Commit Graph

164 Commits

Author SHA1 Message Date
Chris Lattner
6038a6351e some comment improvements.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123243 91177308-0d34-0410-b5e6-96231b3b80d8
2011-01-11 17:11:59 +00:00
Chris Lattner
992efb0378 Step #2 to improve trip count analysis for loops like this:
void f(int* begin, int* end) { std::fill(begin, end, 0); }

which turns into a != exit expression where one pointer is
strided and (thanks to step #1) known to not overflow, and 
the other is loop invariant.

The observation here is that, though the IV is strided by
4 in this case, that the IV *has* to become equal to the
end value.  It cannot "miss" the end value by stepping over
it, because if it did, the strided IV expression would
eventually wrap around.

Handle this by turning A != B into "A-B != 0" where the A-B
part is known to be NUW.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123131 91177308-0d34-0410-b5e6-96231b3b80d8
2011-01-09 22:26:35 +00:00
Chris Lattner
a92ff91a96 implement enough of the memset inference algorithm to recognize and insert
memsets.  This is still missing one important validity check, but this is enough
to compile stuff like this:

void test0(std::vector<char> &X) {
  for (std::vector<char>::iterator I = X.begin(), E = X.end(); I != E; ++I)
    *I = 0;
}

void test1(std::vector<int> &X) {
  for (long i = 0, e = X.size(); i != e; ++i)
    X[i] = 0x01010101;
}

With:
 $ clang t.cpp -S -o - -O2 -emit-llvm | opt -loop-idiom | opt -O3 | llc 

to:

__Z5test0RSt6vectorIcSaIcEE:            ## @_Z5test0RSt6vectorIcSaIcEE
## BB#0:                                ## %entry
	subq	$8, %rsp
	movq	(%rdi), %rax
	movq	8(%rdi), %rsi
	cmpq	%rsi, %rax
	je	LBB0_2
## BB#1:                                ## %bb.nph
	subq	%rax, %rsi
	movq	%rax, %rdi
	callq	___bzero
LBB0_2:                                 ## %for.end
	addq	$8, %rsp
	ret
...
__Z5test1RSt6vectorIiSaIiEE:            ## @_Z5test1RSt6vectorIiSaIiEE
## BB#0:                                ## %entry
	subq	$8, %rsp
	movq	(%rdi), %rax
	movq	8(%rdi), %rdx
	subq	%rax, %rdx
	cmpq	$4, %rdx
	jb	LBB1_2
## BB#1:                                ## %for.body.preheader
	andq	$-4, %rdx
	movl	$1, %esi
	movq	%rax, %rdi
	callq	_memset
LBB1_2:                                 ## %for.end
	addq	$8, %rsp
	ret



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@122573 91177308-0d34-0410-b5e6-96231b3b80d8
2010-12-26 23:42:51 +00:00
Michael J. Spencer
1f6efa3996 Merge System into Support.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120298 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-29 18:16:10 +00:00
Dan Gohman
9c9fcfc719 Introduce memoization for ScalarEvolution dominates and properlyDominates
queries, and SCEVExpander getRelevantLoop queries.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@119595 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-18 00:34:22 +00:00
Dan Gohman
56a7568218 Factor out the code for purging a SCEV from all the various memoization maps.
Some of these maps may merge in the future, but for now it's convenient to have
a utility function for them.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@119587 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-17 23:28:48 +00:00
Dan Gohman
714b5290b0 Merge the implementations of isLoopInvariant and hasComputableLoopEvolution, and
memoize the results. This improves compile time in code which highly complex
expressions which get queried many times.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@119584 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-17 23:21:44 +00:00
Dan Gohman
4ce32db913 Make SCEV::getType() and SCEV::print non-virtual. Move SCEV::hasOperand
to ScalarEvolution. Delete SCEV::~SCEV. SCEV is no longer virtual.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@119578 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-17 22:27:42 +00:00
Dan Gohman
dc0e8fb9f9 Move SCEV::dominates and properlyDominates to ScalarEvolution.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@119570 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-17 21:41:58 +00:00
Dan Gohman
17ead4ff4b Move SCEV::isLoopInvariant and hasComputableLoopEvolution to be member
functions of ScalarEvolution, in preparation for memoization and
other optimizations.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@119562 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-17 21:23:15 +00:00
Dan Gohman
7c0fd8eb72 Fix ScalarEvolution's range memoization to avoid using a
default ctor with ConstantRange.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@119550 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-17 20:23:08 +00:00
Dan Gohman
6678e7b6eb Memoize results from ScalarEvolution's getUnsignedRange and getSignedRange.
This fixes some extreme compile times on unrolled sha512 code.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@119455 91177308-0d34-0410-b5e6-96231b3b80d8
2010-11-17 02:44:44 +00:00
Dan Gohman
e8ac3f3be4 Switch ScalarEvolution's main Value*->SCEV* map from std::map
to DenseMap.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112281 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-27 18:55:03 +00:00
Dan Gohman
081ad68e68 Placate overzealous compiler warnings.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@111152 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-16 17:01:55 +00:00
Dan Gohman
aadb5f5adb Specialize FoldingSetTrait<SCEV>, providing implementations of node
comparison and hash computation which don't require constructing
temporary ID values.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@111131 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-16 15:31:45 +00:00
Dan Gohman
af08a36bd6 Rename and reorder the arguments to isImpliedCond, for consistency and clarity.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110750 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-10 23:46:30 +00:00
Dan Gohman
ab37f50838 Make SCEVUnknown a CallbackVH, so that it can be notified directly
of Value deletions and RAUWs, instead of relying on ScalarEvolution's
Scalars map being notified, as that's complicated at best, and
insufficient in general.

This means SCEVUnknown needs a non-trivial destructor, so introduce
a mechanism to allow ScalarEvolution to locate all the SCEVUnknowns.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110086 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-02 23:49:30 +00:00
Eric Christopher
e6cbfa6b09 Speculatively revert r109705 since it seems to be causing some build bot
angst.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@109718 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-29 01:25:38 +00:00
Dan Gohman
c6743207e2 Factor out some of the code for updating old SCEVUnknown values, and
extend it to handle the case where multiple RAUWs affect a single
SCEVUnknown.

Add a ScalarEvolution unittest to test for this situation.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@109705 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-29 00:17:55 +00:00
Dan Gohman
dc7a235363 Add some extra friend declarations to fix a gcc-4.0 compile error.
This is a temporary fix, until more elaborate changes are ready.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@109593 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-28 17:09:24 +00:00
Dan Gohman
605c14fb11 Replace ScalarEvolution's private copy of getLoopPredecessor
with LoopInfo's public copy.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106603 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-22 23:43:28 +00:00
Dan Gohman
3bf63768e5 Revert r106304 (105548 and friends), which are the SCEVComplexityCompare
optimizations. There is still some nondeterminism remaining.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106306 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-18 19:54:20 +00:00
Dan Gohman
78db186d2d Reapply 105540, 105542, and 105548, and revert r105732.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106304 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-18 19:26:04 +00:00
Dan Gohman
07587a4500 Remove getIntegerSCEV; it's redundant with getConstant, and getConstant
is more consistent with the ConstantInt API.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106281 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-18 14:33:50 +00:00
Evan Cheng
1d451dff63 Revert 105540, 105542, 105544, 105546, and 105548 to unbreak bootstrapping.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@105740 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-09 18:59:43 +00:00
Dan Gohman
fd447eff22 The FoldingSet hash data includes pointer values, so it isn't
determinstic. Instead, give SCEV objects an arbitrary sequence
number.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@105548 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-07 19:36:14 +00:00
Dan Gohman
4d52c6d622 Optimize ScalarEvolution's SCEVComplexityCompare predicate: don't go
scrounging through SCEVUnknown contents and SCEVNAryExpr operands;
instead just do a simple deterministic comparison of the precomputed
hash data.

Also, since this is more precise, it eliminates the need for the slow
N^2 duplicate detection code.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@105540 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-07 19:06:13 +00:00
Dan Gohman
5cc6f9ba47 Add some comments.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@102731 91177308-0d34-0410-b5e6-96231b3b80d8
2010-04-30 18:33:41 +00:00
Dan Gohman
e979650fa5 Add a new utility function SimplifyICmpOperands. Much of this code is
refactored out of ScalarEvolution::isImpliedCond, which will be updated
to use this new utility routine soon.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@102229 91177308-0d34-0410-b5e6-96231b3b80d8
2010-04-24 01:28:42 +00:00
Dan Gohman
005752bbe7 Make getPredecessorWithUniqueSuccessorForBB return the unique successor
in addition to the predecessor.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@101374 91177308-0d34-0410-b5e6-96231b3b80d8
2010-04-15 16:19:08 +00:00
Dan Gohman
53c66eacc4 Enhance ScalarEvolution::isKnownPredicate with support for
loop conditions which are invariants.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@100995 91177308-0d34-0410-b5e6-96231b3b80d8
2010-04-11 22:16:48 +00:00
Dan Gohman
e48172710b Fix a comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@100992 91177308-0d34-0410-b5e6-96231b3b80d8
2010-04-11 22:07:56 +00:00
Dan Gohman
3948d0b8b0 Rename isLoopGuardedByCond to isLoopEntryGuardedByCond, to emphasise
that it's only testing for the entry condition, not full loop-invariant
conditions.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@100979 91177308-0d34-0410-b5e6-96231b3b80d8
2010-04-11 19:27:13 +00:00
Dan Gohman
c93b4cff89 Add the ability to "intern" FoldingSetNodeID data into a
BumpPtrAllocator-allocated region to allow it to be stored in a more
compact form and to avoid the need for a non-trivial destructor call.

Use this new mechanism in ScalarEvolution instead of
FastFoldingSetNode to avoid leaking memory in the case where a
FoldingSetNodeID uses heap storage, and to reduce overall memory
usage.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@98829 91177308-0d34-0410-b5e6-96231b3b80d8
2010-03-18 16:16:38 +00:00
Dan Gohman
3f46a3abee Spelling fixes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@97453 91177308-0d34-0410-b5e6-96231b3b80d8
2010-03-01 17:49:51 +00:00
Dan Gohman
f6d009fb6f Convert a few more backedge-taken count functions to use BackedgeTakenInfo.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@97042 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-24 17:31:30 +00:00
Dale Johannesen
45a2d7d44a recommit 96626, evidence that it broke things appears
to be spurious



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96662 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-19 07:14:22 +00:00
Dale Johannesen
efd9e95051 Revert 96626, which causes build failure on ppc Darwin.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96653 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-19 01:54:37 +00:00
Dan Gohman
aa11defd1c Indvars needs to explicitly notify ScalarEvolution when it is replacing
a loop exit value, so that if a loop gets deleted, ScalarEvolution
isn't stick holding on to dangling SCEVAddRecExprs for that loop. This
fixes PR6339.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96626 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-18 23:26:33 +00:00
Dan Gohman
32efba698d Change the argument to getIntegerSCEV to be an int64_t, rather
than int. This will make it more convenient for LSR, which does
a lot of things with int64_t offsets.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@95281 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-04 02:43:51 +00:00
Dan Gohman
ebdcbc26c7 Fix function names in comments. Thanks Duncan!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@95126 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-02 21:10:27 +00:00
Dan Gohman
4f8eea82d8 Generalize target-independent folding rules for sizeof to handle more
cases, and implement target-independent folding rules for alignof and
offsetof. Also, reassociate reassociative operators when it leads to
more folding.

Generalize ScalarEvolution's isOffsetOf to recognize offsetof on
arrays. Rename getAllocSizeExpr to getSizeOfExpr, and getFieldOffsetExpr
to getOffsetOfExpr, for consistency with analagous ConstantExpr routines.

Make the target-dependent folder promote GEP array indices to
pointer-sized integers, to make implicit casting explicit and exposed
to subsequent folding.

And add a bunch of testcases for this new functionality, and a bunch
of related existing functionality.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@94987 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-01 18:27:38 +00:00
Benjamin Kramer
629bff692a struct/class mismatch.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@94075 91177308-0d34-0410-b5e6-96231b3b80d8
2010-01-21 15:40:14 +00:00
Dan Gohman
1cd9275c8d Give ScalarEvolution access to the DominatorTree. It'll need this
to make more intellegent AddRec folding decisions.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@93930 91177308-0d34-0410-b5e6-96231b3b80d8
2010-01-19 22:21:27 +00:00
Dan Gohman
d281ed2d03 Preserve NSW information in more places.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@91656 91177308-0d34-0410-b5e6-96231b3b80d8
2009-12-18 02:09:29 +00:00
Dan Gohman
4c7279ac72 Rename forgetLoopBackedgeTakenCount to forgetLoop, because it
clears out more information than just the stored backedge taken count.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85664 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-31 15:04:55 +00:00
Chandler Carruth
8b67f774e9 Move DataTypes.h to include/llvm/System, update all users. This breaks the last
direct inclusion edge from System to Support.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85086 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-26 01:35:46 +00:00
Dan Gohman
3645b01002 Add the ability to track HasNSW and HasNUW on more kinds of SCEV expressions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83601 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-09 00:10:36 +00:00
Dan Gohman
6e70e31810 Add a properlyDominates member function to ScalarEvolution.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@82898 91177308-0d34-0410-b5e6-96231b3b80d8
2009-09-27 15:26:03 +00:00
Dan Gohman
1f96e67f78 Teach ScalarEvolution how to reason about no-wrap flags on loops
where the induction variable has a non-unit stride, such as {0,+,2}, and
there are expressions such as {1,+,2} inside the loop formed with
or or add nsw operators.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@82151 91177308-0d34-0410-b5e6-96231b3b80d8
2009-09-17 18:05:20 +00:00