distance away from it.
Reviewers: atrick, hfinkel
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D7980
llvm-svn: 231305
Summary:
DataLayout keeps the string used for its creation.
As a side effect it is no longer needed in the Module.
This is "almost" NFC, the string is no longer
canonicalized, you can't rely on two "equals" DataLayout
having the same string returned by getStringRepresentation().
Get rid of DataLayoutPass: the DataLayout is in the Module
The DataLayout is "per-module", let's enforce this by not
duplicating it more than necessary.
One more step toward non-optionality of the DataLayout in the
module.
Make DataLayout Non-Optional in the Module
Module->getDataLayout() will never returns nullptr anymore.
Reviewers: echristo
Subscribers: resistor, llvm-commits, jholewinski
Differential Revision: http://reviews.llvm.org/D7992
From: Mehdi Amini <mehdi.amini@apple.com>
llvm-svn: 231270
Ultimately, __CxxFrameHandler3 needs us to put a stack offset in a
table, and it will take responsibility for copying the exception object
into that slot. Modelling the exception object as an SSA value returned
by begincatch isn't going to work in general, so make it use an output
parameter.
Reviewers: andrew.w.kaylor
Differential Revision: http://reviews.llvm.org/D7920
llvm-svn: 231086
Essentially the same as the GEP change in r230786.
A similar migration script can be used to update test cases, though a few more
test case improvements/changes were required this time around: (r229269-r229278)
import fileinput
import sys
import re
pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)")
for line in sys.stdin:
sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line))
Reviewers: rafael, dexonsmith, grosser
Differential Revision: http://reviews.llvm.org/D7649
llvm-svn: 230794
One of several parallel first steps to remove the target type of pointers,
replacing them with a single opaque pointer type.
This adds an explicit type parameter to the gep instruction so that when the
first parameter becomes an opaque pointer type, the type to gep through is
still available to the instructions.
* This doesn't modify gep operators, only instructions (operators will be
handled separately)
* Textual IR changes only. Bitcode (including upgrade) and changing the
in-memory representation will be in separate changes.
* geps of vectors are transformed as:
getelementptr <4 x float*> %x, ...
->getelementptr float, <4 x float*> %x, ...
Then, once the opaque pointer type is introduced, this will ultimately look
like:
getelementptr float, <4 x ptr> %x
with the unambiguous interpretation that it is a vector of pointers to float.
* address spaces remain on the pointer, not the type:
getelementptr float addrspace(1)* %x
->getelementptr float, float addrspace(1)* %x
Then, eventually:
getelementptr float, ptr addrspace(1) %x
Importantly, the massive amount of test case churn has been automated by
same crappy python code. I had to manually update a few test cases that
wouldn't fit the script's model (r228970,r229196,r229197,r229198). The
python script just massages stdin and writes the result to stdout, I
then wrapped that in a shell script to handle replacing files, then
using the usual find+xargs to migrate all the files.
update.py:
import fileinput
import sys
import re
ibrep = re.compile(r"(^.*?[^%\w]getelementptr inbounds )(((?:<\d* x )?)(.*?)(| addrspace\(\d\)) *\*(|>)(?:$| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$))")
normrep = re.compile( r"(^.*?[^%\w]getelementptr )(((?:<\d* x )?)(.*?)(| addrspace\(\d\)) *\*(|>)(?:$| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$))")
def conv(match, line):
if not match:
return line
line = match.groups()[0]
if len(match.groups()[5]) == 0:
line += match.groups()[2]
line += match.groups()[3]
line += ", "
line += match.groups()[1]
line += "\n"
return line
for line in sys.stdin:
if line.find("getelementptr ") == line.find("getelementptr inbounds"):
if line.find("getelementptr inbounds") != line.find("getelementptr inbounds ("):
line = conv(re.match(ibrep, line), line)
elif line.find("getelementptr ") != line.find("getelementptr ("):
line = conv(re.match(normrep, line), line)
sys.stdout.write(line)
apply.sh:
for name in "$@"
do
python3 `dirname "$0"`/update.py < "$name" > "$name.tmp" && mv "$name.tmp" "$name"
rm -f "$name.tmp"
done
The actual commands:
From llvm/src:
find test/ -name *.ll | xargs ./apply.sh
From llvm/src/tools/clang:
find test/ -name *.mm -o -name *.m -o -name *.cpp -o -name *.c | xargs -I '{}' ../../apply.sh "{}"
From llvm/src/tools/polly:
find test/ -name *.ll | xargs ./apply.sh
After that, check-all (with llvm, clang, clang-tools-extra, lld,
compiler-rt, and polly all checked out).
The extra 'rm' in the apply.sh script is due to a few files in clang's test
suite using interesting unicode stuff that my python script was throwing
exceptions on. None of those files needed to be migrated, so it seemed
sufficient to ignore those cases.
Reviewers: rafael, dexonsmith, grosser
Differential Revision: http://reviews.llvm.org/D7636
llvm-svn: 230786
accesses are via different types
Noticed this while generalizing the code for loop distribution.
I confirmed with Arnold that this was indeed a bug and managed to create
a testcase.
llvm-svn: 230647
(The change was landed in r230280 and caused the regression PR22674.
This version contains a fix and a test-case for PR22674).
When emitting the increment operation, SCEVExpander marks the
operation as nuw or nsw based on the flags on the preincrement SCEV.
This is incorrect because, for instance, it is possible that {-6,+,1}
is <nuw> while {-6,+,1}+1 = {-5,+,1} is not.
This change teaches SCEV to mark the increment as nuw/nsw only if it
can explicitly prove that the increment operation won't overflow.
Apart from the attached test case, another (more realistic)
manifestation of the bug can be seen in
Transforms/IndVarSimplify/pr20680.ll.
Differential Revision: http://reviews.llvm.org/D7778
llvm-svn: 230533
The bug was a result of getPreStartForExtend interpreting nsw/nuw
flags on an add recurrence more strongly than is legal. {S,+,X}<nsw>
implies S+X is nsw only if the backedge of the loop is taken at least
once.
NOTE: I had accidentally committed an unrelated change with the commit
message of this change in r230275 (r230275 was reverted in r230279).
This is the correct change for this commit message.
Differential Revision: http://reviews.llvm.org/D7808
llvm-svn: 230291
When emitting the increment operation, SCEVExpander marks the
operation as nuw or nsw based on the flags on the preincrement SCEV.
This is incorrect because, for instance, it is possible that {-6,+,1}
is <nuw> while {-6,+,1}+1 = {-5,+,1} is not.
This change teaches SCEV to mark the increment as nuw/nsw only if it
can explicitly prove that the increment operation won't overflow.
Apart from the attached test case, another (more realistic) manifestation
of the bug can be seen in Transforms/IndVarSimplify/pr20680.ll.
NOTE: this change was landed with an incorrect commit message in
rL230275 and was reverted for that reason in rL230279. This commit
message is the correct one.
Differential Revision: http://reviews.llvm.org/D7778
llvm-svn: 230280
230275 got committed with an incorrect commit message due to a mixup
on my side. Will re-land in a few moments with the correct commit
message.
llvm-svn: 230279
The bug was a result of getPreStartForExtend interpreting nsw/nuw
flags on an add recurrence more strongly than is legal. {S,+,X}<nsw>
implies S+X is nsw only if the backedge of the loop is taken at least
once.
Differential Revision: http://reviews.llvm.org/D7808
llvm-svn: 230275
The LoopInfo in combination with depth_first is used to enumerate the
loops.
Right now -analyze is not yet complete. It only prints the result of
the analysis, the report and the run-time checks. Printing the unsafe
depedences will require a bit more reshuffling which I'd like to do in a
follow-on to this patchset. Unsafe dependences are currently checked
via -debug-only=loop-accesses in the new test.
This is part of the patchset that converts LoopAccessAnalysis into an
actual analysis pass.
llvm-svn: 229898
First, don't combine bit masking into vector shuffles (even ones the
target can handle) once operation legalization has taken place. Custom
legalization of vector shuffles may exist for these patterns (making the
predicate return true) but that custom legalization may in some cases
produce the exact bit math this matches. We only really want to handle
this prior to operation legalization.
However, the x86 backend, in a fit of awesome, relied on this. What it
would do is mark VSELECTs as expand, which would turn them into
arithmetic, which this would then match back into vector shuffles, which
we would then lower properly. Amazing.
Instead, the second change is to teach the x86 backend to directly form
vector shuffles from VSELECT nodes with constant conditions, and to mark
all of the vector types we support lowering blends as shuffles as custom
VSELECT lowering. We still mark the forms which actually support
variable blends as *legal* so that the custom lowering is bypassed, and
the legal lowering can even be used by the vector shuffle legalization
(yes, i know, this is confusing. but that's how the patterns are
written).
This makes the VSELECT lowering much more sensible, and in fact should
fix a bunch of bugs with it. However, as you'll see in the test cases,
right now what it does is point out the *hilarious* deficiency of the
new vector shuffle lowering when it comes to blends. Fortunately, my
very next patch fixes that. I can't submit it yet, because that patch,
somewhat obviously, forms the exact and/or pattern that the DAG combine
is matching here! Without this patch, teaching the vector shuffle
lowering to produce the right code infloops in the DAG combiner. With
this patch alone, we produce terrible code but at least lower through
the right paths. With both patches, all the regressions here should be
fixed, and a bunch of the improvements (like using 2 shufps with no
memory loads instead of 2 andps with memory loads and an orps) will
stay. Win!
There is one other change worth noting here. We had hilariously wrong
vectorization cost estimates for vselect because we fell through to the
code path that assumed all "expand" vector operations are scalarized.
However, the "expand" lowering of VSELECT is vector bit math, most
definitely not scalarized. So now we go back to the correct if horribly
naive cost of "1" for "not scalarized". If anyone wants to add actual
modeling of shuffle costs, that would be cool, but this seems an
improvement on its own. Note the removal of 16 and 32 "costs" for doing
a blend. Even in SSE2 we can blend in fewer than 16 instructions. ;] Of
course, we don't right now because of OMG bad code, but I'm going to fix
that. Next patch. I promise.
llvm-svn: 229835
r229622: "[LoopAccesses] Make VectorizerParams global"
r229623: "[LoopAccesses] Stash the report from the analysis rather than emitting it"
r229624: "[LoopAccesses] Cache the result of canVectorizeMemory"
r229626: "[LoopAccesses] Create the analysis pass"
r229628: "[LoopAccesses] Change debug messages from LV to LAA"
r229630: "[LoopAccesses] Add canAnalyzeLoop"
r229631: "[LoopAccesses] Add missing const to APIs in VectorizationReport"
r229632: "[LoopAccesses] Split out LoopAccessReport from VectorizerReport"
r229633: "[LoopAccesses] Add -analyze support"
r229634: "[LoopAccesses] Change LAA:getInfo to return a constant reference"
r229638: "Analysis: fix buildbots"
llvm-svn: 229650
The LoopInfo in combination with depth_first is used to enumerate the
loops.
Right now -analyze is not yet complete. It only prints the result of
the analysis, the report and the run-time checks. Printing the unsafe
depedences will require a bit more reshuffling which I'd like to do in a
follow-on to this patchset. Unsafe dependences are currently checked
via -debug-only=loop-accesses in the new test.
This is part of the patchset that converts LoopAccessAnalysis into an
actual analysis pass.
llvm-svn: 229633
extensions.
This change also removes `DEBUG(dbgs() << "SCEV: untested prestart
overflow check\n");` because that case has a unit test now.
Differential Revision: http://reviews.llvm.org/D7645
llvm-svn: 229600
We would crash if we couldn't locate a Function that either Location's
Value belonged to. Now we just print out a debug message and return
conservatively.
llvm-svn: 228901
Without a valid data layout, deferenceable(N) doesn't get parsed or
propagated. Since this is the key item we are testing, add a dependency
on the pass.
Differential Revision: http://reviews.llvm.org/D7508
llvm-svn: 228611
While a theoretical GC might change dereferenceability on collection,
there is no such known collector and no need to account for the case
with a flag yet.
Differential Revision: http://reviews.llvm.org/D7454
llvm-svn: 228606
When creating a scev for sext({X,+,Y}), scev checks if the expression
is equivalent to {sext X,+,zext Y}. If it can prove that, it also
tags the original {X,+,Y} as <nsw>, which is not correct.
In the test case I run `-scalar-evolution` twice because the bug
manifests only once SCEV has run through and seen the `sext`
expressions (and then does a in-place mutation on {X,+,Y}).
Differential Revision: http://reviews.llvm.org/D7495
llvm-svn: 228586
For the attached test case different types are used in the ICmpInst
and SelectInst that represent the min/max expressions. However, if the
ICmpInst type is smaller a comparison with the sign/zero extended
operands would have yielded the same result. This situation might
arise after the instruction combination pass was applied.
Differential Revision: http://reviews.llvm.org/D7338
llvm-svn: 228572
add recurrences don't overflow.
This change makes the optimization more restrictive. It still assumes
that an overflowing `add nsw` is undefined behavior; and this change
will need revisiting once we have a consistent semantics for poison
values.
Differential Revision: http://reviews.llvm.org/D7331
llvm-svn: 228552
different fields.
We can show that two GEPs off of the same (possibly multidimensional)
array of structs, into different fields, can't alias. Quoting:
For two GEPOperators GEP1 and GEP2, if we find that:
- both GEPs begin indexing from the exact same pointer;
- the last indices in both GEPs are constants, indexing into a struct;
- said indices are different, hence,the pointed-to fields are different;
- and both GEPs only index through arrays prior to that;
this lets us determine that the struct that GEP1 indexes into and the
struct that GEP2 indexes into must either precisely overlap or be
completely disjoint. Because they cannot partially overlap, indexing
into different non-overlapping fields of the struct will never alias.
The other BasicAA::aliasGEP rules worked in some cases, but not all
(for example, the i32x3 struct in the testcase).
We can add this simple ad-hoc rule to complement them.
rdar://19717375
Differential Revision: http://reviews.llvm.org/D7453
llvm-svn: 228498
Since testing the function indirectly is tricky, introduce a direct
print-memderefs pass, in the same spirit as print-memdeps, which prints
dereferenceability information matched by FileCheck.
Differential Revision: http://reviews.llvm.org/D7075
llvm-svn: 228369
Fixes PR22462: two of the tests have regressed for a while,
but were using CHECK-NOT to match "May:". The actual output
was changed to "MayAlias:" at some point, which made the tests
useless.
Two others return MayAlias only because of a lack of analysis;
BasicAA returns PartialAlias in those cases, when a datalayout
is present.
llvm-svn: 228346
type erased interface and a single analysis pass rather than an
extremely complex analysis group.
The end result is that the TTI analysis can contain a type erased
implementation that supports the polymorphic TTI interface. We can build
one from a target-specific implementation or from a dummy one in the IR.
I've also factored all of the code into "mix-in"-able base classes,
including CRTP base classes to facilitate calling back up to the most
specialized form when delegating horizontally across the surface. These
aren't as clean as I would like and I'm planning to work on cleaning
some of this up, but I wanted to start by putting into the right form.
There are a number of reasons for this change, and this particular
design. The first and foremost reason is that an analysis group is
complete overkill, and the chaining delegation strategy was so opaque,
confusing, and high overhead that TTI was suffering greatly for it.
Several of the TTI functions had failed to be implemented in all places
because of the chaining-based delegation making there be no checking of
this. A few other functions were implemented with incorrect delegation.
The message to me was very clear working on this -- the delegation and
analysis group structure was too confusing to be useful here.
The other reason of course is that this is *much* more natural fit for
the new pass manager. This will lay the ground work for a type-erased
per-function info object that can look up the correct subtarget and even
cache it.
Yet another benefit is that this will significantly simplify the
interaction of the pass managers and the TargetMachine. See the future
work below.
The downside of this change is that it is very, very verbose. I'm going
to work to improve that, but it is somewhat an implementation necessity
in C++ to do type erasure. =/ I discussed this design really extensively
with Eric and Hal prior to going down this path, and afterward showed
them the result. No one was really thrilled with it, but there doesn't
seem to be a substantially better alternative. Using a base class and
virtual method dispatch would make the code much shorter, but as
discussed in the update to the programmer's manual and elsewhere,
a polymorphic interface feels like the more principled approach even if
this is perhaps the least compelling example of it. ;]
Ultimately, there is still a lot more to be done here, but this was the
huge chunk that I couldn't really split things out of because this was
the interface change to TTI. I've tried to minimize all the other parts
of this. The follow up work should include at least:
1) Improving the TargetMachine interface by having it directly return
a TTI object. Because we have a non-pass object with value semantics
and an internal type erasure mechanism, we can narrow the interface
of the TargetMachine to *just* do what we need: build and return
a TTI object that we can then insert into the pass pipeline.
2) Make the TTI object be fully specialized for a particular function.
This will include splitting off a minimal form of it which is
sufficient for the inliner and the old pass manager.
3) Add a new pass manager analysis which produces TTI objects from the
target machine for each function. This may actually be done as part
of #2 in order to use the new analysis to implement #2.
4) Work on narrowing the API between TTI and the targets so that it is
easier to understand and less verbose to type erase.
5) Work on narrowing the API between TTI and its clients so that it is
easier to understand and less verbose to forward.
6) Try to improve the CRTP-based delegation. I feel like this code is
just a bit messy and exacerbating the complexity of implementing
the TTI in each target.
Many thanks to Eric and Hal for their help here. I ended up blocked on
this somewhat more abruptly than I expected, and so I appreciate getting
it sorted out very quickly.
Differential Revision: http://reviews.llvm.org/D7293
llvm-svn: 227669
I had already factored this analysis specifically to enable doing this,
but hadn't actually committed the necessary wiring to get at this from
the new pass manager. This also nicely shows how the separate cache
object can be directly managed by the new pass manager.
This analysis didn't have any direct tests and so I've added a printer
pass and a boring test case. I chose to print the i1 value which is
being assumed rather than the call to llvm.assume as that seems much
more useful for testing... but suggestions on an even better printing
strategy welcome. My main goal was to make sure things actually work. =]
llvm-svn: 226868
ScalarEvolution currently lowers a subtraction recurrence to an add
recurrence with the same no-wrap flags as the subtraction. This is
incorrect because `sub nsw X, Y` is not the same as `add nsw X, -Y`
and `sub nuw X, Y` is not the same as `add nuw X, -Y`. This patch
fixes the issue, and adds two test cases demonstrating the bug.
Differential Revision: http://reviews.llvm.org/D7081
llvm-svn: 226755
pass and a LoopPrinterPass with the expected associated wiring.
I've added a RUN line to the only test case (!!!) we have that actually
prints loops. Everything seems to be working.
This is somewhat exciting as this is the first analysis using another
analysis to go in for the new pass manager. =D I also believe it is the
last analysis necessary for porting instcombine, but of course I may yet
discover more.
llvm-svn: 226560
This adds the domtree analysis to the new pass manager. The analysis
returns the same DominatorTree result entity used by the old pass
manager and essentially all of the code is shared. We just have
different boilerplate for running and printing the analysis.
I've converted one test to run in both modes just to make sure this is
exercised while both are live in the tree.
llvm-svn: 225969
Correct, we have *zero* basic testing of the dominator tree in the
regression test suite. There is a single test that even prints it out,
and that test only checks a single line of the output. There are
a handful of tests that check post dominators, but all of those are
looking for bugs rather than just exercising the basic machinery.
This test is super boring and unexciting. But hey, it's something.
I needed there to be something so I could switch the basic test to run
with both the old and new pass manager.
llvm-svn: 225936
doing Load PRE"
It's not really expected to stick around, last time it provoked a weird LTO
build failure that I can't reproduce now, and the bot logs are long gone. I'll
re-revert it if the failures recur.
Original description: Perform Scalar PRE on gep indices that feed loads before
doing Load PRE.
llvm-svn: 225536
Now that `Metadata` is typeless, reflect that in the assembly. These
are the matching assembly changes for the metadata/value split in
r223802.
- Only use the `metadata` type when referencing metadata from a call
intrinsic -- i.e., only when it's used as a `Value`.
- Stop pretending that `ValueAsMetadata` is wrapped in an `MDNode`
when referencing it from call intrinsics.
So, assembly like this:
define @foo(i32 %v) {
call void @llvm.foo(metadata !{i32 %v}, metadata !0)
call void @llvm.foo(metadata !{i32 7}, metadata !0)
call void @llvm.foo(metadata !1, metadata !0)
call void @llvm.foo(metadata !3, metadata !0)
call void @llvm.foo(metadata !{metadata !3}, metadata !0)
ret void, !bar !2
}
!0 = metadata !{metadata !2}
!1 = metadata !{i32* @global}
!2 = metadata !{metadata !3}
!3 = metadata !{}
turns into this:
define @foo(i32 %v) {
call void @llvm.foo(metadata i32 %v, metadata !0)
call void @llvm.foo(metadata i32 7, metadata !0)
call void @llvm.foo(metadata i32* @global, metadata !0)
call void @llvm.foo(metadata !3, metadata !0)
call void @llvm.foo(metadata !{!3}, metadata !0)
ret void, !bar !2
}
!0 = !{!2}
!1 = !{i32* @global}
!2 = !{!3}
!3 = !{}
I wrote an upgrade script that handled almost all of the tests in llvm
and many of the tests in cfe (even handling many `CHECK` lines). I've
attached it (or will attach it in a moment if you're speedy) to PR21532
to help everyone update their out-of-tree testcases.
This is part of PR21532.
llvm-svn: 224257
When a loop gets bundled up, its outgoing edges are quite large, and can
just barely overflow 64-bits. If one successor has multiple incoming
edges -- and that successor is getting all the incoming mass --
combining just its edges can overflow. Handle that by saturating rather
than asserting.
This fixes PR21622.
llvm-svn: 223500
Summary:
Several places in DependenceAnalysis assumes both SCEVs in a subscript pair
share the same integer type. For instance, isKnownPredicate calls
SE->getMinusSCEV(X, Y) which asserts X and Y share the same type. However,
DependenceAnalysis fails to ensure this assumption when producing a subscript
pair, causing tests such as NonCanonicalizedSubscript to crash. With this
patch, DependenceAnalysis runs unifySubscriptType before producing any
subscript pair, ensuring the assumption.
Test Plan:
Added NonCanonicalizedSubscript.ll on which DependenceAnalysis before the fix
crashed because subscripts have different types.
Reviewers: spop, sebpop, jingyue
Reviewed By: jingyue
Subscribers: eliben, meheff, llvm-commits
Differential Revision: http://reviews.llvm.org/D6289
llvm-svn: 222100
HowFarToZero was supposed to use unsigned division in order to calculate
the backedge taken count. However, SCEVDivision::divide performs signed
division. Unless I am mistaken, no users of SCEVDivision actually want
signed arithmetic: switch to udiv and urem.
This fixes PR21578.
llvm-svn: 222093
doing Load PRE"
This commit updates the failing test in
Analysis/TypeBasedAliasAnalysis/gvn-nonlocal-type-mismatch.ll
The failing test is sensitive to the order in which we process loads. This
version turns on the RPO traversal instead of the while DT traversal in GVN.
The new test code is functionally same just the order of loads that are
eliminated is swapped.
This new version also fixes an issue where GVN splits a critical edge and
potentially invalidate the RPO/DT iterator.
llvm-svn: 222039
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!
llvm-svn: 221876
AVX2 is available.
According to IACA, the new lowering has a throughput of 8 cycles instead of 13
with the previous one.
Althought this lowering kicks in some SPECs benchmarks, the performance
improvement was within the noise.
Correctness testing has been done for the whole range of uint32_t with the
following program:
uint4 v = (uint4) {0,1,2,3};
uint32_t i;
//Check correctness over entire range for uint4 -> float4 conversion
for( i = 0; i < 1U << (32-2); i++ )
{
float4 t = test(v);
float4 c = correct(v);
if( 0xf != _mm_movemask_ps( t == c ))
{
printf( "Error @ %vx: %vf vs. %vf\n", v, c, t);
return -1;
}
v += 4;
}
Where "correct" is the old lowering and "test" the new one.
The patch adds a test case for the two custom lowering instruction.
It also modifies the vector cost model, which is why cast.ll and uitofp.ll are
modified.
2009-02-26-MachineLICMBug.ll is also modified because we now hoist 7
instructions instead of 4 (3 more constant loads).
rdar://problem/18153096>
llvm-svn: 221657
In a case where we have a no {un,}signed wrap flag on the increment, if
RHS - Start is constant then we can avoid inserting a max operation bewteen
the two, since we can statically determine which is greater.
This allows us to unroll loops such as:
void testcase3(int v) {
for (int i=v; i<=v+1; ++i)
f(i);
}
llvm-svn: 220960
DSE's overlap checking contained special logic, used only when no DataLayout
was available, which inferred a complete overwrite when the pointee types were
equal. This logic seems fine for regular loads/stores, but does not work for
memcpy and friends. Instead of fixing this, I'm just removing it.
Philosophically, transformations should not contain enhanced behavior used only
when data layout is lacking (data layout should be strictly additive), and
maintaining these rarely-tested code paths seems not worthwhile at this stage.
Credit to Aliaksei Zasenka for the bug report and the diagnosis. The test case
(slightly reduced from that provided by Aliaksei) replaces the original
contents of test/Transforms/DeadStoreElimination/no-targetdata.ll -- a few
other tests have been updated to have a data layout.
llvm-svn: 220035
The CFL-AA implementation was missing a visit* routine for va_arg instructions,
causing it to assert when run on a function that had one. For now, handle these
in a conservative way.
Fixes PR20954.
llvm-svn: 219718
It also makes it more aggressive in querying range information by
adding a call to isKnownPredicateWithRanges to
isLoopBackedgeGuardedByCond and isLoopEntryGuardedByCond.
phabricator: http://reviews.llvm.org/D5638
Reviewed by: atrick, hfinkel
llvm-svn: 219532
ScalarEvolution in the presence of multiple exits. Previously all
loops exits had to have identical counts for a loop trip count to be
considered computable. This pessimization was implemented by calling
getBackedgeTakenCount(L) rather than getExitCount(L, ExitingBlock)
inside of ScalarEvolution::getSmallConstantTripCount() (see the FIXME
in the comments of that function). The pessimization was added to fix
a corner case involving undefined behavior (pr/16130). This patch more
precisely handles the undefined behavior case allowing the pessimization
to be removed.
ControlsExit replaces IsSubExpr to more precisely track the case where
undefined behavior is expected to occur. Because undefined behavior is
tracked more precisely we can remove MustExit from ExitLimit. MustExit
was used to track the case where the limit was computed potentially
assuming undefined behavior even if undefined behavior didn't necessarily
occur.
llvm-svn: 219517
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!
llvm-svn: 219135
We used to return PartialAlias if *either* variable being queried interacted
with arguments or globals. AFAICT, we can change this to only returning
MayAlias iff *both* variables being queried interacted with arguments or
globals.
Also, adding some basic functionality tests: some basic IPA tests, checking
that we give conservative responses with arguments/globals thrown in the mix,
and ensuring that we trace values through stores and loads.
Note that saying that 'x' interacted with arguments or globals means that the
Attributes of the StratifiedSet that 'x' belongs to has any bits set.
Patch by George Burgess IV, thanks!
llvm-svn: 219122
This patch broke 447.dealII on Darwin. I'm currently working on a reduced
test-case, but reverting for now to keep the bots happy.
<rdar://problem/18530107>
llvm-svn: 218944
argument of the llvm.dbg.declare/llvm.dbg.value intrinsics.
Previously, DIVariable was a variable-length field that has an optional
reference to a Metadata array consisting of a variable number of
complex address expressions. In the case of OpPiece expressions this is
wasting a lot of storage in IR, because when an aggregate type is, e.g.,
SROA'd into all of its n individual members, the IR will contain n copies
of the DIVariable, all alike, only differing in the complex address
reference at the end.
By making the complex address into an extra argument of the
dbg.value/dbg.declare intrinsics, all of the pieces can reference the
same variable and the complex address expressions can be uniqued across
the CU, too.
Down the road, this will allow us to move other flags, such as
"indirection" out of the DIVariable, too.
The new intrinsics look like this:
declare void @llvm.dbg.declare(metadata %storage, metadata %var, metadata %expr)
declare void @llvm.dbg.value(metadata %storage, i64 %offset, metadata %var, metadata %expr)
This patch adds a new LLVM-local tag to DIExpressions, so we can detect
and pretty-print DIExpression metadata nodes.
What this patch doesn't do:
This patch does not touch the "Indirect" field in DIVariable; but moving
that into the expression would be a natural next step.
http://reviews.llvm.org/D4919
rdar://problem/17994491
Thanks to dblaikie and dexonsmith for reviewing this patch!
Note: I accidentally committed a bogus older version of this patch previously.
llvm-svn: 218787
argument of the llvm.dbg.declare/llvm.dbg.value intrinsics.
Previously, DIVariable was a variable-length field that has an optional
reference to a Metadata array consisting of a variable number of
complex address expressions. In the case of OpPiece expressions this is
wasting a lot of storage in IR, because when an aggregate type is, e.g.,
SROA'd into all of its n individual members, the IR will contain n copies
of the DIVariable, all alike, only differing in the complex address
reference at the end.
By making the complex address into an extra argument of the
dbg.value/dbg.declare intrinsics, all of the pieces can reference the
same variable and the complex address expressions can be uniqued across
the CU, too.
Down the road, this will allow us to move other flags, such as
"indirection" out of the DIVariable, too.
The new intrinsics look like this:
declare void @llvm.dbg.declare(metadata %storage, metadata %var, metadata %expr)
declare void @llvm.dbg.value(metadata %storage, i64 %offset, metadata %var, metadata %expr)
This patch adds a new LLVM-local tag to DIExpressions, so we can detect
and pretty-print DIExpression metadata nodes.
What this patch doesn't do:
This patch does not touch the "Indirect" field in DIVariable; but moving
that into the expression would be a natural next step.
http://reviews.llvm.org/D4919
rdar://problem/17994491
Thanks to dblaikie and dexonsmith for reviewing this patch!
llvm-svn: 218778
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!
llvm-svn: 218714
The default implementation of getCmpSelInstrCost, which provides the cost of
icmp/fcmp/select instructions, did not deal sensibly with illegal vector types
that were scalarized. We'd ask for the legalization cost of the vector type,
which would return something like (4, f64) given an input of <4 x double>, and
we'd then check the TLI status of the ISD opcode on that scalar type. This would
result in querying (ISD::VSELECT, f64), for example. Amusingly enough,
ISD::VSELECT on scalar types is marked as Legal by default (as with most other
operations), and most backends never change this because VSELECT is never
generated on scalars. However, seeing the resulting operation as Legal, we'd
neglect to add the scalarization cost before returning. The result is that we'd
grossly under-estimate the cost of cmps/selects on illegal vector types.
Now, if type legalization clearly results in scalarization, we skip the early
return and add the scalarization cost.
llvm-svn: 217859
Cross-class copies being expensive is actually a trait of the microarchitecture, but as I haven't yet seen an example of a microarchitecture where they're cheap it seems best to just enable this by default, covering the non-mcpu build case.
llvm-svn: 217674
This adds a basic (but important) use of @llvm.assume calls in ScalarEvolution.
When SE is attempting to validate a condition guarding a loop (such as whether
or not the loop count can be zero), this check should also include dominating
assumptions.
llvm-svn: 217348
This provides an implementation of CFL alias analysis (including some
supporting data structures). Currently, we don't have any extremely fancy
features, sans some interprocedural analysis (i.e. no field sensitivity, etc.),
and we do best sitting behind BasicAA + TBAA. In such a configuration, we take
~0.6-0.8% of total compile time, and give ~7-8% NoAlias responses to queries
TBAA and BasicAA couldn't answer when bootstrapping LLVM. In testing this on
other projects, we've seen up to 10.5% of queries dropped by BasicAA+TBAA
answered with NoAlias by this algorithm.
Patch by George Burgess IV (with minor modifications by me -- mostly adapting
some BasicAA tests), thanks!
llvm-svn: 216970
This is the first commit in a series that add an @llvm.assume intrinsic which
can be used to provide the optimizer with a condition it may assume to be true
(when the control flow would hit the intrinsic call). Some basic properties are added here:
- llvm.invariant(true) is dead.
- llvm.invariant(false) is unreachable (this directly corresponds to the
documented behavior of MSVC's __assume(0)), so is llvm.invariant(undef).
The intrinsic is tagged as writing arbitrarily, in order to maintain control
dependencies. BasicAA has been updated, however, to return NoModRef for any
particular location-based query so that we don't unnecessarily block code
motion.
llvm-svn: 213973
This functionality is currently turned off by default.
Part of the motivation for introducing scoped-noalias metadata is to enable the
preservation of noalias parameter attribute information after inlining.
Sometimes this can be inferred from the code in the caller after inlining, but
often we simply lose valuable information.
The overall process if fairly simple:
1. Create a new unqiue scope domain.
2. For each (used) noalias parameter, create a new alias scope.
3. For each pointer, collect the underlying objects. Add a noalias scope for
each noalias parameter from which we're not derived (and has not been
captured prior to that point).
4. Add an alias.scope for each noalias parameter from which we might be
derived (or has been captured before that point).
Note that the capture checks apply only if one of the underlying objects is not
an identified function-local object.
llvm-svn: 213949
In the process of fixing the noalias parameter -> metadata conversion process
that will take place during inlining (which will be committed soon, but not
turned on by default), I have come to realize that the semantics provided by
yesterday's commit are not really what we want. Here's why:
void foo(noalias a, noalias b, noalias c, bool x) {
*q = x ? a : b;
*c = *q;
}
Generically, we know that *c does not alias with *a and with *b (so there is an
'and' in what we know we're not), and we know that *q might be derived from *a
or from *b (so there is an 'or' in what we know that we are). So we do not want
the semantics currently, where any noalias scope matching any alias.scope
causes a NoAlias return. What we want to know is that the noalias scopes form a
superset of the alias.scope list (meaning that all the things we know we're not
is a superset of all of things the other instruction might be).
Making that change, however, introduces a composibility problem. If we inline
once, adding the noalias metadata, and then inline again adding more, and we
append new scopes onto the noalias and alias.scope lists each time. But, this
means that we could change what was a NoAlias result previously into a MayAlias
result because we appended an additional scope onto one of the alias.scope
lists. So, instead of giving scopes the ability to have parents (which I had
borrowed from the TBAA implementation, but seems increasingly unlikely to be
useful in practice), I've given them domains. The subset/superset condition now
applies within each domain independently, and we only need it to hold in one
domain. Each time we inline, we add the new scopes in a new scope domain, and
everything now composes nicely. In addition, this simplifies the
implementation.
llvm-svn: 213948
This commit adds scoped noalias metadata. The primary motivations for this
feature are:
1. To preserve noalias function attribute information when inlining
2. To provide the ability to model block-scope C99 restrict pointers
Neither of these two abilities are added here, only the necessary
infrastructure. In fact, there should be no change to existing functionality,
only the addition of new features. The logic that converts noalias function
parameters into this metadata during inlining will come in a follow-up commit.
What is added here is the ability to generally specify noalias memory-access
sets. Regarding the metadata, alias-analysis scopes are defined similar to TBAA
nodes:
!scope0 = metadata !{ metadata !"scope of foo()" }
!scope1 = metadata !{ metadata !"scope 1", metadata !scope0 }
!scope2 = metadata !{ metadata !"scope 2", metadata !scope0 }
!scope3 = metadata !{ metadata !"scope 2.1", metadata !scope2 }
!scope4 = metadata !{ metadata !"scope 2.2", metadata !scope2 }
Loads and stores can be tagged with an alias-analysis scope, and also, with a
noalias tag for a specific scope:
... = load %ptr1, !alias.scope !{ !scope1 }
... = load %ptr2, !alias.scope !{ !scope1, !scope2 }, !noalias !{ !scope1 }
When evaluating an aliasing query, if one of the instructions is associated
with an alias.scope id that is identical to the noalias scope associated with
the other instruction, or is a descendant (in the scope hierarchy) of the
noalias scope associated with the other instruction, then the two memory
accesses are assumed not to alias.
Note that is the first element of the scope metadata is a string, then it can
be combined accross functions and translation units. The string can be replaced
by a self-reference to create globally unqiue scope identifiers.
[Note: This overview is slightly stylized, since the metadata nodes really need
to just be numbers (!0 instead of !scope0), and the scope lists are also global
unnamed metadata.]
Existing noalias metadata in a callee is "cloned" for use by the inlined code.
This is necessary because the aliasing scopes are unique to each call site
(because of possible control dependencies on the aliasing properties). For
example, consider a function: foo(noalias a, noalias b) { *a = *b; } that gets
inlined into bar() { ... if (...) foo(a1, b1); ... if (...) foo(a2, b2); } --
now just because we know that a1 does not alias with b1 at the first call site,
and a2 does not alias with b2 at the second call site, we cannot let inlining
these functons have the metadata imply that a1 does not alias with b2.
llvm-svn: 213864
In order to enable the preservation of noalias function parameter information
after inlining, and the representation of block-level __restrict__ pointer
information (etc.), additional kinds of aliasing metadata will be introduced.
This metadata needs to be carried around in AliasAnalysis::Location objects
(and MMOs at the SDAG level), and so we need to generalize the current scheme
(which is hard-coded to just one TBAA MDNode*).
This commit introduces only the necessary refactoring to allow for the
introduction of other aliasing metadata types, but does not actually introduce
any (that will come in a follow-up commit). What it does introduce is a new
AAMDNodes structure to hold all of the aliasing metadata nodes associated with
a particular memory-accessing instruction, and uses that structure instead of
the raw MDNode* in AliasAnalysis::Location, etc.
No functionality change intended.
llvm-svn: 213859
This reverts, "r213024 - Revert r212572 "improve BasicAA CS-CS queries", it
causes PR20303." with a fix for the bug in pr20303. As it turned out, the
relevant code was both wrong and over-conservative (because, as with the code
it replaced, it would return the overall ModRef mask even if just Ref had been
implied by the argument aliasing results). Hopefully, this correctly fixes both
problems.
Thanks to Nick Lewycky for reducing the test case for pr20303 (which I've
cleaned up a little and added in DSE's test directory). The BasicAA test has
also been updated to check for this error.
Original commit message:
BasicAA contains knowledge of certain intrinsics, such as memcpy and memset,
and uses that information to form more-accurate answers to CallSite vs. Loc
ModRef queries. Unfortunately, it did not use this information when answering
CallSite vs. CallSite queries.
Generically, when an intrinsic takes one or more pointers and the intrinsic is
marked only to read/write from its arguments, the offset/size is unknown. As a
result, the generic code that answers CallSite vs. CallSite (and CallSite vs.
Loc) queries in AA uses UnknownSize when forming Locs from an intrinsic's
arguments. While BasicAA's CallSite vs. Loc override could use more-accurate
size information for some intrinsics, it did not do the same for CallSite vs.
CallSite queries.
This change refactors the intrinsic-specific logic in BasicAA into a generic AA
query function: getArgLocation, which is overridden by BasicAA to supply the
intrinsic-specific knowledge, and used by AA's generic implementation. This
allows the intrinsic-specific knowledge to be used by both CallSite vs. Loc and
CallSite vs. CallSite queries, and simplifies the BasicAA implementation.
Currently, only one function, Mac's memset_pattern16, is handled by BasicAA
(all the rest are intrinsics). As a side-effect of this refactoring, BasicAA's
getModRefBehavior override now also returns OnlyAccessesArgumentPointees for
this function (which is an improvement).
llvm-svn: 213219
BasicAA contains knowledge of certain intrinsics, such as memcpy and memset,
and uses that information to form more-accurate answers to CallSite vs. Loc
ModRef queries. Unfortunately, it did not use this information when answering
CallSite vs. CallSite queries.
Generically, when an intrinsic takes one or more pointers and the intrinsic is
marked only to read/write from its arguments, the offset/size is unknown. As a
result, the generic code that answers CallSite vs. CallSite (and CallSite vs.
Loc) queries in AA uses UnknownSize when forming Locs from an intrinsic's
arguments. While BasicAA's CallSite vs. Loc override could use more-accurate
size information for some intrinsics, it did not do the same for CallSite vs.
CallSite queries.
This change refactors the intrinsic-specific logic in BasicAA into a generic AA
query function: getArgLocation, which is overridden by BasicAA to supply the
intrinsic-specific knowledge, and used by AA's generic implementation. This
allows the intrinsic-specific knowledge to be used by both CallSite vs. Loc and
CallSite vs. CallSite queries, and simplifies the BasicAA implementation.
Currently, only one function, Mac's memset_pattern16, is handled by BasicAA
(all the rest are intrinsics). As a side-effect of this refactoring, BasicAA's
getModRefBehavior override now also returns OnlyAccessesArgumentPointees for
this function (which is an improvement).
llvm-svn: 212572
This patch:
1) Improves the cost model for x86 alternate shuffles (originally
added at revision 211339);
2) Teaches the Cost Model Analysis pass how to analyze alternate shuffles.
Alternate shuffles are a special kind of blend; on x86, we can often
easily lowered alternate shuffled into single blend
instruction (depending on the subtarget features).
The existing cost model didn't take into account subtarget features.
Also, it had a couple of "dead" entries for vector types that are never
legal (example: on x86 types v2i32 and v2f32 are not legal; those are
always either promoted or widened to 128-bit vector types).
The new x86 cost model takes into account what target features we have
before returning the shuffle cost (i.e. the number of instructions
after the blend is lowered/expanded).
This patch also teaches the Cost Model Analysis how to identify and analyze
alternate shuffles (i.e. 'SK_Alternate' shufflevector instructions):
- added function 'isAlternateVectorMask';
- added some logic to check if an instruction is a alternate shuffle and, in
case, call the target specific TTI to get the corresponding shuffle cost;
- added a test to verify the cost model analysis on alternate shuffles.
llvm-svn: 212296
Before, we where looking at the size of the pointer type that specifies the
location from which to load the element. This did not make any sense at all.
This change fixes a bug in the delinearization where we failed to delinerize
certain load instructions.
llvm-svn: 210435
The delinearization is needed only to remove the non linearity induced by
expressions involving multiplications of parameters and induction variables.
There is no problem in dealing with constant times parameters, or constant times
an induction variable.
For this reason, the current patch discards all constant terms and multipliers
before running the delinearization algorithm on the terms. The only thing
remaining in the term expressions are parameters and multiply expressions of
parameters: these simplified term expressions are passed to the array shape
recognizer that will not recognize constant dimensions anymore: these will be
recognized as different strides in parametric subscripts.
The only important special case of a constant dimension is the size of elements.
Instead of relying on the delinearization to infer the size of an element,
compute the element size from the base address type. This is a much more precise
way of computing the element size than before, as we would have mixed together
the size of an element with the strides of the innermost dimension.
llvm-svn: 209691
This commit starts with a "git mv ARM64 AArch64" and continues out
from there, renaming the C++ classes, intrinsics, and other
target-local objects for consistency.
"ARM64" test directories are also moved, and tests that began their
life in ARM64 use an arm64 triple, those from AArch64 use an aarch64
triple. Both should be equivalent though.
This finishes the AArch64 merge, and everyone should feel free to
continue committing as normal now.
llvm-svn: 209577
This is a follow-up to r209358: PR19799: Indvars miscompile due to an
incorrect max backedge taken count from SCEV.
That fix was incomplete as pointed out by Arnold and Michael Z. The
code was also too confusing. It needed a careful rewrite with more
unit tests. This version will also happen to optimize more cases.
<rdar://17005101> PR19799: Indvars miscompile...
llvm-svn: 209545
This has to do with the trip count computation for loops with multiple
exits, which is quite subtle. Most passes just ask for a single trip
count number, so we must be conservative assuming any exit could be
taken. Normally, we rely on the "exact" trip count, which was
correctly given as "unknown". However, SCEV also gives a "max"
back-edge taken count. The loops max BE taken count is conservatively
a maximum over the max of each exit's non-exiting iterations
count. Note that some exit tests can be skipped so the max loop
back-edge taken count can actually exceed the max non-exiting
iterations for some exits. However, when we know the loop *latch*
cannot be skipped, we can directly use its max taken count
disregarding other exits. I previously took the minimum here without
checking whether the other exit could be skipped. The correct, and
simpler thing to do here is just to directly use the loop latch's max
non-exiting iterations as the loops max back-edge count.
In the problematic test case, the first loop exit had a max of zero
non-exiting iterations, but could be skipped. The loop latch was known
not to be skipped but had max of one non-exiting iteration. We
incorrectly claimed the loop back-edge could be taken zero times, when
it is actually taken one time.
Fixes Loop %for.body.i: <multiple exits> Unpredictable backedge-taken count.
Loop %for.body.i: max backedge-taken count is 1.
llvm-svn: 209358
Tested by comparing make check VERBOSE=1 before and after to make sure
no tests are missed. (VERBOSE=1 prints the list of tests.)
Only one test :( remains where .cpp is required:
tools/llvm-cov/range_based_for.cpp:// RUN: llvm-cov range_based_for.cpp | FileCheck %s --check-prefix=STDOUT
The topic was discussed in this thread:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20140428/214905.html
llvm-svn: 208621
To compute the dimensions of the array in a unique way, we split the
delinearization analysis in three steps:
- find parametric terms in all memory access functions
- compute the array dimensions from the set of terms
- compute the delinearized access functions for each dimension
The first step is executed on all the memory access functions such that we
gather all the patterns in which an array is accessed. The second step reduces
all this information in a unique description of the sizes of the array. The
third step is delinearizing each memory access function following the common
description of the shape of the array computed in step 2.
This rewrite of the delinearization pass also solves a problem we had with the
previous implementation: because the previous algorithm was by induction on the
structure of the SCEV, it would not correctly recognize the shape of the array
when the memory access was not following the nesting of the loops: for example,
see polly/test/ScopInfo/multidim_only_ivs_3d_reverse.ll
; void foo(long n, long m, long o, double A[n][m][o]) {
;
; for (long i = 0; i < n; i++)
; for (long j = 0; j < m; j++)
; for (long k = 0; k < o; k++)
; A[i][k][j] = 1.0;
Starting with this patch we no longer delinearize access functions that do not
contain parameters, for example in test/Analysis/DependenceAnalysis/GCD.ll
;; for (long int i = 0; i < 100; i++)
;; for (long int j = 0; j < 100; j++) {
;; A[2*i - 4*j] = i;
;; *B++ = A[6*i + 8*j];
these accesses will not be delinearized as the upper bound of the loops are
constants, and their access functions do not contain SCEVUnknown parameters.
llvm-svn: 208232
This reverts commit r207287, reapplying r207286.
I'm hoping that declaring an explicit struct and instantiating
`addBlockEdges()` directly works around the GCC crash from r207286.
This is a lot more boilerplate, though.
llvm-svn: 207438
This reverts commit r207286. It causes an ICE on the
cmake-llvm-x86_64-linux buildbot [1]:
llvm/lib/Analysis/BlockFrequencyInfo.cpp: In lambda function:
llvm/lib/Analysis/BlockFrequencyInfo.cpp:182:1: internal compiler error: in get_expr_operands, at tree-ssa-operands.c:1035
[1]: http://bb.pgr.jp/builders/cmake-llvm-x86_64-linux/builds/12093/steps/build_llvm/logs/stdio
llvm-svn: 207287
Previously, irreducible backedges were ignored. With this commit,
irreducible SCCs are discovered on the fly, and modelled as loops with
multiple headers.
This approximation specifies the headers of irreducible sub-SCCs as its
entry blocks and all nodes that are targets of a backedge within it
(excluding backedges within true sub-loops). Block frequency
calculations act as if we insert a new block that intercepts all the
edges to the headers. All backedges and entries to the irreducible SCC
point to this imaginary block. This imaginary block has an edge (with
even probability) to each header block.
The result is now reasonable enough that I've added a number of
testcases for irreducible control flow. I've outlined in
`BlockFrequencyInfoImpl.h` ways to improve the approximation.
<rdar://problem/14292693>
llvm-svn: 207286
Remove the concepts of "forward" and "general" mass distributions, which
was wrong. The split might have made sense in an early version of the
algorithm, but it's definitely wrong now.
<rdar://problem/14292693>
llvm-svn: 207195
Strip irreducible testcases to pure control flow. The function calls
made the branch weights more believable but cluttered it up a lot.
There isn't going to be any constant analysis here, so just use dumb
branch logic to clarify the important parts.
<rdar://problem/14292693>
llvm-svn: 207192
The branch that skips irreducible backedges was only active when
propagating mass at the top-level. In particular, when propagating mass
through a loop recognized by `LoopInfo` with irreducible control flow
inside, irreducible backedges would not be skipped.
Not sure where that idea came from, but the result was that mass was
lost until after loop exit. Added a testcase that covers this case.
llvm-svn: 206860
This reverts commit r206707, reapplying r206704. The preceding commit
to CalcSpillWeights should have sorted out the failing buildbots.
<rdar://problem/14292693>
llvm-svn: 206766
This reverts commit r206677, reapplying my BlockFrequencyInfo rewrite.
I've done a careful audit, added some asserts, and fixed a couple of
bugs (unfortunately, they were in unlikely code paths). There's a small
chance that this will appease the failing bots [1][2]. (If so, great!)
If not, I have a follow-up commit ready that will temporarily add
-debug-only=block-freq to the two failing tests, allowing me to compare
the code path between what the failing bots and what my machines (and
the rest of the bots) are doing. Once I've triggered those builds, I'll
revert both commits so the bots go green again.
[1]: http://bb.pgr.jp/builders/ninja-x64-msvc-RA-centos6/builds/1816
[2]: http://llvm-amd64.freebsd.your.org/b/builders/clang-i386-freebsd/builds/18445
<rdar://problem/14292693>
llvm-svn: 206704
This reverts commit r206666, as planned.
Still stumped on why the bots are failing. Sanitizer bots haven't
turned anything up. If anyone can help me debug either of the failures
(referenced in r206666) I'll owe them a beer. (In the meantime, I'll be
auditing my patch for undefined behaviour.)
llvm-svn: 206677
This reverts commit r206628, reapplying r206622 (and r206626).
Two tests are failing only on buildbots [1][2]: i.e., I can't reproduce
on Darwin, and Chandler can't reproduce on Linux. Asan and valgrind
don't tell us anything, but we're hoping the msan bot will catch it.
So, I'm applying this again to get more feedback from the bots. I'll
leave it in long enough to trigger builds in at least the sanitizer
buildbots (it was failing for reasons unrelated to my commit last time
it was in), and hopefully a few others.... and then I expect to revert a
third time.
[1]: http://bb.pgr.jp/builders/ninja-x64-msvc-RA-centos6/builds/1816
[2]: http://llvm-amd64.freebsd.your.org/b/builders/clang-i386-freebsd/builds/18445
llvm-svn: 206666
This reverts commit r206622 and the MSVC fixup in r206626.
Apparently the remotely failing tests are still failing, despite my
attempt to fix the nondeterminism in r206621.
llvm-svn: 206628
This reverts commit r206556, effectively reapplying commit r206548 and
its fixups in r206549 and r206550.
In an intervening commit I've added target triples to the tests that
were failing remotely [1] (but passing locally). I'm hoping the mystery
is solved? I'll revert this again if the tests are still failing
remotely.
[1]: http://bb.pgr.jp/builders/ninja-x64-msvc-RA-centos6/builds/1816
llvm-svn: 206622
LazyCallGraph. This is the start of the whole point of this different
abstraction, but it is just the initial bits. Here is a run-down of
what's going on here. I'm planning to incorporate some (or all) of this
into comments going forward, hopefully with better editing and wording.
=]
The crux of the problem with the traditional way of building SCCs is
that they are ephemeral. The new pass manager however really needs the
ability to associate analysis passes and results of analysis passes with
SCCs in order to expose these analysis passes to the SCC passes. Making
this work is kind-of the whole point of the new pass manager. =]
So, when we're building SCCs for the call graph, we actually want to
build persistent nodes that stick around and can be reasoned about
later. We'd also like the ability to walk the SCC graph in more complex
ways than just the traditional postorder traversal of the current CGSCC
walk. That means that in addition to being persistent, the SCCs need to
be connected into a useful graph structure.
However, we still want the SCCs to be formed lazily where possible.
These constraints are quite hard to satisfy with the SCC iterator. Also,
using that would bypass our ability to actually add data to the nodes of
the call graph to facilite implementing the Tarjan walk. So I've
re-implemented things in a more direct and embedded way. This
immediately makes it easy to get the persistence and connectivity
correct, and it also allows leveraging the existing nodes to simplify
the algorithm. I've worked somewhat to make this implementation more
closely follow the traditional paper's nomenclature and strategy,
although it is still a bit obtuse because it isn't recursive, using
an explicit stack and a tail call instead, and it is interruptable,
resuming each time we need another SCC.
The other tricky bit here, and what actually took almost all the time
and trials and errors I spent building this, is exactly *what* graph
structure to build for the SCCs. The naive thing to build is the call
graph in its newly acyclic form. I wrote about 4 versions of this which
did precisely this. Inevitably, when I experimented with them across
various use cases, they became incredibly awkward. It was all
implementable, but it felt like a complete wrong fit. Square peg, round
hole. There were two overriding aspects that pushed me in a different
direction:
1) We want to discover the SCC graph in a postorder fashion. That means
the root node will be the *last* node we find. Using the call-SCC DAG
as the graph structure of the SCCs results in an orphaned graph until
we discover a root.
2) We will eventually want to walk the SCC graph in parallel, exploring
distinct sub-graphs independently, and synchronizing at merge points.
This again is not helped by the call-SCC DAG structure.
The structure which, quite surprisingly, ended up being completely
natural to use is the *inverse* of the call-SCC DAG. We add the leaf
SCCs to the graph as "roots", and have edges to the caller SCCs. Once
I switched to building this structure, everything just fell into place
elegantly.
Aside from general cleanups (there are FIXMEs and too few comments
overall) that are still needed, the other missing piece of this is
support for iterating across levels of the SCC graph. These will become
useful for implementing #2, but they aren't an immediate priority.
Once SCCs are in good shape, I'll be working on adding mutation support
for incremental updates and adding the pass manager that this analysis
enables.
llvm-svn: 206581
Rewrite the shared implementation of BlockFrequencyInfo and
MachineBlockFrequencyInfo entirely.
The old implementation had a fundamental flaw: precision losses from
nested loops (or very wide branches) compounded past loop exits (and
convergence points).
The @nested_loops testcase at the end of
test/Analysis/BlockFrequencyAnalysis/basic.ll is motivating. This
function has three nested loops, with branch weights in the loop headers
of 1:4000 (exit:continue). The old analysis gives non-sensical results:
Printing analysis 'Block Frequency Analysis' for function 'nested_loops':
---- Block Freqs ----
entry = 1.0
for.cond1.preheader = 1.00103
for.cond4.preheader = 5.5222
for.body6 = 18095.19995
for.inc8 = 4.52264
for.inc11 = 0.00109
for.end13 = 0.0
The new analysis gives correct results:
Printing analysis 'Block Frequency Analysis' for function 'nested_loops':
block-frequency-info: nested_loops
- entry: float = 1.0, int = 8
- for.cond1.preheader: float = 4001.0, int = 32007
- for.cond4.preheader: float = 16008001.0, int = 128064007
- for.body6: float = 64048012001.0, int = 512384096007
- for.inc8: float = 16008001.0, int = 128064007
- for.inc11: float = 4001.0, int = 32007
- for.end13: float = 1.0, int = 8
Most importantly, the frequency leaving each loop matches the frequency
entering it.
The new algorithm leverages BlockMass and PositiveFloat to maintain
precision, separates "probability mass distribution" from "loop
scaling", and uses dithering to eliminate probability mass loss. I have
unit tests for these types out of tree, but it was decided in the review
to make the classes private to BlockFrequencyInfoImpl, and try to shrink
them (or remove them entirely) in follow-up commits.
The new algorithm should generally have a complexity advantage over the
old. The previous algorithm was quadratic in the worst case. The new
algorithm is still worst-case quadratic in the presence of irreducible
control flow, but it's linear without it.
The key difference between the old algorithm and the new is that control
flow within a loop is evaluated separately from control flow outside,
limiting propagation of precision problems and allowing loop scale to be
calculated independently of mass distribution. Loops are visited
bottom-up, their loop scales are calculated, and they are replaced by
pseudo-nodes. Mass is then distributed through the function, which is
now a DAG. Finally, loops are revisited top-down to multiply through
the loop scales and the masses distributed to pseudo nodes.
There are some remaining flaws.
- Irreducible control flow isn't modelled correctly. LoopInfo and
MachineLoopInfo ignore irreducible edges, so this algorithm will
fail to scale accordingly. There's a note in the class
documentation about how to get closer. See also the comments in
test/Analysis/BlockFrequencyInfo/irreducible.ll.
- Loop scale is limited to 4096 per loop (2^12) to avoid exhausting
the 64-bit integer precision used downstream.
- The "bias" calculation proposed on llvmdev is *not* incorporated
here. This will be added in a follow-up commit, once comments from
this review have been handled.
llvm-svn: 206548
Previously, BranchProbabilityInfo::calcLoopBranchHeuristics would determine the weights of basic blocks inside loops even when it didn't have enough information to estimate the branch probabilities correctly. This patch fixes the function to exit early if it doesn't see any exit edges or back edges and let the later heuristics determine the weights.
This fixes PR18705 and <rdar://problem/15991090>.
Differential Revision: http://reviews.llvm.org/D3363
llvm-svn: 206194
BasicTTI::getMemoryOpCost must explicitly check for non-simple types; setting
AllowUnknown=true with TLI->getSimpleValueType is not sufficient because, for
example, non-power-of-two vector types return non-simple EVTs (not MVT::Other).
llvm-svn: 206150
This provides more realistic costs for the insert/extractelement instructions
(which are load/store pairs), accounts for the cheap unaligned Altivec load
sequence, and for unaligned VSX load/stores.
Bad news:
MultiSource/Applications/sgefa/sgefa - 35% slowdown (this will require more investigation)
SingleSource/Benchmarks/McGill/queens - 20% slowdown (we no longer vectorize this, but it was a constant store that was scalarized)
MultiSource/Benchmarks/FreeBench/pcompress2/pcompress2 - 2% slowdown
Good news:
SingleSource/Benchmarks/Shootout/ary3 - 54% speedup
SingleSource/Benchmarks/Shootout-C++/ary - 40% speedup
MultiSource/Benchmarks/Ptrdist/ks/ks - 35% speedup
MultiSource/Benchmarks/FreeBench/neural/neural - 30% speedup
MultiSource/Benchmarks/TSVC/Symbolics-flt/Symbolics-flt - 20% speedup
Unfortunately, estimating the costs of the stack-based scalarization sequences
is hard, and adjusting these costs is like a game of whac-a-mole :( I'll
revisit this again after we have better codegen for vector extloads and
truncstores and unaligned load/stores.
llvm-svn: 205658
When a vector type legalizes to a larger vector type, and the target does not
support the associated extending load (or truncating store), then legalization
will scalarize the load (or store) resulting in an associated scalarization
cost. BasicTTI::getMemoryOpCost needs to account for this.
Between this, and r205487, PowerPC on the P7 with VSX enabled shows:
MultiSource/Benchmarks/PAQ8p/paq8p: 43% speedup
SingleSource/Benchmarks/BenchmarkGame/puzzle: 51% speedup
SingleSource/UnitTests/Vectorizer/gcc-loops 28% speedup
(some of these are new; some of these, such as PAQ8p, just reverse regressions
that VSX support would trigger)
llvm-svn: 205495
For an cast (extension, etc.), the currently logic predicts a low cost if the
associated operation (keyed on the destination type) is legal (or promoted).
This is not true when the number of values required to legalize the type is
changing. For example, <8 x i16> being sign extended by <8 x i32> is not
generically cheap on PPC with VSX, even though sign extension to v4i32 is
legal, because two output v4i32 values are required compared to the single
v8i16 input value, and without custom logic in the target, this conversion will
scalarize.
llvm-svn: 205487
This adds a second implementation of the AArch64 architecture to LLVM,
accessible in parallel via the "arm64" triple. The plan over the
coming weeks & months is to merge the two into a single backend,
during which time thorough code review should naturally occur.
Everything will be easier with the target in-tree though, hence this
commit.
llvm-svn: 205090
This commit consist of two parts.
The first part fix the PR15967. The wrong conclusion was made when the MaxLookup
limit was reached. The fix introduce a out parameter (MaxLookupReached) to
DecomposeGEPExpression that the function aliasGEP can act upon.
The second part is introducing the constant MaxLookupSearchDepth to make sure
that DecomposeGEPExpression and GetUnderlyingObject use the same search depth.
This is a small cleanup to clarify the original algorithm.
Patch by Karl-Johan Karlsson!
llvm-svn: 204859
If we have a loop of the form
for (unsigned n = 0; n != (k & -32); n += 32) {}
then we know that n is always divisible by 32 and the loop must
terminate. Even if we have a condition where the loop counter will
overflow it'll always hold this invariant.
PR19183. Our loop vectorizer creates this pattern and it's also
occasionally formed by loop counters derived from pointers.
llvm-svn: 204728
On ELF and COFF an alias is just another name for a position in the file.
There is no way to refer to a position in another file, so an alias to
undefined is meaningless.
MachO currently doesn't support aliases. The spec has a N_INDR, which when
implemented will have a different set of restrictions. Adding support for
it shouldn't be harder than any other IR extension.
For now, having the IR represent what is actually possible with current
tools makes it easier to fix the design of GlobalAlias.
llvm-svn: 203705
the legalization cost must be included to get an accurate
estimation of the total cost of the scalarized vector.
The inaccurate cost triggered unprofitable SLP vectorization on
32-bit X86.
Summary:
Include legalization overhead when computing scalarization cost
Reviewers: hfinkel, nadav
CC: chandlerc, rnk, llvm-commits
Differential Revision: http://llvm-reviews.chandlerc.com/D2992
llvm-svn: 203509
'OK_NonUniformConstValue' to identify operands which are constants but
not constant splats.
The cost model now allows returning 'OK_NonUniformConstValue'
for non splat operands that are instances of ConstantVector or
ConstantDataVector.
With this change, targets are now able to compute different costs
for instructions with non-uniform constant operands.
For example, On X86 the cost of a vector shift may vary depending on whether
the second operand is a uniform or non-uniform constant.
This patch applies the following changes:
- The cost model computation now takes into account non-uniform constants;
- The cost of vector shift instructions has been improved in
X86TargetTransformInfo analysis pass;
- BBVectorize, SLPVectorizer and LoopVectorize now know how to distinguish
between non-uniform and uniform constant operands.
Added a new test to verify that the output of opt
'-cost-model -analyze' is valid in the following configurations: SSE2,
SSE4.1, AVX, AVX2.
llvm-svn: 201272
The most important part of this is probably adding any cost at all for
operations like zext <8 x i8> to <8 x i32>. Before they were being
recorded as extremely costly (24, I believe) which made LLVM fall back
on a 4-wide vectorisation of a loop.
It also rebalances the values for sext, zext and trunc. Lacking any
other sane metric that might work across CPU microarchitectures I went
for instructions. This seems to be in reasonable accord with the rest
of the table (sitofp, ...) though no doubt at least one value is
sub-optimal for some bizarre reason.
Finally, separate AVX and AVX2 values are provided where appropriate.
The CodeGen is quite different in many cases.
rdar://problem/15981990
llvm-svn: 200928
The primary motivation for this pass is to separate the call graph
analysis used by the new pass manager's CGSCC pass management from the
existing call graph analysis pass. That analysis pass is (somewhat
unfortunately) over-constrained by the existing CallGraphSCCPassManager
requirements. Those requirements make it *really* hard to cleanly layer
the needed functionality for the new pass manager on top of the existing
analysis.
However, there are also a bunch of things that the pass manager would
specifically benefit from doing differently from the existing call graph
analysis, and this new implementation tries to address several of them:
- Be lazy about scanning function definitions. The existing pass eagerly
scans the entire module to build the initial graph. This new pass is
significantly more lazy, and I plan to push this even further to
maximize locality during CGSCC walks.
- Don't use a single synthetic node to partition functions with an
indirect call from functions whose address is taken. This node creates
a huge choke-point which would preclude good parallelization across
the fanout of the SCC graph when we got to the point of looking at
such changes to LLVM.
- Use a memory dense and lightweight representation of the call graph
rather than value handles and tracking call instructions. This will
require explicit update calls instead of some updates working
transparently, but should end up being significantly more efficient.
The explicit update calls ended up being needed in many cases for the
existing call graph so we don't really lose anything.
- Doesn't explicitly model SCCs and thus doesn't provide an "identity"
for an SCC which is stable across updates. This is essential for the
new pass manager to work correctly.
- Only form the graph necessary for traversing all of the functions in
an SCC friendly order. This is a much simpler graph structure and
should be more memory dense. It does limit the ways in which it is
appropriate to use this analysis. I wish I had a better name than
"call graph". I've commented extensively this aspect.
This is still very much a WIP, in fact it is really just the initial
bits. But it is about the fourth version of the initial bits that I've
implemented with each of the others running into really frustrating
problms. This looks like it will actually work and I'd like to split the
actual complexity across commits for the sake of my reviewers. =] The
rest of the implementation along with lots of wiring will follow
somewhat more rapidly now that there is a good path forward.
Naturally, this doesn't impact any of the existing optimizer. This code
is specific to the new pass manager.
A bunch of thanks are deserved for the various folks that have helped
with the design of this, especially Nick Lewycky who actually sat with
me to go through the fundamentals of the final version here.
llvm-svn: 200903
Unfortunately, this in turn led to some lower quality SCEVs due to some different paths through expression simplification, so add getUDivExactExpr and use it. This fixes all instances of the problems that I found, but we can make that function smarter as necessary.
Merge test "xor-and.ll" into "and-xor.ll" since I needed to update it anyways. Test 'nsw-offset.ll' analyzes a little deeper, %n now gets a scev in terms of %no instead of a SCEVUnknown.
llvm-svn: 200203
cycles
This allows the value equality check to work even if we don't have a dominator
tree. Also add some more comments.
I was worried about compile time impacts and did not implement reachability but
used the dominance check in the initial patch. The trade-off was that the
dominator tree was required.
The llvm utility function isPotentiallyReachable cuts off the recursive search
after 32 visits. Testing did not show any compile time regressions showing my
worries unjustfied.
No compile time or performance regressions at O3 -flto -mavx on test-suite +
externals.
Addresses review comments from r198290.
llvm-svn: 198400
When there are cycles in the value graph we have to be careful interpreting
"Value*" identity as "value" equivalence. We interpret the value of a phi node
as the value of its operands.
When we check for value equivalence now we make sure that the "Value*" dominates
all cycles (phis).
%0 = phi [%noaliasval, %addr2]
%l = load %ptr
%addr1 = gep @a, 0, %l
%addr2 = gep @a, 0, (%l + 1)
store %ptr ...
Before this patch we would return NoAlias for (%0, %addr1) which is wrong
because the value of the load is from different iterations of the loop.
Tested on x86_64 -mavx at O3 and O3 -flto with no performance or compile time
regressions.
PR18068
radar://15653794
llvm-svn: 198290
The tests just hit this with a different sized
address space since I haven't figured out how
to use this to break it.
I thought I committed this a long time ago,
and I'm not sure why missing this hasn't caused
any problems.
llvm-svn: 194903
Patch by Michele Scandale!
Rewrite of the functions used to compute the backedge taken count of a
loop on LT and GT comparisons.
I decided to split the handling of LT and GT cases becasue the trick
"a > b == -a < -b" in some cases prevents the trip count computation
due to the multiplication by -1 on the two operands of the
comparison. This issue comes from the conservative computation of
value range of SCEVs: taking the negative SCEV of an expression that
have a small positive range (e.g. [0,31]), we would have a SCEV with a
fullset as value range.
Indeed, in the new rewritten function I tried to better handle the
maximum backedge taken count computation when MAX/MIN expression are
used to handle the cases where no entry guard is found.
Some test have been modified in order to check the new value correctly
(I manually check them and reasoning on possible overflow the new
values seem correct).
I finally added a new test case related to the multiplication by -1
issue on GT comparisons.
llvm-svn: 194116
This adds another heuristic to BPI, similar to the existing heuristic that
considers (x == 0) unlikely to be true. As suggested in the PACT'98 paper by
Deitrich, Cheng, and Hwu, -1 is often used to indicate an invalid index, and
equality comparisons with -1 are also unlikely to succeed. Local
experimentation supports this hypothesis: This yields a 1-2% speedup in the
test-suite sqlite benchmark on the PPC A2 core, with no significant
regressions.
llvm-svn: 193855
We can't do this for the general case as saying a GEP with a negative index
doesn't have unsigned wrap isn't valid for negative indices.
%gep = getelementptr inbounds i32* %p, i64 -1
But an inbounds GEP cannot run past the end of address space. So we check for
the very common case of a positive index and make GEPs derived from that NUW.
Together with Andy's recent non-unit stride work this lets us analyze loops
like
void foo3(int *a, int *b) {
for (; a < b; a++) {}
}
PR12375, PR12376.
Differential Revision: http://llvm-reviews.chandlerc.com/D2033
llvm-svn: 193514
Major steps include:
1). introduces a not-addr-taken bit-field in GlobalVariable
2). GlobalOpt pass sets "not-address-taken" if it proves a global varirable
dosen't have its address taken.
3). AA use this info for disambiguation.
llvm-svn: 193251
We can have a struct type with a single field and the field does not start
with 0. In that case, we should correctly update the offset.
llvm-svn: 193137
The test before wasn't successfully testing this
since it was missing the datalayout piece to change
the size of the second address space.
llvm-svn: 193102
SCEV currently fails to compute loop counts for nonunit stride
loops. This comes up frequently. It prevents loop optimization and
forces vectorization to insert extra loop checks.
For example:
void foo(int n, int *x) {
for (int i = 0; i < n; i += 3) {
x[i] = i;
x[i+1] = i+1;
x[i+2] = i+2;
}
}
We need to properly handle the case in which limit > INT_MAX-stride. In
the above case: n > INT_MAX-3. In this case the loop counter will step
beyond the limit and overflow at the same time. However, knowing that
signed integer overlow in undefined, we can assume the loop test
behavior is arbitrary after overflow. This obeys both C undefined
behavior rules, and the more strict LLVM poison value rules.
I'm finally fixing this in response to Hal Finkel's persistence.
The most probable reason that we never optimized this before is that
we were being careful to handle case where the developer expected a
side-effect free infinite loop relying on overflow:
for (int i = 0; i < n; i += s) {
++j;
}
return j;
If INT_MAX+1 is a multiple of s and n > INT_MAX-s, then we might
expect an infinite loop. However there are plenty of ways to achieve
this effect without relying on undefined behavior of signed overflow.
llvm-svn: 193015
infrastructure.
This was essentially work toward PGO based on a design that had several
flaws, partially dating from a time when LLVM had a different
architecture, and with an effort to modernize it abandoned without being
completed. Since then, it has bitrotted for several years further. The
result is nearly unusable, and isn't helping any of the modern PGO
efforts. Instead, it is getting in the way, adding confusion about PGO
in LLVM and distracting everyone with maintenance on essentially dead
code. Removing it paves the way for modern efforts around PGO.
Among other effects, this removes the last of the runtime libraries from
LLVM. Those are being developed in the separate 'compiler-rt' project
now, with somewhat different licensing specifically more approriate for
runtimes.
llvm-svn: 191835
Remove the command line argument "struct-path-tbaa" since we should not depend
on command line argument to decide which format the IR file is using. Instead,
we check the first operand of the tbaa tag node, if it is a MDNode, we treat
it as struct-path aware TBAA format, otherwise, we treat it as scalar TBAA
format.
When clang starts to use struct-path aware TBAA format no matter whether
struct-path-tbaa is no, and we can auto-upgrade existing bc files, the support
for scalar TBAA format can be dropped.
Existing testing cases are updated to use the struct-path aware TBAA format.
llvm-svn: 191538
Upcoming SLP vectorization improvements will want to be able to estimate costs
of horizontal reductions. Add infrastructure to support this.
We model reductions as a series of (shufflevector,add) tuples ultimately
followed by an extractelement. For example, for an add-reduction of <4 x float>
we could generate the following sequence:
(v0, v1, v2, v3)
\ \ / /
\ \ /
+ +
(v0+v2, v1+v3, undef, undef)
\ /
((v0+v2) + (v1+v3), undef, undef)
%rdx.shuf = shufflevector <4 x float> %rdx, <4 x float> undef,
<4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
%bin.rdx = fadd <4 x float> %rdx, %rdx.shuf
%rdx.shuf7 = shufflevector <4 x float> %bin.rdx, <4 x float> undef,
<4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
%bin.rdx8 = fadd <4 x float> %bin.rdx, %rdx.shuf7
%r = extractelement <4 x float> %bin.rdx8, i32 0
This commit adds a cost model interface "getReductionCost(Opcode, Ty, Pairwise)"
that will allow clients to ask for the cost of such a reduction (as backends
might generate more efficient code than the cost of the individual instructions
summed up). This interface is excercised by the CostModel analysis pass which
looks for reduction patterns like the one above - starting at extractelements -
and if it sees a matching sequence will call the cost model interface.
We will also support a second form of pairwise reduction that is well supported
on common architectures (haddps, vpadd, faddp).
(v0, v1, v2, v3)
\ / \ /
(v0+v1, v2+v3, undef, undef)
\ /
((v0+v1)+(v2+v3), undef, undef, undef)
%rdx.shuf.0.0 = shufflevector <4 x float> %rdx, <4 x float> undef,
<4 x i32> <i32 0, i32 2 , i32 undef, i32 undef>
%rdx.shuf.0.1 = shufflevector <4 x float> %rdx, <4 x float> undef,
<4 x i32> <i32 1, i32 3, i32 undef, i32 undef>
%bin.rdx.0 = fadd <4 x float> %rdx.shuf.0.0, %rdx.shuf.0.1
%rdx.shuf.1.0 = shufflevector <4 x float> %bin.rdx.0, <4 x float> undef,
<4 x i32> <i32 0, i32 undef, i32 undef, i32 undef>
%rdx.shuf.1.1 = shufflevector <4 x float> %bin.rdx.0, <4 x float> undef,
<4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
%bin.rdx.1 = fadd <4 x float> %rdx.shuf.1.0, %rdx.shuf.1.1
%r = extractelement <4 x float> %bin.rdx.1, i32 0
llvm-svn: 190876
- Instead of setting the suffixes in a bunch of places, just set one master
list in the top-level config. We now only modify the suffix list in a few
suites that have one particular unique suffix (.ml, .mc, .yaml, .td, .py).
- Aside from removing the need for a bunch of lit.local.cfg files, this enables
4 tests that were inadvertently being skipped (one in
Transforms/BranchFolding, a .s file each in DebugInfo/AArch64 and
CodeGen/PowerPC, and one in CodeGen/SI which is now failing and has been
XFAILED).
- This commit also fixes a bunch of config files to use config.root instead of
older copy-pasted code.
llvm-svn: 188513
This patch fixes the multiple breakages on ARM test-suite after the SLP
vectorizer was introduced by default on O3. The problem was an illegal
vector type on ARMTTI::getCmpSelInstrCost() <3 x i1> which is not simple.
The guard protects this code from breaking (cause of the problems) but
doesn't fix the issue that is generating the odd vector in the first
place, which also needs to be investigated.
llvm-svn: 187658
This fixes an oversight that Intrinsic::nearbyint was not being mapped to
ISD::FNEARBYINT (thus fixing the over-optimistic cost we were assigning to
nearbyint calls for some targets).
llvm-svn: 185783
This is easier to read than the internal fixed-point representation.
If anybody knows the correct algorithm for converting fixed-point
numbers to base 10, feel free to fix it.
llvm-svn: 184881
This is a band-aid to fix the most severe regressions we're seeing from basing
spill decisions on block frequencies, until we have a better solution.
llvm-svn: 184835
Zero is used by BlockFrequencyInfo as a special "don't know" value. It also
causes a sink for frequencies as you can't ever get off a zero frequency with
more multiplies.
This recovers a 10% regression on MultiSource/Benchmarks/7zip. A zero frequency
was propagated into an inner loop causing excessive spilling.
PR16402.
llvm-svn: 184584
Other than recognizing the attribute, the patch does little else.
It changes the branch probability analyzer so that edges into
blocks postdominated by a cold function are given low weight.
Added analysis and code generation tests. Added documentation for the
new attribute.
llvm-svn: 182638
We switch the order of offset and field type to make TBAAStructType node
(name, parent node, offset) similar to scalar TBAA node (name, parent node).
TypeIsImmutable is added to TBAAStructTag node.
llvm-svn: 180654
Rather than just splitting the input type and hoping for the best, apply
a bit more cleverness. Just splitting the types until the source is
legal often leads to an illegal result time, which is then widened and a
scalarization step is introduced which leads to truly horrible code
generation. With the loop vectorizer, these sorts of operations are much
more common, and so it's worth extra effort to do them well.
Add a legalization hook for the operands of a TRUNCATE node, which will
be encountered after the result type has been legalized, but if the
operand type is still illegal. If simple splitting of both types
ends up with the result type of each half still being legal, just
do that (v16i16 -> v16i8 on ARM, for example). If, however, that would
result in an illegal result type (v8i32 -> v8i8 on ARM, for example),
we can get more clever with power-two vectors. Specifically,
split the input type, but also widen the result element size, then
concatenate the halves and truncate again. For example on ARM,
To perform a "%res = v8i8 trunc v8i32 %in" we transform to:
%inlo = v4i32 extract_subvector %in, 0
%inhi = v4i32 extract_subvector %in, 4
%lo16 = v4i16 trunc v4i32 %inlo
%hi16 = v4i16 trunc v4i32 %inhi
%in16 = v8i16 concat_vectors v4i16 %lo16, v4i16 %hi16
%res = v8i8 trunc v8i16 %in16
This allows instruction selection to generate three VMOVN instructions
instead of a sequences of moves, stores and loads.
Update the ARMTargetTransformInfo to take this improved legalization
into account.
Consider the simplified IR:
define <16 x i8> @test1(<16 x i32>* %ap) {
%a = load <16 x i32>* %ap
%tmp = trunc <16 x i32> %a to <16 x i8>
ret <16 x i8> %tmp
}
define <8 x i8> @test2(<8 x i32>* %ap) {
%a = load <8 x i32>* %ap
%tmp = trunc <8 x i32> %a to <8 x i8>
ret <8 x i8> %tmp
}
Previously, we would generate the truly hideous:
.syntax unified
.section __TEXT,__text,regular,pure_instructions
.globl _test1
.align 2
_test1: @ @test1
@ BB#0:
push {r7}
mov r7, sp
sub sp, sp, #20
bic sp, sp, #7
add r1, r0, #48
add r2, r0, #32
vld1.64 {d24, d25}, [r0:128]
vld1.64 {d16, d17}, [r1:128]
vld1.64 {d18, d19}, [r2:128]
add r1, r0, #16
vmovn.i32 d22, q8
vld1.64 {d16, d17}, [r1:128]
vmovn.i32 d20, q9
vmovn.i32 d18, q12
vmov.u16 r0, d22[3]
strb r0, [sp, #15]
vmov.u16 r0, d22[2]
strb r0, [sp, #14]
vmov.u16 r0, d22[1]
strb r0, [sp, #13]
vmov.u16 r0, d22[0]
vmovn.i32 d16, q8
strb r0, [sp, #12]
vmov.u16 r0, d20[3]
strb r0, [sp, #11]
vmov.u16 r0, d20[2]
strb r0, [sp, #10]
vmov.u16 r0, d20[1]
strb r0, [sp, #9]
vmov.u16 r0, d20[0]
strb r0, [sp, #8]
vmov.u16 r0, d18[3]
strb r0, [sp, #3]
vmov.u16 r0, d18[2]
strb r0, [sp, #2]
vmov.u16 r0, d18[1]
strb r0, [sp, #1]
vmov.u16 r0, d18[0]
strb r0, [sp]
vmov.u16 r0, d16[3]
strb r0, [sp, #7]
vmov.u16 r0, d16[2]
strb r0, [sp, #6]
vmov.u16 r0, d16[1]
strb r0, [sp, #5]
vmov.u16 r0, d16[0]
strb r0, [sp, #4]
vldmia sp, {d16, d17}
vmov r0, r1, d16
vmov r2, r3, d17
mov sp, r7
pop {r7}
bx lr
.globl _test2
.align 2
_test2: @ @test2
@ BB#0:
push {r7}
mov r7, sp
sub sp, sp, #12
bic sp, sp, #7
vld1.64 {d16, d17}, [r0:128]
add r0, r0, #16
vld1.64 {d20, d21}, [r0:128]
vmovn.i32 d18, q8
vmov.u16 r0, d18[3]
vmovn.i32 d16, q10
strb r0, [sp, #3]
vmov.u16 r0, d18[2]
strb r0, [sp, #2]
vmov.u16 r0, d18[1]
strb r0, [sp, #1]
vmov.u16 r0, d18[0]
strb r0, [sp]
vmov.u16 r0, d16[3]
strb r0, [sp, #7]
vmov.u16 r0, d16[2]
strb r0, [sp, #6]
vmov.u16 r0, d16[1]
strb r0, [sp, #5]
vmov.u16 r0, d16[0]
strb r0, [sp, #4]
ldm sp, {r0, r1}
mov sp, r7
pop {r7}
bx lr
Now, however, we generate the much more straightforward:
.syntax unified
.section __TEXT,__text,regular,pure_instructions
.globl _test1
.align 2
_test1: @ @test1
@ BB#0:
add r1, r0, #48
add r2, r0, #32
vld1.64 {d20, d21}, [r0:128]
vld1.64 {d16, d17}, [r1:128]
add r1, r0, #16
vld1.64 {d18, d19}, [r2:128]
vld1.64 {d22, d23}, [r1:128]
vmovn.i32 d17, q8
vmovn.i32 d16, q9
vmovn.i32 d18, q10
vmovn.i32 d19, q11
vmovn.i16 d17, q8
vmovn.i16 d16, q9
vmov r0, r1, d16
vmov r2, r3, d17
bx lr
.globl _test2
.align 2
_test2: @ @test2
@ BB#0:
vld1.64 {d16, d17}, [r0:128]
add r0, r0, #16
vld1.64 {d18, d19}, [r0:128]
vmovn.i32 d16, q8
vmovn.i32 d17, q9
vmovn.i16 d16, q8
vmov r0, r1, d16
bx lr
llvm-svn: 179989
The costs are overfitted so that I can still use the legalization factor.
For example the following kernel has about half the throughput vectorized than
unvectorized when compiled with SSE2. Before this patch we would vectorize it.
unsigned short A[1024];
double B[1024];
void f() {
int i;
for (i = 0; i < 1024; ++i) {
B[i] = (double) A[i];
}
}
radar://13599001
llvm-svn: 179033
The code in getTypeConversion attempts to promote the element vector type
before it trys to split or widen the vector.
After it failed finding a legal vector type by promoting it would continue using
the promoted vector element type. Thereby missing legal splitted vector types.
For example the type v32i32 that has a legal split of 4 x v3i32 on x86/sse2
would be transformed to: v32i256 and from there on successively split to:
v16i256, v8i256, v1i256 and then finally ends up as an i64 type.
By resetting the vector element type to the original vector element type that
existed before the promotion the code will attempt to split the vector type to
smaller vector widths of the same type.
llvm-svn: 178999
SSE2 has efficient support for shifts by a scalar. My previous change of making
shifts expensive did not take this into account marking all shifts as expensive.
This would prevent vectorization from happening where it is actually beneficial.
With this change we differentiate between shifts of constants and other shifts.
radar://13576547
llvm-svn: 178808
The default logic does not correctly identify costs of casts because they are
marked as custom on x86.
For some cases, where the shift amount is a scalar we would be able to generate
better code. Unfortunately, when this is the case the value (the splat) will get
hoisted out of the loop, thereby making it invisible to ISel.
radar://13130673
radar://13537826
llvm-svn: 178703
Fixes PR15570: SEGV: SCEV back-edge info invalid after dead code removal.
Indvars creates a SCEV expression for the loop's back edge taken
count, then determines that the comparison is always true and
removes it.
When loop-unroll asks for the expression, it contains a NULL
SCEVUnknkown (as a CallbackVH).
forgetMemoizedResults should invalidate the loop back edges expression.
llvm-svn: 177986
Chandler Carruth