Summary:
The immediate post dominator of the loop header may be part of the divergent loop.
Since this /was/ the divergence propagation bound the SDA would not detect joins of divergent paths outside the loop.
Reviewers: nhaehnle
Reviewed By: nhaehnle
Subscribers: mmasten, arsenm, jvesely, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D59042
llvm-svn: 358681
isSafeToExpand was making a common, but dangerously wrong, mistake in assuming that if any instruction within a basic block executes, that all instructions within that block must execute. This can be trivially shown to be false by considering the following small example:
bb:
add x, y <-- InsertionPoint
call @throws()
udiv x, y <-- SCEV* S
br ...
It's clearly not legal to expand S above the throwing call, but the previous logic would do so since S dominates (but not properlyDominates) the block containing the InsertionPoint.
Since iterating instructions w/in a block is expensive, this change special cases two cases: 1) S is an operand of InsertionPoint, and 2) InsertionPoint is the terminator of it's block. These two together are enough to keep all current optimizations triggering while fixing the latent correctness issue.
As best I can tell, this is a silent bug in current ToT. Given that, there's no tests with this change. It was noticed in an upcoming optimization change (D60093), and was reviewed as part of that. That change will include the test which caused me to notice the issue. I'm submitting this seperately so that anyone bisecting a problem gets a clear explanation.
llvm-svn: 358680
If a branch is conditional on extractvalue(op.with.overflow(%x, C), 1)
then we can constrain the value of %x inside the branch based on
makeGuaranteedNoWrapRegion(). We do this by extending the edge-value
handling in LVI. This allows CVP to then fold comparisons against %x,
as illustrated in the tests.
Differential Revision: https://reviews.llvm.org/D60650
llvm-svn: 358597
This adds a WithOverflowInst class with a few helper methods to get
the underlying binop, signedness and nowrap type and makes use of it
where sensible. There will be two more uses in D60650/D60656.
The refactorings are all NFC, though I left some TODOs where things
could be improved. In particular we have two places where add/sub are
handled but mul isn't.
Differential Revision: https://reviews.llvm.org/D60668
llvm-svn: 358512
Summary:
When inserting a new Def, MemorySSA may be have non-minimal number of Phis.
While inserting, the walk to find the previous definition may cleanup minimal Phis.
When the last definition is trivial to obtain, we do not cache it.
It is possible while getting the previous definition for a Def to get two different answers:
- one that was straight-forward to find when walking the first path (a trivial phi in this case), and
- another that follows a cleanup of the trivial phi, it determines it may need additional Phi nodes, it inserts them and returns a new phi in the same position as the former trivial one.
While the Phis added for the second path are all redundant, they are not complete (the walk is only done upwards), and they are not properly cleaned up afterwards.
A way to fix this problem is to cache the straight-forward answer we got on the first walk.
The caching is only kept for the duration of a getPreviousDef call, and for Phis we use TrackingVH, so removing the trivial phi will lead to replacing it with the next dominating phi in the cache.
Resolves PR40749.
Reviewers: george.burgess.iv
Subscribers: jlebar, Prazek, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D60634
llvm-svn: 358313
Summary:
Create a method to forget everything in SCEV.
Add a cl::opt and PassManagerBuilder option to use this in LoopUnroll.
Motivation: Certain Halide applications spend a very long time compiling in forgetLoop, and prefer to forget everything and rebuild SCEV from scratch.
Sample difference in compile time reduction: 21.04 to 14.78 using current ToT release build.
Testcase showcasing this cannot be opensourced and is fairly large.
The option disabled by default, but it may be desirable to enable by
default. Evidence in favor (two difference runs on different days/ToT state):
File Before (s) After (s)
clang-9.bc 7267.91 6639.14
llvm-as.bc 194.12 194.12
llvm-dis.bc 62.50 62.50
opt.bc 1855.85 1857.53
File Before (s) After (s)
clang-9.bc 8588.70 7812.83
llvm-as.bc 196.20 194.78
llvm-dis.bc 61.55 61.97
opt.bc 1739.78 1886.26
Reviewers: sanjoy
Subscribers: mehdi_amini, jlebar, zzheng, javed.absar, dmgreen, jdoerfert, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D60144
llvm-svn: 358304
Summary:
After introducing the limit for clobber walking, `walkToPhiOrClobber` would assert that the limit is at least 1 on entry.
The test included triggered that assert.
The callsite in `tryOptimizePhi` making the calls to `walkToPhiOrClobber` is structured like this:
```
while (true) {
if (getBlockingAccess()) { // calls walkToPhiOrClobber
}
for (...) {
walkToPhiOrClobber();
}
}
```
The cleanest fix is to check if the limit was reached inside `walkToPhiOrClobber`, and give an allowence of 1.
This approach not make any alias() calls (no calls to instructionClobbersQuery), so the performance condition is enforced.
The limit is set back to 0 if not used, as this provides info on the fact that we stopped before reaching a true clobber.
Reviewers: george.burgess.iv
Subscribers: jlebar, Prazek, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D60479
llvm-svn: 358303
This is a follow-up patch to D60504 to further improve
performance issues in computeKnownBitsFromAssume.
The patch is NFC, but may improve compile-time performance
if the compiler isn't clever enough to do the optimization
itself.
llvm-svn: 358163
If the ObjectSizeOffsetEvaluator fails to fold the object size call, then it may
litter some unused instructions in the function. When done repeatably in
InstCombine, this results in an infinite loop. Fix this by tracking the set of
instructions that were inserted, then removing them on failure.
rdar://49172227
Differential revision: https://reviews.llvm.org/D60298
llvm-svn: 358146
This patch changes the order of pattern matching by first testing
a compare instruction's predicate, before doing the pattern
match for the whole expression tree.
Patch by Paul Walker.
Reviewed By: spatel
Differential Revision: https://reviews.llvm.org/D60504
llvm-svn: 358097
This is the same change as D60420 but for signed sub rather than
signed add: Range information is intersected into the known bits
result, allows to detect more no/always overflow conditions.
Differential Revision: https://reviews.llvm.org/D60469
llvm-svn: 358020
This is D59386 for the signed add case. The computeConstantRange()
result is now intersected into the existing known bits information,
allowing to detect additional no-overflow/always-overflow conditions
(though the latter isn't used yet).
This (finally...) covers the motivating case from D59071.
Differential Revision: https://reviews.llvm.org/D60420
llvm-svn: 358014
This patch factors out mappings of scalar maths functions to their vector
counterparts from TargetLibraryInfo.cpp to a separate VecFuncs.def file. Such
mappings are currently available for Accelerate framework, and SVML library.
This is in support of the follow-up: https://reviews.llvm.org/D59881
Patch by pjeeva01
Differential revision: https://reviews.llvm.org/D60211
llvm-svn: 358001
Switch part of the computeOverflowForSignedAdd() implementation to
use Range.isAllNegative() rather than KnownBits.isNegative() and
similar. They do the same thing, but using the ConstantRange methods
allows dropping the KnownBits variables more easily in D60420.
llvm-svn: 357969
Add support for min/max flavor selects in computeConstantRange(),
which allows us to fold comparisons of a min/max against a constant
in InstSimplify. This fixes an infinite InstCombine loop, with the
test case taken from D59378.
Relative to the previous iteration, this contains some adjustments for
AMDGPU med3 tests: The AMDGPU target runs InstSimplify prior to codegen,
which ends up constant folding some existing med3 tests after this
change. To preserve these tests a hidden -amdgpu-scalar-ir-passes option
is added, which allows disabling scalar IR passes (that use InstSimplify)
for testing purposes.
Differential Revision: https://reviews.llvm.org/D59506
llvm-svn: 357870
Current LCG doesn't check aliased functions. So if an internal function has a public alias it will not be added to CG SCC, but it is still reachable from outside through the alias.
So this patch adds aliased functions to SCC.
Differential Revision: https://reviews.llvm.org/D59898
llvm-svn: 357795
A block reachable from the entry block can't have any route to a block that's not reachable from the entry block (if it did, that route would make it reachable from the entry block). That is the intended performance optimization for isPotentiallyReachable. For the case where we ask whether an unreachable from entry block has a route to a reachable from entry block, we can't conclude one way or the other. Fix a bug where we claimed there could be no such route.
The fix in rL357425 ironically reintroduced the very bug it was fixing but only when a DominatorTree is provided. This fixes the remaining bug.
llvm-svn: 357734
Create method `optForNone()` testing for the function level equivalent of
`-O0` and refactor appropriately.
Differential revision: https://reviews.llvm.org/D59852
llvm-svn: 357638
The Emscripten OS provides a definition of __EMSCRIPTEN__, and also that it
supports iprintf optimizations.
Also define small_printf optimizations, which is a printf with float support
but not long double (which in wasm can be useful since long doubles are 128
bit and force linking of float128 emulation code). This part is based on
sunfish's https://reviews.llvm.org/D57620 (which can't land yet since
the WASI integration isn't ready yet).
Differential Revision: https://reviews.llvm.org/D60167
llvm-svn: 357552
The code was failing to actually check for the presence of the call to widenable_condition. The whole point of specifying the widenable_condition intrinsic was allowing widening transforms. A normal branch is not widenable. A normal branch leading to a deopt is not widenable (in general).
I added a test case via LoopPredication, but GuardWidening has an analogous bug. Those are the only two passes actually using this utility just yet. Noticed while working on LoopPredication for non-widenable branches; POC in D60111.
llvm-svn: 357493
Summary:
This lets us avoid e.g. checking if A >=s B in getSMaxExpr(A, B) if we've
already established that (A smax B) is the best we can do.
Fixes PR41225.
Reviewers: asbirlea
Subscribers: mcrosier, jlebar, bixia, jdoerfert, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D60010
llvm-svn: 357320
Updated to use DenseMap::insert instead of [] operator for insertion, to
avoid a crash caused by epoch checks.
This reverts commit 2b85de4383.
llvm-svn: 357257
By extending OrderedBB to allow removing and replacing cached
instructions, we can preserve OrderedBBs in DSE easily. This eliminates
one source of quadratic compile time in DSE.
Fixes PR38829.
Reviewers: rnk, efriedma, hfinkel
Reviewed By: efriedma
Differential Revision: https://reviews.llvm.org/D59789
llvm-svn: 357208
If the caller can preserve the OBB, we can avoid recomputing the order
for each getDependency call.
Reviewers: efriedma, rnk, hfinkel
Reviewed By: rnk
Differential Revision: https://reviews.llvm.org/D59788
llvm-svn: 357206
The issue here is that we actually allow CGSCC passes to mutate IR (and
therefore invalidate analyses) outside of the current SCC. At a minimum,
we need to support mutating parent and ancestor SCCs to support the
ArgumentPromotion pass which rewrites all calls to a function.
However, the analysis invalidation infrastructure is heavily based
around not needing to invalidate the same IR-unit at multiple levels.
With Loop passes for example, they don't invalidate other Loops. So we
need to customize how we handle CGSCC invalidation. Doing this without
gratuitously re-running analyses is even harder. I've avoided most of
these by using an out-of-band preserved set to accumulate the cross-SCC
invalidation, but it still isn't perfect in the case of re-visiting the
same SCC repeatedly *but* it coming off the worklist. Unclear how
important this use case really is, but I wanted to call it out.
Another wrinkle is that in order for this to successfully propagate to
function analyses, we have to make sure we have a proxy from the SCC to
the Function level. That requires pre-creating the necessary proxy.
The motivating test case now works cleanly and is added for
ArgumentPromotion.
Thanks for the review from Philip and Wei!
Differential Revision: https://reviews.llvm.org/D59869
llvm-svn: 357137
This adds ConstantRange::getFull(BitWidth) and
ConstantRange::getEmpty(BitWidth) named constructors as more readable
alternatives to the current ConstantRange(BitWidth, /* full */ false)
and similar. Additionally private getFull() and getEmpty() member
functions are added which return a full/empty range with the same bit
width -- these are commonly needed inside ConstantRange.cpp.
The IsFullSet argument in the ConstantRange(BitWidth, IsFullSet)
constructor is now mandatory for the few usages that still make use of it.
Differential Revision: https://reviews.llvm.org/D59716
llvm-svn: 356852
We're already computing the known bits of the operands here. If the
known bits of the operands can determine the sign bit of the result,
we'll already catch this in signedAddMayOverflow(). The only other
way (and as the comment already indicates) we'll get new information
from computing known bits on the whole add, is if there's an assumption
on it.
As such, we change the code to only compute known bits from assumptions,
instead of computing full known bits on the add (which would unnecessarily
recompute the known bits of the operands as well).
Differential Revision: https://reviews.llvm.org/D59473
llvm-svn: 356785
Summary:
Adding contained caching to AliasAnalysis. BasicAA is currently the only one using it.
AA changes:
- This patch is pulling the caches from BasicAAResults to AAResults, meaning the getModRefInfo call benefits from the IsCapturedCache as well when in "batch mode".
- All AAResultBase implementations add the QueryInfo member to all APIs. AAResults APIs maintain wrapper APIs such that all alias()/getModRefInfo call sites are unchanged.
- AA now provides a BatchAAResults type as a wrapper to AAResults. It keeps the AAResults instance and a QueryInfo instantiated to batch mode. It delegates all work to the AAResults instance with the batched QueryInfo. More API wrappers may be needed in BatchAAResults; only the minimum needed is currently added.
MemorySSA changes:
- All walkers are now templated on the AA used (AliasAnalysis=AAResults or BatchAAResults).
- At build time, we optimize uses; now we create a local walker (lives only as long as OptimizeUses does) using BatchAAResults.
- All Walkers have an internal AA and only use that now, never the AA in MemorySSA. The Walkers receive the AA they will use when built.
- The walker we use for queries after the build is instantiated on AliasAnalysis and is built after building MemorySSA and setting AA.
- All static methods doing walking are now templated on AliasAnalysisType if they are used both during build and after. If used only during build, the method now only takes a BatchAAResults. If used only after build, the method now takes an AliasAnalysis.
Subscribers: sanjoy, arsenm, jvesely, nhaehnle, jlebar, george.burgess.iv, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D59315
llvm-svn: 356783
Summary:
In C++, the behavior of casting a double value that is beyond the range
of a single precision floating-point to a float value is undefined. This
change replaces such a cast with APFloat::convert to convert the value,
which is consistent with how we convert a double value to a half value.
Reviewers: sanjoy
Subscribers: lebedev.ri, sanjoy, jlebar, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D59500
llvm-svn: 356781
This adds support for scalarizing these intrinsics as well the X86TargetTransformInfo support to avoid scalarizing them in the cases X86 can handle.
I've omitted handling special cases for constant masks for this first pass. Though CodeGenPrepare can constant fold the branch conditions and remove some of the control flow anyway.
Fixes PR40994 and is covers most of PR3666. Might want to implement constant masks to close that.
Differential Revision: https://reviews.llvm.org/D59180
llvm-svn: 356687
This is D59450, but for signed sub. This case is not NFC, because
the overflow logic in ConstantRange is more powerful than the existing
check. This resolves the TODO in the function.
I've added two tests to show that this indeed catches more cases than
the previous logic, but the main correctness test coverage here is in
the existing ConstantRange unit tests.
Differential Revision: https://reviews.llvm.org/D59617
llvm-svn: 356685
Summary:
This is a refactoring patch.
- Reduce the number of map searches by reusing the iterator.
- Add asserts to check that the entry is in the cache, as this is something BasicAA relies on to avoid infinite recursion.
Reviewers: chandlerc, aschwaighofer
Subscribers: sanjoy, jlebar, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D59151
llvm-svn: 356644
Nicolai Haehnle