Do not call SplitBlockPredecessors on a loop preheader when one of the
predecessors is an indirectbr. Otherwise, you will hit this assert:
!isa<IndirectBrInst>(Preds[i]->getTerminator()) && "Cannot split an edge from an IndirectBrInst"
llvm-svn: 153134
alignment. If that's the case, then we want to make sure that we don't increase
the alignment of the store instruction. Because if we increase it to be "more
aligned" than the pointer, code-gen may use instructions which require a greater
alignment than the pointer guarantees.
<rdar://problem/11043589>
llvm-svn: 152907
It was added in 2007 as the first cut at supporting no-inline
attributes, but we didn't have function attributes of any form at the
time. However, it was added without any mention in the LangRef or other
documentation.
Later on, in 2008, Devang added function notes for 'inline=never' and
then turned them into proper function attributes. From that point
onward, as far as I can tell, the world moved on, and no one has touched
'llvm.noinline' in any meaningful way since.
It's time has now come. We have had better mechanisms for doing this for
a long time, all the frontends I'm aware of use them, and this is just
holding back progress. Given that it was never a documented feature of
the IR, I've provided no auto-upgrade support. If people know of real,
in-the-wild bitcode that relies on this, yell at me and I'll add it, but
I *seriously* doubt anyone cares.
llvm-svn: 152904
directly query the function information which this set was representing.
This simplifies the interface of the inline cost analysis, and makes the
always-inline pass significantly more efficient.
Previously, always-inline would first make a single set of every
function in the module *except* those marked with the always-inline
attribute. It would then query this set at every call site to see if the
function was a member of the set, and if so, refuse to inline it. This
is quite wasteful. Instead, simply check the function attribute directly
when looking at the callsite.
The normal inliner also had similar redundancy. It added every function
in the module with the noinline attribute to its set to ignore, even
though inside the cost analysis function we *already tested* the
noinline attribute and produced the same result.
The only tricky part of removing this is that we have to be able to
correctly remove only the functions inlined by the always-inline pass
when finalizing, which requires a bit of a hack. Still, much less of
a hack than the set of all non-always-inline functions was. While I was
touching this function, I switched a heavy-weight set to a vector with
sort+unique. The algorithm already had a two-phase insert and removal
pattern, we were just needlessly paying the uniquing cost on every
insert.
This probably speeds up some compiles by a small amount (-O0 compiles
with lots of always-inline, so potentially heavy libc++ users), but I've
not tried to measure it.
I believe there is no functional change here, but yell if you spot one.
None are intended.
Finally, the direction this is going in is to greatly simplify the
inline cost query interface so that we can replace its implementation
with a much more clever one. Along the way, all the APIs get simplified,
so it seems incrementally good.
llvm-svn: 152903
Only record IVUsers that are dominated by simplified loop
headers. Otherwise SCEVExpander will crash while looking for a
preheader.
I previously tried to work around this in LSR itself, but that was
insufficient. This way, LSR can continue to run if some uses are not
in simple loops, as long as we don't attempt to analyze those users.
Fixes <rdar://problem/11049788> Segmentation fault: 11 in LoopStrengthReduce
llvm-svn: 152892
which are small enough to themselves be inlined. Delaying in this manner
can be harmful if the function is inelligible for inlining in some (or
many) contexts as it pessimizes the code of the function itself in the
event that inlining does not eventually happen.
Previously the check was written to only do this delaying of inlining
for static functions in the hope that they could be entirely deleted and
in the knowledge that all callers of static functions will have the
opportunity to inline if it is in fact profitable. However, with C++ we
get two other important sources of functions where the definition is
always available for inlining: inline functions and templated functions.
This patch generalizes the inliner to allow linkonce-ODR (the linkage
such C++ routines receive) to also qualify for this delay-based
inlining.
Benchmarking across a range of large real-world applications shows
roughly 2% size increase across the board, but an average speedup of
about 0.5%. Some benhcmarks improved over 2%, and the 'clang' binary
itself (when bootstrapped with this feature) shows a 1% -O0 performance
improvement when run over all Sema, Lex, and Parse source code smashed
into a single file. A clean re-build of Clang+LLVM with a bootstrapped
Clang shows approximately 2% improvement, but that measurement is often
noisy.
llvm-svn: 152737
candidate set for subsequent inlining, try to simplify the arguments to
the inner call site now that inlining has been performed.
The goal here is to propagate and fold constants through deeply nested
call chains. Without doing this, we loose the inliner bonus that should
be applied because the arguments don't match the exact pattern the cost
estimator uses.
Reviewed on IRC by Benjamin Kramer.
llvm-svn: 152556
Renamed methods caseBegin, caseEnd and caseDefault with case_begin, case_end, and case_default.
Added some notes relative to case iterators.
llvm-svn: 152532
traversal, consider nodes for which the only successors are backedges
which the traversal is ignoring to be exit nodes. This fixes a problem
where the bottom-up traversal was failing to visit split blocks along
split loop backedges. This fixes rdar://10989035.
llvm-svn: 152421
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20120130/136146.html
Implemented CaseIterator and it solves almost all described issues: we don't need to mix operand/case/successor indexing anymore. Base iterator class is implemented as a template since it may be initialized either from "const SwitchInst*" or from "SwitchInst*".
ConstCaseIt is just a read-only iterator.
CaseIt is read-write iterator; it allows to change case successor and case value.
Usage of iterator allows totally remove resolveXXXX methods. All indexing convertions done automatically inside the iterator's getters.
Main way of iterator usage looks like this:
SwitchInst *SI = ... // intialize it somehow
for (SwitchInst::CaseIt i = SI->caseBegin(), e = SI->caseEnd(); i != e; ++i) {
BasicBlock *BB = i.getCaseSuccessor();
ConstantInt *V = i.getCaseValue();
// Do something.
}
If you want to convert case number to TerminatorInst successor index, just use getSuccessorIndex iterator's method.
If you want initialize iterator from TerminatorInst successor index, use CaseIt::fromSuccessorIndex(...) method.
There are also related changes in llvm-clients: klee and clang.
llvm-svn: 152297
This implicitly fixes a nasty bug in the GVN hashing (that thankfully
could only manifest as a performance bug): actually include the opcode
in the hash. The old code started the hash off with the opcode, but then
overwrote it with the type pointer.
Since this is likely to be pretty hot (GVN being already pretty
expensive) I've included a micro-optimization to just not bother with
the varargs hashing if they aren't present. I can't measure any change
in GVN performance due to this, even with a big test case like Duncan's
sqlite one. Everything I see is in the noise floor. That said, this
closes a loop hole for a potential scaling problem due to collisions if
the opcode were the differentiating aspect of the expression.
llvm-svn: 152025
equalities into phi node operands for which the equality is known to
hold in the incoming basic block. That's because replaceAllDominatedUsesWith
wasn't handling phi nodes correctly in general (that this didn't give wrong
results was just luck: the specific way GVN uses replaceAllDominatedUsesWith
precluded wrong changes to phi nodes).
llvm-svn: 152006
Some BBs can become dead after codegen preparation. If we delete them here, it
could help enable tail-call optimizations later on.
<rdar://problem/10256573>
llvm-svn: 152002
This change replaces getTypeStoreSize with getTypeAllocSize in AddressSanitizer
instrumentation for stack allocations.
One case where old behaviour produced undesired results is an optimization in
InstCombine pass (PromoteCastOfAllocation), which can replace alloca(T) with
alloca(S), where S has the same AllocSize, but a smaller StoreSize. Another
case is memcpy(long double => long double), where ASan will poison bytes 10-15
of a stack-allocated long double (StoreSize 10, AllocSize 16,
sizeof(long double) = 16).
See http://llvm.org/bugs/show_bug.cgi?id=12047 for more context.
llvm-svn: 151887
value numbers to be assigned when calculating any particular value number.
Enhance the logic that detects new value numbers to take this into account,
for a tiny compile time speedup. Fix a comment typo while there.
llvm-svn: 151522
%cmp (eg: A==B) we already replace %cmp with "true" under the true edge, and
with "false" under the false edge. This change enhances this to replace the
negated compare (A!=B) with "false" under the true edge and "true" under the
false edge. Reported to improve perlbench results by 1%.
llvm-svn: 151517
are optimization hints, but at -O0 we're not optimizing. This becomes a problem
when the alwaysinline attribute is abused.
rdar://10921594
llvm-svn: 151429
they'll be simple enough to simulate, and to reduce the chance we'll encounter
equal but different simple pointer constants.
This removes the symptoms from PR11352 but is not a full fix. A proper fix would
either require a guarantee that two constant objects we simulate are folded
when equal, or a different way of handling equal pointers (ie., trying a
constantexpr icmp on them to see whether we know they're equal or non-equal or
unsure).
llvm-svn: 151093
This transformation is not safe in some pathological cases (signed icmp of pointers should be an
extremely rare thing, but it's valid IR!). Add an explanatory comment.
Kudos to Duncan for pointing out this edge case (and not giving up explaining it until I finally got it).
llvm-svn: 151055
- Ignore pointer casts.
- Also expand GEPs that aren't constantexprs when they have one use or only constant indices.
- We now compile "&foo[i] - &foo[j]" into "i - j".
llvm-svn: 150961
metadata may still unwind, but only in ways that the ARC
optimizer doesn't need to consider. This permits more
aggressive optimization.
llvm-svn: 150829
useful to represent a variable that is const in the source but can't be constant
in the IR because of a non-trivial constructor. If globalopt evaluates the
constructor, and there was an invariant.start with no matching invariant.end
possible, it will mark the global constant afterwards.
llvm-svn: 150794
This folds a simple loop tail into a loop latch. It covers the common (in fortran) case of postincrement loops. It's a "free" way to expose this type of loop to downstream loop optimizations that bail out on non-canonical loops (getLoopLatch is a heavily used check).
llvm-svn: 150439
- Use unsigned literals when the desired result is unsigned. This mostly allows unsigned/signed mismatch warnings to be less noisy even if they aren't on by default.
- Remove misplaced llvm_unreachable.
- Add static to a declaration of a function on MSVC x86 only.
- Change some instances of calling a static function through a variable to simply calling that function while removing the unused variable.
llvm-svn: 150364
This allows BBVectorize to check the "unknown instruction" list in the
alias sets. This is important to prevent instruction fusing from reordering
function calls. Resolves PR11920.
llvm-svn: 150250
GlobalOpt runs early in the pipeline (before inlining) and complex class
hierarchies often introduce bitcasts or GEPs which weren't optimized away.
Teach it to ignore side-effect free instructions instead of depending on
other passes to remove them.
llvm-svn: 150174
* Most of the transforms come through intact by having each transformed load or
store copy the ordering and synchronization scope of the original.
* The transform that turns a global only accessed in main() into an alloca
(since main is non-recursive) with a store of the initial value uses an
unordered store, since it's guaranteed to be the first thing to happen in main.
(Threads may have started before main (!) but they can't have the address of a
function local before the point in the entry block we insert our code.)
* The heap-SRoA transforms are disabled in the face of atomic operations. This
can probably be improved; it seems odd to have atomic accesses to an alloca
that doesn't have its address taken.
AnalyzeGlobal keeps track of the strongest ordering found in any use of the
global. This is more information than we need right now, but it's cheap to
compute and likely to be useful.
llvm-svn: 149847
logic by half: isOnlyReachableViaThisEdge was trying to be clever and
handle the case of a branch to a basic block which is contained in a
loop. This costs a domtree lookup and is completely useless due to
GVN's position in the pass pipeline: all loops have preheaders at this
point, which means it is enough for isOnlyReachableViaThisEdge to check
that Dst has only one predecessor. (I checked this theoretical argument
by running over the entire nightly testsuite, and indeed it is so!).
llvm-svn: 149838
but with a critical fix to the SelectionDAG code that optimizes copies
from strings into immediate stores: the previous code was stopping reading
string data at the first nul. Address this by adding a new argument to
llvm::getConstantStringInfo, preserving the behavior before the patch.
llvm-svn: 149800
By default, boost the chain depth contribution of loads and stores. This will allow a load/store pair to vectorize even when it would not otherwise be long enough to satisfy the chain depth requirement.
llvm-svn: 149761
As suggested by Nick Lewycky, the tree traversal queues have been changed to SmallVectors and the associated loops have been rotated. Also, an 80-col violation was fixed.
llvm-svn: 149607
Long basic blocks with many candidate pairs (such as in the SHA implementation in Perl 5.14; thanks to Roman Divacky for the example) used to take an unacceptably-long time to compile. Instead, break long blocks into groups so that no group has too many candidate pairs.
llvm-svn: 149595
The purpose of refactoring is to hide operand roles from SwitchInst user (programmer). If you want to play with operands directly, probably you will need lower level methods than SwitchInst ones (TerminatorInst or may be User). After this patch we can reorganize SwitchInst operands and successors as we want.
What was done:
1. Changed semantics of index inside the getCaseValue method:
getCaseValue(0) means "get first case", not a condition. Use getCondition() if you want to resolve the condition. I propose don't mix SwitchInst case indexing with low level indexing (TI successors indexing, User's operands indexing), since it may be dangerous.
2. By the same reason findCaseValue(ConstantInt*) returns actual number of case value. 0 means first case, not default. If there is no case with given value, ErrorIndex will returned.
3. Added getCaseSuccessor method. I propose to avoid usage of TerminatorInst::getSuccessor if you want to resolve case successor BB. Use getCaseSuccessor instead, since internal SwitchInst organization of operands/successors is hidden and may be changed in any moment.
4. Added resolveSuccessorIndex and resolveCaseIndex. The main purpose of these methods is to see how case successors are really mapped in TerminatorInst.
4.1 "resolveSuccessorIndex" was created if you need to level down from SwitchInst to TerminatorInst. It returns TerminatorInst's successor index for given case successor.
4.2 "resolveCaseIndex" converts low level successors index to case index that curresponds to the given successor.
Note: There are also related compatability fix patches for dragonegg, klee, llvm-gcc-4.0, llvm-gcc-4.2, safecode, clang.
llvm-svn: 149481
This is the initial checkin of the basic-block autovectorization pass along with some supporting vectorization infrastructure.
Special thanks to everyone who helped review this code over the last several months (especially Tobias Grosser).
llvm-svn: 149468
Changing arguments from being passed as fixed to varargs is unsafe, as
the ABI may require they be handled differently (stack vs. register, for
example).
Remove two tests which rely on the bitcast being folded into the direct
call, which is exactly the transformation that's unsafe.
llvm-svn: 149457
Problem: LLVM needs more function attributes than currently available (32 bits).
One such proposed attribute is "address_safety", which shows that a function is being checked for address safety (by AddressSanitizer, SAFECode, etc).
Solution:
- extend the Attributes from 32 bits to 64-bits
- wrap the object into a class so that unsigned is never erroneously used instead
- change "unsigned" to "Attributes" throughout the code, including one place in clang.
- the class has no "operator uint64 ()", but it has "uint64_t Raw() " to support packing/unpacking.
- the class has "safe operator bool()" to support the common idiom: if (Attributes attr = getAttrs()) useAttrs(attr);
- The CTOR from uint64_t is marked explicit, so I had to add a few explicit CTOR calls
- Add the new attribute "address_safety". Doing it in the same commit to check that attributes beyond first 32 bits actually work.
- Some of the functions from the Attribute namespace are worth moving inside the class, but I'd prefer to have it as a separate commit.
Tested:
"make check" on Linux (32-bit and 64-bit) and Mac (10.6)
built/run spec CPU 2006 on Linux with clang -O2.
This change will break clang build in lib/CodeGen/CGCall.cpp.
The following patch will fix it.
llvm-svn: 148553
LSR has gradually been improved to more aggressively reuse existing code, particularly existing phi cycles. This exposed problems with the SCEVExpander's sloppy treatment of its insertion point. I applied some rigor to the insertion point problem that will hopefully avoid an endless bug cycle in this area. Changes:
- Always used properlyDominates to check safe code hoisting.
- The insertion point provided to SCEV is now considered a lower bound. This is usually a block terminator or the use itself. Under no cirumstance may SCEVExpander insert below this point.
- LSR is reponsible for finding a "canonical" insertion point across expansion of different expressions.
- Robust logic to determine whether IV increments are in "expanded" form and/or can be safely hoisted above some insertion point.
Fixes PR11783: SCEVExpander assert.
llvm-svn: 148535
It's becoming clear that LoopSimplify needs to unconditionally create loop preheaders. But that is a bigger fix. For now, continuing to hack LSR.
Fixes rdar://10701050 "Cannot split an edge from an IndirectBrInst" assert.
llvm-svn: 148288
Message for r148132:
LoopUnswitch: All helper data that is collected during loop-unswitch iterations was moved to separated class (LUAnalysisCache).
llvm-svn: 148215
the optimizer doesn't eliminate objc_retainBlock calls which are needed
for their side effect of copying blocks onto the heap.
This implements rdar://10361249.
llvm-svn: 148076
1. Size heuristics changed. Now we calculate number of unswitching
branches only once per loop.
2. Some checks was moved from UnswitchIfProfitable to
processCurrentLoop, since it is not changed during processCurrentLoop
iteration. It allows decide to skip some loops at an early stage.
Extended statistics:
- Added total number of instructions analyzed.
llvm-svn: 147935
with other symbols.
An object in the __cfstring section is suppoed to be filled with CFString
objects, which have a pointer to ___CFConstantStringClassReference followed by a
pointer to a __cstring. If we allow the object in the __cstring section to be
merged with another global, then it could end up in any section. Because the
linker is going to remove these symbols in the final executable, we shouldn't
bother to merge them.
<rdar://problem/10564621>
llvm-svn: 147899
These heuristics are sufficient for enabling IV chains by
default. Performance analysis has been done for i386, x86_64, and
thumbv7. The optimization is rarely important, but can significantly
speed up certain cases by eliminating spill code within the
loop. Unrolled loops are prime candidates for IV chains. In many
cases, the final code could still be improved with more target
specific optimization following LSR. The goal of this feature is for
LSR to make the best choice of induction variables.
Instruction selection may not completely take advantage of this
feature yet. As a result, there could be cases of slight code size
increase.
Code size can be worse on x86 because it doesn't support postincrement
addressing. In fact, when chains are formed, you may see redundant
address plus stride addition in the addressing mode. GenerateIVChains
tries to compensate for the common cases.
On ARM, code size increase can be mitigated by using postincrement
addressing, but downstream codegen currently misses some opportunities.
llvm-svn: 147826
After collecting chains, check if any should be materialized. If so,
hide the chained IV users from the LSR solver. LSR will only solve for
the head of the chain. GenerateIVChains will then materialize the
chained IV users by computing the IV relative to its previous value in
the chain.
In theory, chained IV users could be exposed to LSR's solver. This
would be considerably complicated to implement and I'm not aware of a
case where we need it. In practice it's more important to
intelligently prune the search space of nontrivial loops before
running the solver, otherwise the solver is often forced to prune the
most optimal solutions. Hiding the chained users does this well, so
that LSR is more likely to find the best IV for the chain as a whole.
llvm-svn: 147801
This collects a set of IV uses within the loop whose values can be
computed relative to each other in a sequence. Following checkins will
make use of this information.
llvm-svn: 147797
This will be more important as we extend the LSR pass in ways that don't rely on the formula solver. In particular, we need it for constructing IV chains.
llvm-svn: 147724
LoopSimplify may not run on some outer loops, e.g. because of indirect
branches. SCEVExpander simply cannot handle outer loops with no preheaders.
Fixes rdar://10655343 SCEVExpander segfault.
llvm-svn: 147718
present in the bottom of the CFG triangle, as the transformation isn't
ever valuable if the branch can't be eliminated.
Also, unify some heuristics between SimplifyCFG's multiple
if-converters, for consistency.
This fixes rdar://10627242.
llvm-svn: 147630
code can incorrectly move the load across a store. This never
happens in practice today, but only because the current
heuristics accidentally preclude it.
llvm-svn: 147623
captured. This allows the tracker to look at the specific use, which may be
especially interesting for function calls.
Use this to fix 'nocapture' deduction in FunctionAttrs. The existing one does
not iterate until a fixpoint and does not guarantee that it produces the same
result regardless of iteration order. The new implementation builds up a graph
of how arguments are passed from function to function, and uses a bottom-up walk
on the argument-SCCs to assign nocapture. This gets us nocapture more often, and
does so rather efficiently and independent of iteration order.
llvm-svn: 147327
This has the obvious advantage of being commutable and is always a win on x86 because
const - x wastes a register there. On less weird architectures this may lead to
a regression because other arithmetic doesn't fuse with it anymore. I'll address that
problem in a followup.
llvm-svn: 147254
performance regressions (both execution-time and compile-time) on our
nightly testers.
Original commit message:
Fix for bug #11429: Wrong behaviour for switches. Small improvement for code
size heuristics.
llvm-svn: 147131
into Analysis as a standalone function, since there's no need for
it to be in VMCore. Also, update it to use isKnownNonZero and
other goodies available in Analysis, making it more precise,
enabling more aggressive optimization.
llvm-svn: 146610
This should always be done as a matter of principal. I don't have a
case that exposes the problem. I just noticed this recently while
scanning the code and realized I meant to fix it long ago.
llvm-svn: 146438
subdirectories to traverse into.
- Originally I wanted to avoid this and just autoscan, but this has one key
flaw in that new subdirectories can not automatically trigger a rerun of the
llvm-build tool. This is particularly a pain when switching back and forth
between trees where one has added a subdirectory, as the dependencies will
tend to be wrong. This will also eliminates FIXME implicitly.
llvm-svn: 146436
detected in the forward-CFG DFS. This prevents the reverse-CFG from
visiting blocks inside loops after blocks that dominate them in the
case where loops have multiple exits.
No testcase, because this fixes a bug which in practice only shows
up in a full optimizer run, due to the use-list order.
This fixes rdar://10422791 and others.
llvm-svn: 146408
indicates whether the intrinsic has a defined result for a first
argument equal to zero. This will eventually allow these intrinsics to
accurately model the semantics of GCC's __builtin_ctz and __builtin_clz
and the X86 instructions (prior to AVX) which implement them.
This patch merely sets the stage by extending the signature of these
intrinsics and establishing auto-upgrade logic so that the old spelling
still works both in IR and in bitcode. The upgrade logic preserves the
existing (inefficient) semantics. This patch should not change any
behavior. CodeGen isn't updated because it can use the existing
semantics regardless of the flag's value.
Note that this will be followed by API updates to Clang and DragonEgg.
Reviewed by Nick Lewycky!
llvm-svn: 146357
Since we're not rewriting IVs in other loops, there's not much reason
to consider their stride when generating formulae.
This should reduce the number of useless formulas considered by LSR.
llvm-svn: 146302
Patch by Brendon Cahoon!
This extends the existing LoopUnroll and LoopUnrollPass. Brendon
measured no regressions in the llvm test suite with -unroll-runtime
enabled. This implementation works by using the existing loop
unrolling code to unroll the loop by a power-of-two (default 8). It
generates an if-then-else sequence of code prior to the loop to
execute the extra iterations before entering the unrolled loop.
llvm-svn: 146245
- Walking over pred_begin/pred_end is an expensive operation.
- PHINodes contain a value for each predecessor anyway.
- While it may look like we used to save a few iterations with the set,
be aware that getIncomingValueForBlock does a linear search on
the values of the phi node.
- Another -5% on ARMDisassembler.cpp (Release build). This was the last
entry in the profile that was obviously wasting time.
llvm-svn: 145937
It's always good to prune early, but formulae that are unsatisfactory
in their own right need to be removed before running any other pruning
heuristics. We easily avoid generating such formulae, but we need them
as an intermediate basis for forming other good formulae.
llvm-svn: 145906
- Calling getUser in a loop is much more expensive than iterating over a few instructions.
- Use it instead of the open-coded loop in AddrModeMatcher.
- 5% speedup on ARMDisassembler.cpp Release builds.
llvm-svn: 145810
The callee is usually smaller than the caller, too. This reduces the compile
time of ARMDisassembler.cpp by 32% (Release build). It still takes ages to
compile though.
llvm-svn: 145690
weak variable are compiled by different compilers, such as GCC and LLVM, while
LLVM may increase the alignment to the preferred alignment there is no reason to
think that GCC will use anything more than the ABI alignment. Since it is the
GCC version that might end up in the final program (as the linkage is weak), it
is wrong to increase the alignment of loads from the global up to the preferred
alignment as the alignment might only be the ABI alignment.
Increasing alignment up to the ABI alignment might be OK, but I'm not totally
convinced that it is. It seems better to just leave the alignment of weak
globals alone.
llvm-svn: 145413
gcc, though I thought it was older (my gcc 4.4 has it as a local patch. Whoops!)
This fixes PR10589.
Also add some debugging statements.
Remove GcnoFiles, the mapping from CompilationUnit to raw_ostream. Now that we
start by iterating over each CU and descending into them, there's no need to
maintain a mapping.
llvm-svn: 145208
The right way to check for a binary operation is
cast<BinaryOperator>. The original check: cast<Instruction> &&
numOperands() == 2 would match phi "instructions", leading to an
infinite loop in extreme corner case: a useless phi with operands
[self, constant] that prior optimization passes failed to remove,
being used in the loop by another useless phi, in turn being used by an
lshr or udiv.
Fixes PR11350: runaway iteration assertion.
llvm-svn: 144935
lead to it trying to re-mark a value marked as a constant with a different value. It also appears to trigger very rarely.
Fixes PR11357.
llvm-svn: 144352
Size of data being pointed to wasn't always being checked so some small writes were killing big writes
Fixes <rdar://problem/10426753>
llvm-svn: 144312
Currently checks alignment and killing stores on a power of 2 boundary as this is likely
to trim the size of the earlier store without breaking large vector stores into scalar ones.
Fixes <rdar://problem/10140300>
llvm-svn: 144239
Only currently done if the later store is writing to a power of 2 address or
has the same alignment as the earlier store as then its likely to not break up
large stores into smaller ones
Fixes <rdar://problem/10140300>
llvm-svn: 143630
We've been hitting asserts in this code due to the many supported
combintions of modes (iv-rewrite/no-iv-rewrite) and IV types. This
second rewrite of the code attempts to deal with these cases systematically.
llvm-svn: 143546
instructions.
This doesn't introduce any optimizations we weren't doing before (except
potentially due to pass ordering issues), now passes will eliminate them sooner
as part of their own cleanups.
llvm-svn: 142787
element types, even though the element extraction code does. It is surprising
that this bug has been here for so long. Fixes <rdar://problem/10318778>.
llvm-svn: 142740
combining of the landingpad instruction. The ObjC personality function acts
almost identically to the C++ personality function. In particular, it uses
"null" as a "catch-all" value.
llvm-svn: 142256
Some code want to check that *any* call within a function has the 'returns
twice' attribute, not just that the current function has one.
llvm-svn: 142221
profile metadata at the same time. Use it to preserve metadata attached
to a branch when re-writing it in InstCombine.
Add metadata to the canonicalize_branch InstCombine test, and check that
it is tranformed correctly.
Reviewed by Nick Lewycky!
llvm-svn: 142168
I rewrote the algorithm a while back so it doesn't require map lookup,
but neglected to change the data structure. This was caught by
llvm-gcc self host, not because there's anything special about
llvm-gcc, but because it is the only test for nondeterminism we
currently have. Unit tests don't work well for everything; we should
always try to have a nondeterminism stress test running.
Fixes PR11133: llvm-gcc self host .o mismatch after enable-iv-rewrite=false
llvm-svn: 142036
Someone more familiar with LSR should double-check that the extra cast is actually doing the right thing in the overflow cases; I'm not completely confident that's that case.
llvm-svn: 141916
would have never worked, since the element type of a vector type is never a
vector type. Also fix the conditional to be more direct in checking whether
EltTy is a vector type.
llvm-svn: 141713
IVs.
Indvars previously chose randomly between congruent IVs. Now it will
bias the decision toward IVs that SCEVExpander likes to create. This
was not done to fix any problem, it's just a welcome side effect of
factoring code.
llvm-svn: 141633
promoting allocas to preferred alignments that exceed the natural
alignment. This avoids some potentially expensive dynamic stack realignments.
The natural stack alignment is set in target data strings via the "S<size>"
option. Size is in bits and must be a multiple of 8. The natural stack alignment
defaults to "unspecified" (represented by a zero value), and the "unspecified"
value does not prevent any alignment promotions. Target maintainers that care
about avoiding promotions should explicitly add the "S<size>" option to their
target data strings.
llvm-svn: 141599
switch (n) {
case 27:
do_something(x);
...
}
the call do_something(x) will be replaced with do_something(27). In
gcc-as-one-big-file this results in the removal of about 500 lines of
bitcode (about 0.02%), so has about 1/10 of the effect of propagating
branch conditions.
llvm-svn: 141360
While I'm here, fix the related issue with strncmp, add some actual tests for strcmp and strncmp, and start using StringRef::compare for constant folding instead of using strcmp/strncmp so that the optimized IR isn't dependent on the host's implementation of strcmp.
llvm-svn: 141227
Just pull the instruction name, but don't change the order of anything
else. That keeps --debug happy and non-crashing, but doesn't change
how the worklist gets built.
llvm-svn: 141210
When updating the worklist for InstCombine, the Add/AddUsersToWorklist
functions may access the instruction(s) being added, for debug output for
example. If the instructions aren't yet added to the basic block, this
can result in a crash. Finish the instruction transformation before
adjusting the worklist instead.
rdar://10238555
llvm-svn: 141203
branch "br i1 %x, label %if_true, label %if_false" then it replaces
"%x" with "true" in places only reachable via the %if_true arm, and
with "false" in places only reachable via the %if_false arm. Except
that actually it doesn't: if value numbering shows that %y is equal
to %x then, yes, %y will be turned into true/false in this way, but
any occurrences of %x itself are not transformed. Fix this. What's
more, it's often the case that %x is an equality comparison such as
"%x = icmp eq %A, 0", in which case every occurrence of %A that is
only reachable via the %if_true arm can be replaced with 0. Implement
this and a few other variations on this theme. This reduces the number
of lines of LLVM IR in "GCC as one big file" by 0.2%. It has a bigger
impact on Ada code, typically reducing the number of lines of bitcode
by around 0.4% by removing repeated compiler generated checks. Passes
the LLVM nightly testsuite and the Ada ACATS testsuite.
llvm-svn: 141177
it's OK for the false/true destination to have multiple
predecessors as long as the extra ones are dominated by
the branch destination.
llvm-svn: 141176
This handles the case in which LSR rewrites an IV user that is a phi and
splits critical edges originating from a switch.
Fixes <rdar://problem/6453893> LSR is not splitting edges "nicely"
llvm-svn: 141059
We want heuristics to be based on accurate data, but more importantly
we don't want llvm to behave randomly. A benign trunc inserted by an
upstream pass should not cause a wild swings in optimization
level. See PR11034. It's a general problem with threshold-based
heuristics, but we can make it less bad.
llvm-svn: 140919
catch or repeated filter clauses. Teach instcombine a bunch
of tricks for simplifying landingpad clauses. Currently the
code only recognizes the GNU C++ and Ada personality functions,
but that doesn't stop it doing a bunch of "generic" transforms
which are hopefully fine for any real-world personality function.
If these "generic" transforms turn out not to be generic, they
can always be conditioned on the personality function. Probably
someone should add the ObjC++ personality function. I didn't as
I don't know anything about it.
llvm-svn: 140852
Rewriting the entire loop nest now requires -enable-lsr-nested.
See PR11035 for some performance data.
A few unit tests specifically test nested LSR, and are now under a flag.
llvm-svn: 140762
The minor bug heuristic was noticed by inspection. I added the
isLoser/isValid helpers because they will become more
important with subsequent checkins.
llvm-svn: 140580
The landing pad must accompany the invoke when it's extracted. However, if it
does, then the loop isn't properly extracted. I.e., the resulting extraction has
a loop in it. The extracted function is then extracted, etc. resulting in an
infinite loop.
llvm-svn: 140193
extract its associated landing pad block as well. However, that landing pad
block may have more than one predecessor. So split the landing pad block so that
individual landing pads have only one predecessor.
This type of transformation may produce a false positive with bugpoint.
llvm-svn: 140173
extract the landing pad block. Otherwise, there will be a situation where the
invoke's unwind edge lands on a non-landing pad.
We also forbid the user from extracting the landing pad block by itself. Again,
this is not a valid transformation.
llvm-svn: 140083
No tests; these changes aren't really interesting in the sense that the logic is the same for volatile and atomic.
I believe this completes all of the changes necessary for the optimizer to handle loads and stores correctly. I'm going to try and come up with some additional testing, though.
llvm-svn: 139533
better.
Don't immediately give up when an add operation can't be trivially
sign/zero-extended within a loop. If it has NSW/NUW flags, generate a
new expression with sign extended (non-recurrent) operand. As before,
if SCEV says that all sign extends are loop invariant, then we can
widen the operation.
llvm-svn: 139453
init.trampoline and adjust.trampoline intrinsics, into two intrinsics
like in GCC. While having one combined intrinsic is tempting, it is
not natural because typically the trampoline initialization needs to
be done in one function, and the result of adjust trampoline is needed
in a different (nested) function. To get around this llvm-gcc hacks the
nested function lowering code to insert an additional parent variable
holding the adjust.trampoline result that can be accessed from the child
function. Dragonegg doesn't have the luxury of tweaking GCC code, so it
stored the result of adjust.trampoline in the memory GCC set aside for
the trampoline itself (this is always available in the child function),
and set up some new memory (using an alloca) to hold the trampoline.
Unfortunately this breaks Go which allocates trampoline memory on the
heap and wants to use it even after the parent has exited (!). Rather
than doing even more hacks to get Go working, it seemed best to just use
two intrinsics like in GCC. Patch mostly by Sanjoy Das.
llvm-svn: 139140
This changes loop unrolling to use the same mechanism for trip count
computation as indvars. This is a stronger check that tends to unroll
more loops. A very common side-effect is that many single iteration
loops will be removed sooner. The real goal was simply to remove
dependence on canonical IVs.
x86 is break even.
ARM performance changes to expect (+ is good):
External/SPEC/CFP2000/183.equake/183.equake +13%
SingleSource/Benchmarks/Dhrystone/fldry +21%
MultiSource/Applications/spiff/spiff +3%
SingleSource/Benchmarks/Stanford/Puzzle -14%
The Puzzle regression is actually an improvement in loop optimization
that defeats GVN: rdar://problem/10065079.
llvm-svn: 139009
The landingpad instruction is required in the landing pad block. Because we're
not deleting terminating instructions, the invoke may still jump to here (see
Transforms/SCCP/2004-11-16-DeadInvoke.ll). Remove all uses of the landingpad
instruction, but keep it around until code-gen can remove the basic block.
llvm-svn: 138890
In theory this could be extended to other instructions, eg. division by zero, but it's likely that it will "miscompile" some code because people depend on div by zero not trapping. NULL pointer dereference usually leads to a crash so we should be on the safe side.
This shrinks the size of a Release clang by 16k on x86_64.
llvm-svn: 138618
We have to be careful when splitting the landing pad block, because the
landingpad instruction is required to remain as the first non-PHI of an invoke's
unwind edge. To retain this, we split the block into two blocks, moving the
predecessors within the loop to one block and the remaining predecessors to the
other. The landingpad instruction is cloned into the new blocks.
llvm-svn: 138015
SplitLandingPadPredecessors is similar to SplitBlockPredecessors in that it
splits the current block and attaches a set of predecessors to the new basic
block. However, it differs from SplitBlockPredecessors in that it's specifically
designed to handle landing pad blocks.
Two new basic blocks are created: one that is has the vector of predecessors as
its predecessors and one that has the remaining predecessors as its
predecessors. Those two new blocks then receive a cloned copy of the landingpad
instruction from the original block. The landingpad instructions are joined in a
PHI, etc. Like SplitBlockPredecessors, it updates the LLVM IR, AliasAnalysis,
DominatorTree, DominanceFrontier, LoopInfo, and LCCSA analyses.
llvm-svn: 138014
PRE needs the landing pads to have their critical edges split. Doing this for a
landing pad is non-trivial. Abandon the attempt to perform PRE when we come
across a landing pad. (Reviewed by Owen!)
llvm-svn: 137876
One way to exit the loop is through an unwind edge. However, that may involve
splitting the critical edge of the landing pad, which is non-trivial. Prevent
the transformation from rewriting the landing pad exit loop block.
llvm-svn: 137871
making random bad assumptions about instructions which are not explicitly listed.
Includes fix for rdar://9956541, a version of "undef ^ undef should return
0 because it's easier than arguing with users".
llvm-svn: 137777
This commit includes a mention of the landingpad instruction, but it's not
changing the behavior around it. I think the current behavior is correct,
though. Bill, can you double-check that?
llvm-svn: 137691
This builds off of the current scheme, but instead of llvm.eh.exception and
llvm.eh.selector, it uses the landingpad instruction. And instead of
llvm.eh.resume, it uses the resume instruction.
Because of the invariants in the landing pad instruction, a lot of code that's
currently needed to find the appropriate intrinsic calls for an invoke
instruction won't be needed once we go to the new EH scheme. The "FIXME"s tell
us what to remove after we switch.
llvm-svn: 137576
This implements the 'landingpad' instruction. It's used to indicate that a basic
block is a landing pad. There are several restrictions on its use (see
LangRef.html for more detail). These restrictions allow the exception handling
code to gather the information it needs in a much more sane way.
This patch has the definition, implementation, C interface, parsing, and bitcode
support in it.
llvm-svn: 137501
the retains and releases all use the same SSA pointer value.
Also, don't let CFG hazards disrupt nested retain+release pair
optimizations.
llvm-svn: 137399
SCEV unrolling can unroll loops with arbitrary induction variables. It
is a prerequisite for -disable-iv-rewrite performance. It is also
easily handles loops of arbitrary structure including multiple exits
and is generally more robust.
This is under a temporary option to avoid affecting default
behavior for the next couple of weeks. It is needed so that I can
checkin unit tests for updateUnloop.
llvm-svn: 137384
based on ScalarEvolution without changing the induction variable phis.
This utility is the main tool of IndVarSimplifyPass, but the pass also
restructures induction variables in strange ways that are sensitive to
pass ordering. This provides a way for other loop passes to simplify
new uses of induction variables created during transformation. The
utility may be used by any pass that preserves ScalarEvolution. Soon
LoopUnroll will use it.
The net effect in this checkin is to cleanup the IndVarSimplify pass
by factoring out the SimplifyIndVar algorithm into a standalone utility.
llvm-svn: 137197
These are not individual bug fixes. I had to rewrite a good chunk of
the unroller to make it sane. I think it was getting lucky on trivial
completely unrolled loops with no early exits. I included some fairly
simple unit tests for partial unrolling. I didn't do much stress
testing, so it may not be perfect, but should be usable now.
llvm-svn: 137190
The 'unwind' instruction was acting essentially as a placeholder, because it
would be replaced at the end of this function by a branch to the "unwind
handler". The 'unwind' instruction is going away, so use 'unreachable' instead,
which serves the same purpose as a placeholder.
llvm-svn: 137098
recurrence, the initial values low bits can sometimes be ignored.
To take advantage of this, added FoldIVUser to IndVarSimplify to fold
an IV operand into a udiv/lshr if the operator doesn't affect the
result.
-indvars -disable-iv-rewrite now transforms
i = phi i4
i1 = i0 + 1
idx = i1 >> (2 or more)
i4 = i + 4
into
i = phi i4
idx = i0 >> ...
i4 = i + 4
llvm-svn: 137013
inlined variable, based on the discussion in PR10542.
This explodes the runtime of several passes down the pipeline due to
a large number of "copies" remaining live across a large function. This
only shows up with both debug and opt, but when it does it creates
a many-minute compile when self-hosting LLVM+Clang. There are several
other cases that show these types of regressions.
All of this is tracked in PR10542, and progress is being made on fixing
the issue. Once its addressed, the re-instated, but until then this
restores the performance for self-hosting and other opt+debug builds.
Devang, let me know if this causes any trouble, or impedes fixing it in
any way, and thanks for working on this!
llvm-svn: 136953
- use SmallVectorImpl& for the function argument.
- ignore the operands on the GEP, even if they aren't constant! Much as we
pretend the malloc succeeds, we pretend that malloc + whatever-you-GEP'd-by
is not null. It's magic!
llvm-svn: 136757
Don't replace a gep/bitcast with 'undef' because that will form a "free(undef)"
which in turn means "unreachable". What we wanted was a no-op. Instead, analyze
the whole tree and look for all the instructions we need to delete first, then
delete them second, not relying on the use_list to stay consistent.
llvm-svn: 136752
This adds the 'resume' instruction class, IR parsing, and bitcode reading and
writing. The 'resume' instruction resumes propagation of an existing (in-flight)
exception whose unwinding was interrupted with a 'landingpad' instruction (to be
added later).
llvm-svn: 136589
working on x86 (at least for trivial testcases); other architectures will
need more work so that they actually emit the appropriate instructions for
orderings stricter than 'monotonic'. (As far as I can tell, the ARM, PPC,
Mips, and Alpha backends need such changes.)
llvm-svn: 136457
specified in the same file that the library itself is created. This is
more idiomatic for CMake builds, and also allows us to correctly specify
dependencies that are missed due to bugs in the GenLibDeps perl script,
or change from compiler to compiler. On Linux, this returns CMake to
a place where it can relably rebuild several targets of LLVM.
I have tried not to change the dependencies from the ones in the current
auto-generated file. The only places I've really diverged are in places
where I was seeing link failures, and added a dependency. The goal of
this patch is not to start changing the dependencies, merely to move
them into the correct location, and an explicit form that we can control
and change when necessary.
This also removes a serialization point in the build because we don't
have to scan all the libraries before we begin building various tools.
We no longer have a step of the build that regenerates a file inside the
source tree. A few other associated cleanups fall out of this.
This isn't really finished yet though. After talking to dgregor he urged
switching to a single CMake macro to construct libraries with both
sources and dependencies in the arguments. Migrating from the two macros
to that style will be a follow-up patch.
Also, llvm-config is still generated with GenLibDeps.pl, which means it
still has slightly buggy dependencies. The internal CMake
'llvm-config-like' macro uses the correct explicitly specified
dependencies however. A future patch will switch llvm-config generation
(when using CMake) to be based on these deps as well.
This may well break Windows. I'm getting a machine set up now to dig
into any failures there. If anyone can chime in with problems they see
or ideas of how to solve them for Windows, much appreciated.
llvm-svn: 136433
The new EH is more simple in many respects. Mainly, we don't have to worry about
the "llvm.eh.exception" and "llvm.eh.selector" calls being in weird places.
llvm-svn: 136339
This takes the new 'resume' instruction and turns it into a direct jump to the
caller's landing pad code. The caller's landingpad instruction is merged with
the landingpad instructions of the callee. This is a bit rough and makes some
assumptions in how the code works. But it passes a simple test.
llvm-svn: 136313
size but different element types, so that it filters out the cases
that CreateShuffleVectorCast doesn't handle. This fixes rdar://9786827.
llvm-svn: 135721
For -disable-iv-rewrite, perform LFTR without generating a new
"canonical" induction variable. Instead find the "best" existing
induction variable for use in the loop exit test and compute the final
value of that IV for use in the new loop exit test. In short,
convert to a simple eq/ne exit test as long as it's cheap to do so.
llvm-svn: 135420
is named after a common idiom (i.e., memset/memcpy). Otherwise, we can run into
infinite recursion. Ideally, the user should use the correct -fno-builtin flag,
but in case they don't we should play nicely.
rdar://9763412
llvm-svn: 135286
an assert on Darwin llvm-gcc builds.
Assertion failed: (castIsValid(op, S, Ty) && "Invalid cast!"), function Create, file /Users/buildslave/zorg/buildbot/smooshlab/slave-0.8/build.llvm-gcc-i386-darwin9-RA/llvm.src/lib/VMCore/Instructions.cpp, li\
ne 2067.
etc.
http://smooshlab.apple.com:8013/builders/llvm-gcc-i386-darwin9-RA/builds/2354
--- Reverse-merging r134893 into '.':
U include/llvm/Target/TargetData.h
U include/llvm/DerivedTypes.h
U tools/bugpoint/ExtractFunction.cpp
U unittests/Support/TypeBuilderTest.cpp
U lib/Target/ARM/ARMGlobalMerge.cpp
U lib/Target/TargetData.cpp
U lib/VMCore/Constants.cpp
U lib/VMCore/Type.cpp
U lib/VMCore/Core.cpp
U lib/Transforms/Utils/CodeExtractor.cpp
U lib/Transforms/Instrumentation/ProfilingUtils.cpp
U lib/Transforms/IPO/DeadArgumentElimination.cpp
U lib/CodeGen/SjLjEHPrepare.cpp
--- Reverse-merging r134888 into '.':
G include/llvm/DerivedTypes.h
U include/llvm/Support/TypeBuilder.h
U include/llvm/Intrinsics.h
U unittests/Analysis/ScalarEvolutionTest.cpp
U unittests/ExecutionEngine/JIT/JITTest.cpp
U unittests/ExecutionEngine/JIT/JITMemoryManagerTest.cpp
U unittests/VMCore/PassManagerTest.cpp
G unittests/Support/TypeBuilderTest.cpp
U lib/Target/MBlaze/MBlazeIntrinsicInfo.cpp
U lib/Target/Blackfin/BlackfinIntrinsicInfo.cpp
U lib/VMCore/IRBuilder.cpp
G lib/VMCore/Type.cpp
U lib/VMCore/Function.cpp
G lib/VMCore/Core.cpp
U lib/VMCore/Module.cpp
U lib/AsmParser/LLParser.cpp
U lib/Transforms/Utils/CloneFunction.cpp
G lib/Transforms/Utils/CodeExtractor.cpp
U lib/Transforms/Utils/InlineFunction.cpp
U lib/Transforms/Instrumentation/GCOVProfiling.cpp
U lib/Transforms/Scalar/ObjCARC.cpp
U lib/Transforms/Scalar/SimplifyLibCalls.cpp
U lib/Transforms/Scalar/MemCpyOptimizer.cpp
G lib/Transforms/IPO/DeadArgumentElimination.cpp
U lib/Transforms/IPO/ArgumentPromotion.cpp
U lib/Transforms/InstCombine/InstCombineCompares.cpp
U lib/Transforms/InstCombine/InstCombineAndOrXor.cpp
U lib/Transforms/InstCombine/InstCombineCalls.cpp
U lib/CodeGen/DwarfEHPrepare.cpp
U lib/CodeGen/IntrinsicLowering.cpp
U lib/Bitcode/Reader/BitcodeReader.cpp
llvm-svn: 134949
LinearFunctionTestReplace rewrite. No functionality.
I've been wanting to group the indvar subphases into sections and
order them by their logical sequence. My next checkin adds functions
related to LFTR, and doing the reorg now should help reviewers. Since,
most of the code in IndVarSimplify.cpp has recently been replaced or
will be replaced soon, obscuring blame should not be an issue. This
seems like an ideal time to shuffle the code around.
I'm happy to take more suggestions for cleaning up the code. Or if
you've been wanting to cleanup anything in this file yourself, now is
a good time.
llvm-svn: 134941
patch brings numerous advantages to LLVM. One way to look at it
is through diffstat:
109 files changed, 3005 insertions(+), 5906 deletions(-)
Removing almost 3K lines of code is a good thing. Other advantages
include:
1. Value::getType() is a simple load that can be CSE'd, not a mutating
union-find operation.
2. Types a uniqued and never move once created, defining away PATypeHolder.
3. Structs can be "named" now, and their name is part of the identity that
uniques them. This means that the compiler doesn't merge them structurally
which makes the IR much less confusing.
4. Now that there is no way to get a cycle in a type graph without a named
struct type, "upreferences" go away.
5. Type refinement is completely gone, which should make LTO much MUCH faster
in some common cases with C++ code.
6. Types are now generally immutable, so we can use "Type *" instead
"const Type *" everywhere.
Downsides of this patch are that it removes some functions from the C API,
so people using those will have to upgrade to (not yet added) new API.
"LLVM 3.0" is the right time to do this.
There are still some cleanups pending after this, this patch is large enough
as-is.
llvm-svn: 134829
The promotion code lost any alignment information, when hoisting loads and
stores out of the loop. This lead to incorrect aligned memory accesses. We now
use the largest alignment we can prove to be correct.
llvm-svn: 134520
alloca that only holds a copy of a global and we're going to replace the users
of the alloca with that global, just nuke the lifetime intrinsics. Part of
PR10121.
llvm-svn: 133905
"Reinstate r133435 and r133449 (reverted in r133499) now that the clang
self-hosted build failure has been fixed (r133512)."
Due to some additional warnings.
llvm-svn: 133700
ops.
This is a rewrite of the IV simplification algorithm used by
-disable-iv-rewrite. To avoid perturbing the default mode, I
temporarily split the driver and created SimplifyIVUsersNoRewrite. The
idea is to avoid doing opcode/pattern matching inside
IndVarSimplify. SCEV already does it. We want to optimize with the
full generality of SCEV, but optimize def-use chains top down on-demand rather
than rewriting the entire expression bottom-up. This was easy to do
for operations that SCEV can prove are identity function. So we're now
eliminating bitmasks and zero extends this way.
A result of this rewrite is that indvars -disable-iv-rewrite no longer
requires IVUsers.
llvm-svn: 133502
Change PHINodes to store simple pointers to their incoming basic blocks,
instead of full-blown Uses.
Note that this loses an optimization in SplitCriticalEdge(), because we
can no longer walk the use list of a BasicBlock to find phi nodes. See
the comment I removed starting "However, the foreach loop is slow for
blocks with lots of predecessors".
Extend replaceAllUsesWith() on a BasicBlock to also update any phi
nodes in the block's successors. This mimics what would have happened
when PHINodes were proper Users of their incoming blocks. (Note that
this only works if OldBB->replaceAllUsesWith(NewBB) is called when
OldBB still has a terminator instruction, so it still has some
successors.)
llvm-svn: 133435
Change various bits of code to make better use of the existing PHINode
API, to insulate them from forthcoming changes in how PHINodes store
their operands.
llvm-svn: 133434
all over the place in different styles and variants. Standardize on two
preferred entrypoints: one that takes a StructType and ArrayRef, and one that
takes StructType and varargs.
In cases where there isn't a struct type convenient, we now add a
ConstantStruct::getAnon method (whose name will make more sense after a few
more patches land).
It would be "really really nice" if the ConstantStruct::get and
ConstantVector::get methods didn't make temporary std::vectors.
llvm-svn: 133412
In cases such as the attached test, where the case value for a switch
destination is used in a phi node that follows the destination, it
might be better to replace that value with the condition value of the
switch, so that more blocks can be folded away with
TryToSimplifyUncondBranchFromEmptyBlock because there are less
conflicts in the phi node.
llvm-svn: 133344
type's bitwidth matches the (allocated) size of the alloca. This severely
pessimizes vector scalar replacement when the only vector type being used is
something like <3 x float> on x86 or ARM whose allocated size matches a
<4 x float>.
I hope to fix some of the flawed assumptions about allocated size throughout
scalar replacement and reenable this in most cases.
llvm-svn: 133338
spartan right now, but I plan to encode more information in this enum to improve
the correctness and reliability of SRoA. At least this first pass makes it
possible to make VectorTy an actual VectorType.
llvm-svn: 132937
might overflow. Re-typing the alloca to a larger type (e.g. double)
hoists a shift into the alloca, potentially exposing overflow in the
expression. rdar://problem/9265821
llvm-svn: 132926
intrinsics. In fact, we'll optimize a bitcast to that when possible. Detect it
when looking for the lifetime intrinsics.
No test case, noticed by inspection.
llvm-svn: 132906
pad, separating the exception and selector calls from the new lpad. Teaching
it not to do that, or to properly adjust the CFG afterwards, is out of
scope because it would require the other edges to the landing pad to be split
as well (effectively). Instead, just recover from the most likely cases
during inlining. The best long-term solution is to change the exception
representation and commit to either requiring or not requiring the more
complex edge-splitting logic; this is just a shorter-term hack.
llvm-svn: 132799
assuming that all offsets are legal vector accesses, and thus trying to access
the float member of { <2 x float>, float } as the 3rd element of the first
member.
llvm-svn: 132766
former was using the size of the entire alloca, whereas the latter was correctly using
the allocated size of the immediate type being converted (which may differ from the size
of the alloca). This fixes PR10082.
llvm-svn: 132759
then we don't want to set the destination in the indirect branch to the
destination. This is because the indirect branch needs its destinations to have
had their block addresses taken. This isn't so of the new critical edge that's
split during this process. If it turns out that the destination block has only
one predecessor, and that being a BB with an indirect branch, then it won't be
marked as 'used' and may be removed.
PR10072
llvm-svn: 132638
which edge to split by pred/succ pair, which means that we can end up splitting
the wrong edge (by case value) in the switch statement entirely. Fixes PR10031!
llvm-svn: 132535
variable. Noticed by inspection.
Simulate memset in EvaluateFunction where the target of the memset and the
value we're setting are both the null value. Fixes PR10047!
llvm-svn: 132288
transformed by the inliner into a branch to the enclosing landing pad
(when inlined through an invoke). If not so optimized, it is lowered
DWARF EH preparation into a call to _Unwind_Resume (or _Unwind_SjLj_Resume
as appropriate). Its chief advantage is that it takes both the
exception value and the selector value as arguments, meaning that there
is zero effort in recovering these; however, the frontend is required
to pass these down, which is not actually particularly difficult.
Also document the behavior of landing pads a bit better, and make it
clearer that it's okay that personality functions don't always land at
landing pads. This is just a fact of life. Don't write optimizations that
rely on pushing things over an unwind edge.
llvm-svn: 132253
- the selector for the landing pad must provide all available information
about the handlers, filters, and cleanups within that landing pad
- calls to _Unwind_Resume must be converted to branches to the enclosing
lpad so as to avoid re-entering the unwinder when the lpad claimed it
was going to handle the exception in some way
This is quite specific to libUnwind-based unwinding. In an effort to not
interfere too badly with other unwinders, and with existing hacks in frontends,
this only triggers on _Unwind_Resume (not _Unwind_Resume_or_Rethrow) and does
nothing with selectors if it cannot find a selector call for either lpad.
llvm-svn: 132200
This looks like it flagged an actual bug. Devang, please review. I added
the parentheses that change behavior, but make the behavior more closely
match commit log's intent.
llvm-svn: 132165
Use a proper worklist for use-def traversal without holding onto an
iterator. Now that we process all IV uses, we need complete logic for
resusing existing derived IV defs. See HoistStep.
llvm-svn: 132103
case of a switch instruction. Back off this optimization when this would
eliminate all of the predecessors to the latch.
Sorry, I am unable to reduce a reasonably sized test case.
rdar://9486843
llvm-svn: 132022
aligned.
Teach memcpyopt to not give up all hope when confonted with an underaligned
memcpy feeding an overaligned byval. If the *source* of the memcpy can be
determined to be adequeately aligned, or if it can be forced to be, we can
eliminate the memcpy.
This addresses PR9794. We now compile the example into:
define i32 @f(%struct.p* nocapture byval align 8 %q) nounwind ssp {
entry:
%call = call i32 @g(%struct.p* byval align 8 %q) nounwind
ret i32 %call
}
in both x86-64 and x86-32 mode. We still don't get a tailcall though,
because tailcalls apparently can't handle byval.
llvm-svn: 131884
result is non-zero. Implement an example optimization (PR9814), which allows us to
transform:
A / ((1 << B) >>u 2)
into:
A >>u (B-2)
which we compile into:
_divu3: ## @divu3
leal -2(%rsi), %ecx
shrl %cl, %edi
movl %edi, %eax
ret
instead of:
_divu3: ## @divu3
movb %sil, %cl
movl $1, %esi
shll %cl, %esi
shrl $2, %esi
movl %edi, %eax
xorl %edx, %edx
divl %esi, %eax
ret
llvm-svn: 131860
failing to form a memset, then having to delete it" but my approximation
isn't safe for self recurrent loops. Instead of doign a hack, just
do it the right way.
llvm-svn: 131858
I also changed -simplifycfg, -jump-threading and -codegenprepare to use this to produce slightly better code without any extra cleanup passes (AFAICT this was the only place in -simplifycfg where now-dead conditions of replaced terminators weren't being cleaned up). The only other user of this function is -sccp, but I didn't read that thoroughly enough to figure out whether it might be holding pointers to instructions that could be deleted by this.
llvm-svn: 131855
No functionality enabled by default. Use -disable-iv-rewrite.
Extended IVUsers to keep track of the phi that represents the users' IV.
Added the WidenIV transform to replace a narrow IV with a wide IV
by doing a one-for-one replacement of IV users instead of expanding the
SCEV expressions. [sz]exts are removed and truncs are inserted.
llvm-svn: 131744
As an example, the change to InstCombineCalls catches a common case where a call to a bitcast of a function is rewritten.
Chris, does this approach look reasonable?
llvm-svn: 131516
This adds functionality to remove size/zero extension during indvars
without generating a canonical IV and rewriting all IV users. It's
disabled by default so should have no effect on codegen. Work in progress.
llvm-svn: 130829
Only create a canonical IV for backedge taken count if it will
actually be used by LinearFunctionTestReplace. And some related
cleanup, preparing to reduce dependence on canonical IVs.
No significant effect on x86 or arm in the test-suite.
llvm-svn: 130799
model constants which can be added to base registers via add-immediate
instructions which don't require an additional register to materialize
the immediate.
llvm-svn: 130743
This obviously helps a lot if the division would be turned into a libcall
(think i64 udiv on i386), but div is also one of the few remaining instructions
on modern CPUs that become more expensive when the bitwidth gets bigger.
This also helps register pressure on i386 when dividing chars, divb needs
two 8-bit parts of a 16 bit register as input where divl uses two registers.
int foo(unsigned char a) { return a/10; }
int bar(unsigned char a, unsigned char b) { return a/b; }
compiles into (x86_64)
_foo:
imull $205, %edi, %eax
shrl $11, %eax
ret
_bar:
movzbl %dil, %eax
divb %sil, %al
movzbl %al, %eax
ret
llvm-svn: 130615
This shouldn't happen in practice because the icmp would be a constant.
Add a check so we don't miscompile code if something goes wrong.
llvm-svn: 130446
between two reads (threading).
Fix an off-by-one in the indirect counter table that I meant to revert after an
earlier experiment. Whoops!
Implement GCOV_PREFIX. Doesn't handle GCOV_PREFIX_STRIP yet.
Fix an off-by-one in string emission. Extra whoops!
Tolerate DISubprograms that have null Function*'s attached to them. I don't yet
understand what this means, but it happens when you have a global static with
a non-trivial constructor/destructor.
Fix a crash on switch statements with a single successor (default-only).
llvm-svn: 130443
a nice and tidy:
%x1 = load i32* %0, align 4
%1 = icmp eq i32 %x1, 1179403647
br i1 %1, label %if.then, label %if.end
instead of doing lots of loads and branches. May the FreeBSD bootloader
long fit in its allocated space.
llvm-svn: 130416
wider load would allow elimination of subsequent loads, and when the wider
load is still a native integer type. This eliminates a ton of loads on
various benchmarks involving struct fields, though it is somewhat hobbled
by clang not being very aggressive about field alignment.
This is yet another step along the way towards resolving PR6627.
llvm-svn: 130390
Modified LinearFunctionTestReplace to push the condition on the dead
list instead of eagerly deleting it. This can cause unnecessary
IV rewrites, which should have no effect on codegen and will not be an
issue once we stop generating canonical IVs.
llvm-svn: 130340
effective in avoiding recomputation of LCSSA form; the widespread
use of instsimplify (which looks through phi nodes) means it was
not preserving LCSSA form anyway; and instcombine is no longer
scheduled in the middle of the loop passes so this doesn't matter
anymore.
llvm-svn: 130301
when X has multiple uses. This is useful for exposing secondary optimizations,
but the X86 backend isn't ready for this when X has a single use. For example,
this can disable load folding.
This is inching towards resolving PR6627.
llvm-svn: 130238
Add support for switch and indirectbr edges. This works by densely numbering
all blocks which have such terminators, and then separately numbering the
possible successors. The predecessors write down a number, the successor knows
its own number (as a ConstantInt) and sends that and the pointer to the number
the predecessor wrote down to the runtime, who looks up the counter in a
per-function table.
Coverage data should now be functional, but I haven't tested it on anything
other than my 2-file synthetic test program for coverage.
llvm-svn: 130186
return it as a clobber. This allows GVN to do smart things.
Enhance GVN to be smart about the case when a small load is clobbered
by a larger overlapping load. In this case, forward the value. This
allows us to compile stuff like this:
int test(void *P) {
int tmp = *(unsigned int*)P;
return tmp+*((unsigned char*)P+1);
}
into:
_test: ## @test
movl (%rdi), %ecx
movzbl %ch, %eax
addl %ecx, %eax
ret
which has one load. We already handled the case where the smaller
load was from a must-aliased base pointer.
llvm-svn: 130180
necessary since gcov counts transitions between blocks. It can't see if you've
run every line in a straight-line function, so we add an edge for it to notice.
llvm-svn: 129905
Break the arc-profile code out to a function like the notes emission code is,
and reorder the functions in the file.
The only functionality change is that we no longer modify the Module when the
Module has no debug info to use.
llvm-svn: 129631
instruction around, reducing work.
Greatly simplify handling of debug instructions. There is no need to
build up a vector of them and then move them into the one predecessor
if we're processing a block. Instead just rescan the block and *copy*
them into the pred. If a block gets merged into multiple preds, this
will retain more debug info.
llvm-svn: 129502
the same allocation size but different primitive sizes(e.g., <3xi32> and
<4xi32>). When ScalarRepl promotes them, it can't use a bit cast but
should use a shuffle vector instead.
llvm-svn: 129472
will allow multiple context with different loop unroll parameters to run. This is a minor change and no effect
on existing application.
llvm-svn: 129449
Now that we have a first-class way to represent unaligned loads, the unaligned
load intrinsics are superfluous.
First part of <rdar://problem/8460511>.
llvm-svn: 129401
Use debug info in the IR to find the directory/file:line:col. Each time that location changes, bump a counter.
Unlike the existing profiling system, we don't try to look at argv[], and thusly don't require main() to be present in the IR. This matches GCC's technique where you specify the profiling flag when producing each .o file.
The runtime library is minimal, currently just calling printf at program shutdown time. The API is designed to make it possible to emit GCOV data later on.
llvm-svn: 129340
reassociation opportunities are exposed. This fixes a bug where
the nested reassociation expects to be the IR to be consistent,
but it isn't, because the outer reassociation has disconnected
some of the operands. rdar://9167457
llvm-svn: 129324
mean that it has to be ConstantArray of ConstantStruct. We might have
ConstantAggregateZero, at either level, so don't crash on that.
Also, semi-deprecate the sentinal value. The linker isn't aware of sentinals so
we end up with the two lists appended, each with their "sentinals" on them.
Different parts of LLVM treated sentinals differently, so make them all just
ignore the single entry and continue on with the rest of the list.
llvm-svn: 129307
is equivalent to any other relevant value; it isn't true in general.
If it is equivalent, the LoopPromoter will tell the AST the equivalence.
Also, delete the PreheaderLoad if it is unused.
Chris, since you were the last one to make major changes here, can you check
that this is sane?
llvm-svn: 129049
space info. We crash with an assert in this case. This change checks that the
address space of the bitcasted pointer is the same as the gep ptr.
llvm-svn: 128884
after the given instruction; make sure to handle that case correctly.
(It's difficult to trigger; the included testcase involves a dead
block, but I don't think that's a requirement.)
While I'm here, get rid of the unnecessary warning about
SimplifyInstructionsInBlock, since it should work correctly as far as I know.
llvm-svn: 128782
It's possible to craft an input that hits the recursion limits in a way
that SimplifyDemandedBits doesn't simplify the icmp but ComputeMaskedBits
can infer which bits are zero.
No test case as it depends on too many other things. Fixes PR9609.
llvm-svn: 128777
- Localize the check if an icmp has one use to a place where we know we're
introducing something that's likely more expensive than a sext from i1.
- Add an assert to make sure a case that would lead to a miscompilation is
folded away earlier.
- Fix a typo.
llvm-svn: 128744
that one of the numbers is signed while the other is unsigned. This could lead
to a wrong result when the signed was promoted to an unsigned int.
* Add the data layout line to the testcase so that it will test the appropriate
thing.
Patch by David Terei!
llvm-svn: 128577
removes one use of X which helps it pass the many hasOneUse() checks.
In my analysis, this turns up very often where X = A >>exact B and that can't be
simplified unless X has one use (except by increasing the lifetime of A which is
generally a performance loss).
llvm-svn: 128373
There are two ways that a later store can comletely overlap a previous store:
1. They both start at the same offset, but the earlier store's size is <= the
later's size, or
2. The earlier store's offset is > the later's offset, but it's offset + size
doesn't extend past the later's offset + size.
llvm-svn: 128332
to have single return block (at least getting there) for optimizations. This
is general goodness but it would prevent some tailcall optimizations.
One specific case is code like this:
int f1(void);
int f2(void);
int f3(void);
int f4(void);
int f5(void);
int f6(void);
int foo(int x) {
switch(x) {
case 1: return f1();
case 2: return f2();
case 3: return f3();
case 4: return f4();
case 5: return f5();
case 6: return f6();
}
}
=>
LBB0_2: ## %sw.bb
callq _f1
popq %rbp
ret
LBB0_3: ## %sw.bb1
callq _f2
popq %rbp
ret
LBB0_4: ## %sw.bb3
callq _f3
popq %rbp
ret
This patch teaches codegenprep to duplicate returns when the return value
is a phi and where the phi operands are produced by tail calls followed by
an unconditional branch:
sw.bb7: ; preds = %entry
%call8 = tail call i32 @f5() nounwind
br label %return
sw.bb9: ; preds = %entry
%call10 = tail call i32 @f6() nounwind
br label %return
return:
%retval.0 = phi i32 [ %call10, %sw.bb9 ], [ %call8, %sw.bb7 ], ... [ 0, %entry ]
ret i32 %retval.0
This allows codegen to generate better code like this:
LBB0_2: ## %sw.bb
jmp _f1 ## TAILCALL
LBB0_3: ## %sw.bb1
jmp _f2 ## TAILCALL
LBB0_4: ## %sw.bb3
jmp _f3 ## TAILCALL
rdar://9147433
llvm-svn: 127953
SCEV may generate expressions composed of multiple pointers, which can
lead to invalid GEP expansion. Until we can teach SCEV to follow strict
pointer rules, make sure no bad GEPs creep into IR.
Fixes rdar://problem/9038671.
llvm-svn: 127839
chose is having a non-memcpy/memset use and being larger than any native integer
type. Originally I chose having an access of a size smaller than the total size
of the alloca, but this caused some minor issues on the spirit benchmark where
SRoA runs again after some inlining.
This fixes <rdar://problem/8613163>.
llvm-svn: 127718
properties.
Added the self-wrap flag for SCEV::AddRecExpr.
A slew of temporary FIXMEs indicate the intention of the no-self-wrap flag
without changing behavior in this revision.
llvm-svn: 127590
load and store reference same memory location, the memory location
is represented by getelementptr with two uses (load and store) and
the getelementptr's base is alloca with single use. At this point,
instructions from alloca to store can be removed.
(this pattern is generated when bitfield is accessed.)
For example,
%u = alloca %struct.test, align 4 ; [#uses=1]
%0 = getelementptr inbounds %struct.test* %u, i32 0, i32 0;[#uses=2]
%1 = load i8* %0, align 4 ; [#uses=1]
%2 = and i8 %1, -16 ; [#uses=1]
%3 = or i8 %2, 5 ; [#uses=1]
store i8 %3, i8* %0, align 4
llvm-svn: 127565
Optimize trivial branches in CodeGenPrepare, which often get created from the
lowering of objectsize intrinsics. Unfortunately, a number of tests were relying
on llc not optimizing trivial branches, so I had to add an option to allow them
to continue to test what they originally tested.
This fixes <rdar://problem/8785296> and <rdar://problem/9112893>.
llvm-svn: 127498
lowering of objectsize intrinsics. Unfortunately, a number of tests were relying
on llc not optimizing trivial branches, so I had to add an option to allow them
to continue to test what they originally tested.
This fixes <rdar://problem/8785296> and <rdar://problem/9112893>.
llvm-svn: 127459
Value, not an Instruction, so casting is not necessary. Also,
it's theoretically possible that the Value is not an
Instruction, since WeakVH follows RAUWs.
llvm-svn: 127427
after it has finished all of its reassociations, because its
habit of unlinking operands and holding them in a datastructure
while working means that it's not easy to determine when an
instruction is really dead until after all its regular work is
done. rdar://9096268.
llvm-svn: 127424
This happens a lot in clang-compiled C++ code because it adds overflow checks to operator new[]:
unsigned *foo(unsigned n) { return new unsigned[n]; }
We can optimize away the overflow check on 64 bit targets because (uint64_t)n*4 cannot overflow.
llvm-svn: 127418
alloca as both integer and floating-point vectors of the same size. Bugpoint is
not cooperating with me, but I'll try to find a manual testcase tomorrow.
llvm-svn: 127320
a union of a float, <2 x float>, and <4 x float>. This mostly comes up with the
use of vector intrinsics, especially in NEON when programmers know the layout of
the register file. This enables codegen to eliminate a lot of the subregister
traffic it would otherwise generate.
This commit only enables this for a small number of floating-point cases, but a
lot more integer cases. I assume this is okay for all ports, but I did not do
extensive testing of the quality of code involving i512 vectors and the like. If
there is a use case where this generates worse code than before, let me know and
we can scale it back.
This fixes <rdar://problem/9036264>.
llvm-svn: 127317
reachable uses, but there still might be uses in dead blocks. Use the
standard solution of replacing all the uses with undef. This is
a rare case because it's very sensitive to phase ordering in SimplifyCFG.
llvm-svn: 127299
the value splatted into every element. Extend this to getTrue and getFalse which
by providing new overloads that take Types that are either i1 or <N x i1>. Use
it in InstCombine to add vector support to some code, fixing PR8469!
llvm-svn: 127116
possible. This goes into instcombine and instsimplify because instsimplify
doesn't need to check hasOneUse since it returns (almost exclusively) constants.
This fixes PR9343 #4#5 and #8!
llvm-svn: 127064
addressing code. On 403.gcc this almost halves CodeGenPrepare time and reduces
total llc time by 9.5%. Unfortunately, getNumUses() is still the hottest function
in llc.
llvm-svn: 126782
intersection of the LHS and RHS ConstantRanges and return "false" when
the range is empty.
This simplifies some code and catches some extra cases.
llvm-svn: 126744
Yes, there are other types than i8* and GEPs on them can produce an add+multiply.
We don't consider that cheap enough to be speculatively executed.
llvm-svn: 126481
function prototype into a call to a varargs prototype. We do
allow the xform if we have a definition, but otherwise we don't
want to risk that we're changing the abi in a subtle way. On
X86-64, for example, varargs require passing stuff in %al.
llvm-svn: 126363
itself without going via a phi node then we could return false here in
spite of making a change. Also, tweak the comment because this method
can (and always could) return true without deleting the original phi node.
For example, if the phi node was used by a read-only invoke instruction
which is used by another phi node phi2 which is only used by and only uses
the invoke, then phi2 would be deleted but not the invoke instruction and
not the original phi node.
llvm-svn: 126129
should be that if the phi is used by a side-effect free instruction with
no uses then the phi and the instruction now get zapped (checked by the
unittest).
llvm-svn: 126124
test for that. With this change, test/CodeGen/X86/codegen-dce.ll no longer finds
any instructions to DCE, so delete the test.
Also renamed J and JP to I and IP in RecursivelyDeleteDeadPHINode.
llvm-svn: 126088
We usually catch this kind of optimization through InstSimplify's distributive
magic, but or doesn't distribute over xor in general.
"A | ~(A | B) -> A | ~B" hits 24 times on gcc.c.
llvm-svn: 126081
one Value set. This is faster because we only need to use the set when there
isn't already an entry in the map. No functionality change!
llvm-svn: 126076
constant, including globals. This makes us generate much more "pretty" pattern
globals as well because it doesn't break it down to an array of bytes all the
time.
This enables us to handle stores of relocatable globals. This kicks in about
48 times in 254.gap, giving us stuff like this:
@.memset_pattern40 = internal constant [2 x %struct.TypHeader* (%struct.TypHeader*, %struct.TypHeader*)*] [%struct.TypHeader* (%struct.TypHeader*, %struct
.TypHeader*)* @IsFalse, %struct.TypHeader* (%struct.TypHeader*, %struct.TypHeader*)* @IsFalse], align 16
...
call void @memset_pattern16(i8* %scevgep5859, i8* bitcast ([2 x %struct.TypHeader* (%struct.TypHeader*, %struct.TypHeader*)*]* @.memset_pattern40 to i8*
), i64 %tmp75) nounwind
llvm-svn: 126044
unsplatable values into memset_pattern16 when it is available
(recent darwins). This transforms lots of strided loop stores
of ints for example, like 5 in vpr:
Formed memset: call void @memset_pattern16(i8* %4, i8* getelementptr inbounds ([16 x i8]* @.memset_pattern9, i32 0, i32 0), i64 %tmp25)
from store to: {%3,+,4}<%11> at: store i32 3, i32* %scevgep, align 4, !tbaa !4
llvm-svn: 126040
taken (and used!). This prevents merging the blocks (invalidating
the block addresses) in a case like this:
#define _THIS_IP_ ({ __label__ __here; __here: (unsigned long)&&__here; })
void foo() {
printf("%p\n", _THIS_IP_);
printf("%p\n", _THIS_IP_);
printf("%p\n", _THIS_IP_);
}
which fixes PR4151.
llvm-svn: 125829
This is part of a futile attempt to not "break" bizzaro
code like this:
l1:
printf("l1: %p\n", &&l1);
++x;
if( x < 3 ) goto l1;
Previously we'd fold &&l1 to 1, which is fine per our semantics
but not helpful to the user.
llvm-svn: 125827
gep to explicit addressing, we know that none of the intermediate
computation overflows.
This could use review: it seems that the shifts certainly wouldn't
overflow, but could the intermediate adds overflow if there is a
negative index?
Previously the testcase would instcombine to:
define i1 @test(i64 %i) {
%p1.idx.mask = and i64 %i, 4611686018427387903
%cmp = icmp eq i64 %p1.idx.mask, 1000
ret i1 %cmp
}
now we get:
define i1 @test(i64 %i) {
%cmp = icmp eq i64 %i, 1000
ret i1 %cmp
}
llvm-svn: 125271
exact/nsw/nuw shifts and have instcombine infer them when it can prove
that the relevant properties are true for a given shift without them.
Also, a variety of refactoring to use the new patternmatch logic thrown
in for good luck. I believe that this takes care of a bunch of related
code quality issues attached to PR8862.
llvm-svn: 125267
optimizations to be much more aggressive in the face of
exact/nsw/nuw div and shifts. For example, these (which
are the same except the first is 'exact' sdiv:
define i1 @sdiv_icmp4_exact(i64 %X) nounwind {
%A = sdiv exact i64 %X, -5 ; X/-5 == 0 --> x == 0
%B = icmp eq i64 %A, 0
ret i1 %B
}
define i1 @sdiv_icmp4(i64 %X) nounwind {
%A = sdiv i64 %X, -5 ; X/-5 == 0 --> x == 0
%B = icmp eq i64 %A, 0
ret i1 %B
}
compile down to:
define i1 @sdiv_icmp4_exact(i64 %X) nounwind {
%1 = icmp eq i64 %X, 0
ret i1 %1
}
define i1 @sdiv_icmp4(i64 %X) nounwind {
%X.off = add i64 %X, 4
%1 = icmp ult i64 %X.off, 9
ret i1 %1
}
This happens when you do something like:
(ptr1-ptr2) == 42
where the pointers are pointers to non-unit types.
llvm-svn: 125266
and generally tidying things up. Only very trivial functionality changes
like now doing (-1 - A) -> (~A) for vectors too.
InstCombineAddSub.cpp | 296 +++++++++++++++++++++-----------------------------
1 file changed, 126 insertions(+), 170 deletions(-)
llvm-svn: 125264
Natural Loop Information
Loop Pass Manager
Canonicalize natural loops
Scalar Evolution Analysis
Loop Pass Manager
Induction Variable Users
Canonicalize natural loops
Induction Variable Users
Loop Strength Reduction
into this:
Scalar Evolution Analysis
Loop Pass Manager
Canonicalize natural loops
Induction Variable Users
Loop Strength Reduction
This fixes <rdar://problem/8869639>. I also filed PR9184 on doing this sort of
thing automatically, but it seems easier to just change the ordering of the
passes if this is the only case.
llvm-svn: 125254
versions of creation functions. Eventually, the "insertion point" versions
of these should just be removed, we do have IRBuilder afterall.
Do a massive rewrite of much of pattern match. It is now shorter and less
redundant and has several other widgets I will be using in other patches.
Among other changes, m_Div is renamed to m_IDiv (since it only matches
integer divides) and m_Shift is gone (it used to match all binops!!) and
we now have m_LogicalShift for the one client to use.
Enhance IRBuilder to have "isExact" arguments to things like CreateUDiv
and reduce redundancy within IRbuilder by having these methods chain to
each other more instead of duplicating code.
llvm-svn: 125194
could end up removing a different function than we intended because it was
functionally equivalent, then end up with a comparison of a function against
itself in the next round of comparisons (the one in the function set and the
one on the deferred list). To fix this, I introduce a choice in the form of
comparison for ComparableFunctions, either normal or "pointer only" used to
find exact Function*'s in lookups.
Also add some debugging statements.
llvm-svn: 125180
the active loop. This is generally desirable, and it avoids trouble
in situations such as the testcase in PR9123, though the failure
mode depends on use-list order, so it is infeasible to test.
llvm-svn: 125065
This makes the job of the later optzn passes easier, allowing the vast amount of
icmp transforms to chew on it.
We transform 840 switches in gcc.c, leading to a 16k byte shrink of the resulting
binary on i386-linux.
The testcase from README.txt now compiles into
decl %edi
cmpl $3, %edi
sbbl %eax, %eax
andl $1, %eax
ret
llvm-svn: 124724
that might have changed been affected by a merge elsewhere will have been
removed from the function set, and it isn't needed for performance because we
call grow() ahead of time to prevent reallocations.
llvm-svn: 124717
Modified patch by Adam Preuss.
This builds on the existing framework for block tracing, edge profiling and optimal edge profiling.
See -help-hidden for new flags.
For documentation, see the technical report "Implementation of Path Profiling..." in llvm.org/pubs.
llvm-svn: 124515
benchmarks, and that it can be simplified to X/Y. (In general you can only
simplify (Z*Y)/Y to Z if the multiplication did not overflow; if Z has the
form "X/Y" then this is the case). This patch implements that transform and
moves some Div logic out of instcombine and into InstructionSimplify.
Unfortunately instcombine gets in the way somewhat, since it likes to change
(X/Y)*Y into X-(X rem Y), so I had to teach instcombine about this too.
Finally, thanks to the NSW/NUW flags, sometimes we know directly that "Z*Y"
does not overflow, because the flag says so, so I added that logic too. This
eliminates a bunch of divisions and subtractions in 447.dealII, and has good
effects on some other benchmarks too. It seems to have quite an effect on
tramp3d-v4 but it's hard to say if it's good or bad because inlining decisions
changed, resulting in massive changes all over.
llvm-svn: 124487
operand being factorized (and erased) could occur several times in Ops,
resulting in freed memory being used when the next occurrence in Ops was
analyzed.
llvm-svn: 124287
merge vector<intptr_t>::push_back() and vector<void*>::push_back() because
Enumerate() doesn't realize that "i64* null" and "i8** null" are equivalent.
llvm-svn: 124285
with BasicAA's DecomposeGEPExpression, which recently began
using a TargetData. This fixes PR8968, though the testcase
is awkward to reduce.
Also, update several off GetUnderlyingObject's users
which happen to have a TargetData handy to pass it in.
llvm-svn: 124134
occurs because instcombine sinks loads and inserts phis. This kicks in
on such apps as 175.vpr, eon, 403.gcc, xalancbmk and a bunch of times in
spec2006 in some app that uses std::deque.
This resolves the last of rdar://7339113.
llvm-svn: 124090
common cases. This triggers a surprising number of times in SPEC2K6
because min/max idioms end up doing this. For example, code from the
STL ends up looking like this to SRoA:
%202 = load i64* %__old_size, align 8, !tbaa !3
%203 = load i64* %__old_size, align 8, !tbaa !3
%204 = load i64* %__n, align 8, !tbaa !3
%205 = icmp ult i64 %203, %204
%storemerge.i = select i1 %205, i64* %__n, i64* %__old_size
%206 = load i64* %storemerge.i, align 8, !tbaa !3
We can now promote both the __n and the __old_size allocas.
This addresses another chunk of rdar://7339113, poor codegen on
stringswitch.
llvm-svn: 124088
clang's -Wuninitialized-experimental warning.
While these don't look like real bugs, clang's
-Wuninitialized-experimental analysis is stricter
than GCC's, and these fixes have the benefit
of being general nice cleanups.
llvm-svn: 124073
that have PHI or select uses of their element pointers. This can often happen
when instcombine sinks two loads into a successor, inserting a phi or select.
With this patch, we can scalarize the alloca, but the pinned elements are not
yet promoted. This is still a win for large aggregates where only one element
is used. This fixes rdar://8904039 and part of rdar://7339113 (poor codegen
on stringswitch).
llvm-svn: 124070
handle the "Transformation preventing inst" printing,
so that -scalarrepl -debug will always print the rejected
instruction. No functionality change.
llvm-svn: 124066
a select. A vector select is pairwise on each element so we'd need a new
condition with the right number of elements to select on. Fixes PR8994.
llvm-svn: 123963
auto-simplier the transform most missed by early-cse is (zext X) != 0 -> X != 0.
This patch adds this transform and some related logic to InstructionSimplify
and removes some of the logic from instcombine (unfortunately not all because
there are several situations in which instcombine can improve things by making
new instructions, whereas instsimplify is not allowed to do this). At -O2 this
often results in more than 15% more simplifications by early-cse, and results in
hundreds of lines of bitcode being eliminated from the testsuite. I did see some
small negative effects in the testsuite, for example a few additional instructions
in three programs. One program, 483.xalancbmk, got an additional 35 instructions,
which seems to be due to a function getting an additional instruction and then
being inlined all over the place.
llvm-svn: 123911
without whatever this was trying to do. When/if someone has the time to do some empirical
evaluations, it might be worth it to figure out what this code was trying to do and see if
it's worth resurrecting/fixing.
llvm-svn: 123684
checks enabled:
1) Use '<' to compare integers in a comparison function rather than '<='.
2) Use the uniqued set DefBlocks rather than Info.DefiningBlocks to initialize
the priority queue.
The speedup of scalarrepl on test-suite + SPEC2000 + SPEC2006 is a bit less, at
just under 16% rather than 17%.
llvm-svn: 123662
eliminating a potentially quadratic data structure, this also gives a 17%
speedup when running -scalarrepl on test-suite + SPEC2000 + SPEC2006. My initial
experiment gave a greater speedup around 25%, but I moved the dominator tree
level computation from dominator tree construction to PromoteMemToReg.
Since this approach to computing IDFs has a much lower overhead than the old
code using precomputed DFs, it is worth looking at using this new code for the
second scalarrepl pass as well.
llvm-svn: 123609
This fixes the original testcase in PR8927. It also causes a clang
binary built with a patched clang to increase in size by 0.21%.
We can probably get some of the size back by writing a pass that
detects that a global never has its pointer compared and adds
unnamed_addr to it (maybe extend global opt). It is also possible that
there are some other cases clang could add unnamed_addr to.
I will investigate extending globalopt next.
llvm-svn: 123584
then don't try to decimate it into its individual pieces. This will just make a mess of the
IR and is pointless if none of the elements are individually accessed. This was generating
really terrible code for std::bitset (PR8980) because it happens to be lowered by clang
as an {[8 x i8]} structure instead of {i64}.
The testcase now is optimized to:
define i64 @test2(i64 %X) {
br label %L2
L2: ; preds = %0
ret i64 %X
}
before we generated:
define i64 @test2(i64 %X) {
%sroa.store.elt = lshr i64 %X, 56
%1 = trunc i64 %sroa.store.elt to i8
%sroa.store.elt8 = lshr i64 %X, 48
%2 = trunc i64 %sroa.store.elt8 to i8
%sroa.store.elt9 = lshr i64 %X, 40
%3 = trunc i64 %sroa.store.elt9 to i8
%sroa.store.elt10 = lshr i64 %X, 32
%4 = trunc i64 %sroa.store.elt10 to i8
%sroa.store.elt11 = lshr i64 %X, 24
%5 = trunc i64 %sroa.store.elt11 to i8
%sroa.store.elt12 = lshr i64 %X, 16
%6 = trunc i64 %sroa.store.elt12 to i8
%sroa.store.elt13 = lshr i64 %X, 8
%7 = trunc i64 %sroa.store.elt13 to i8
%8 = trunc i64 %X to i8
br label %L2
L2: ; preds = %0
%9 = zext i8 %1 to i64
%10 = shl i64 %9, 56
%11 = zext i8 %2 to i64
%12 = shl i64 %11, 48
%13 = or i64 %12, %10
%14 = zext i8 %3 to i64
%15 = shl i64 %14, 40
%16 = or i64 %15, %13
%17 = zext i8 %4 to i64
%18 = shl i64 %17, 32
%19 = or i64 %18, %16
%20 = zext i8 %5 to i64
%21 = shl i64 %20, 24
%22 = or i64 %21, %19
%23 = zext i8 %6 to i64
%24 = shl i64 %23, 16
%25 = or i64 %24, %22
%26 = zext i8 %7 to i64
%27 = shl i64 %26, 8
%28 = or i64 %27, %25
%29 = zext i8 %8 to i64
%30 = or i64 %29, %28
ret i64 %30
}
In this case, instcombine was able to eliminate the nonsense, but in PR8980 enough
PHIs are in play that instcombine backs off. It's better to not generate this stuff
in the first place.
llvm-svn: 123571
multiple uses. In some cases, all the uses are the same operation,
so instcombine can go ahead and promote the phi. In the testcase
this pushes an add out of the loop.
llvm-svn: 123568
The basic issue is that isel (very reasonably!) expects conditional branches
to be folded, so CGP leaving around a bunch dead computation feeding
conditional branches isn't such a good idea. Just fold branches on constants
into unconditional branches.
llvm-svn: 123526
have objectsize folding recursively simplify away their result when it
folds. It is important to catch this here, because otherwise we won't
eliminate the cross-block values at isel and other times.
llvm-svn: 123524
instead of DomTree/DomFrontier. This may be interesting for reducing compile
time. This is currently disabled, but seems to work just fine.
When this is enabled, we eliminate two runs of dominator frontier, one in the
"early per-function" optimizations and one in the "interlaced with inliner"
function passes.
llvm-svn: 123434
While there, I noticed that the transform "undef >>a X -> undef" was wrong.
For example if X is 2 then the top two bits must be equal, so the result can
not be anything. I fixed this in the constant folder as well. Also, I made
the transform for "X << undef" stronger: it now folds to undef always, even
though X might be zero. This is in accordance with the LangRef, but I must
admit that it is fairly aggressive. Also, I added "i32 X << 32 -> undef"
following the LangRef and the constant folder, likewise fairly aggressive.
llvm-svn: 123417
This is a minor extension of SROA to handle a special case that is
important for some ARM NEON operations. Some of the NEON intrinsics
return multiple values, which are handled as struct types containing
multiple elements of the same vector type. The corresponding return
types declared in the arm_neon.h header have equivalent arrays. We
need SROA to recognize that it can split up those arrays and structs
into separate vectors, even though they are not always accessed with
the same type. SROA already handles loads and stores of an entire
alloca by using insertvalue/extractvalue to access the individual
pieces, and that code works the same regardless of whether the type
is a struct or an array. So, all that needs to be done is to check
for compatible arrays and homogeneous structs.
llvm-svn: 123381
SROA only split up structs and arrays one level at a time, so padding can
only cause trouble if it is located in between the struct or array elements.
llvm-svn: 123380
DT->changeImmediateDominator() trivially ignores identity updates, so there is
really no need for the uniqueing provided by SmallPtrSet.
I expect this to fix PR8954.
llvm-svn: 123286
phi nodes. It is called from MergeBlockIntoPredecessor which is
called from GVN, which claims to preserve these.
I'm skeptical that this is the actual problem behind PR8954, but
this is a stab in the right direction.
llvm-svn: 123222
without informing memdep. This could cause nondeterminstic weirdness
based on where instructions happen to get allocated, and will hopefully
breath some life into some broken testers.
llvm-svn: 123124
larger memsets. Among other things, this fixes rdar://8760394 and
allows us to handle "Example 2" from http://blog.regehr.org/archives/320,
compiling it into a single 4096-byte memset:
_mad_synth_mute: ## @mad_synth_mute
## BB#0: ## %entry
pushq %rax
movl $4096, %esi ## imm = 0x1000
callq ___bzero
popq %rax
ret
llvm-svn: 123089
that it was leaving in loops after rotation (between the original latch
block and the original header.
With this change, it is possible for rotated loops to have just a single
basic block, which is useful.
llvm-svn: 123075
1. Rip out LoopRotate's domfrontier updating code. It isn't
needed now that LICM doesn't use DF and it is super complex
and gross.
2. Make DomTree updating code a lot simpler and faster. The
old loop over all the blocks was just to find a block??
3. Change the code that inserts the new preheader to just use
SplitCriticalEdge instead of doing an overcomplex
reimplementation of it.
No behavior change, except for the name of the inserted preheader.
llvm-svn: 123072
they all ready do). This removes two dominator recomputations prior to isel,
which is a 1% improvement in total llc time for 403.gcc.
The only potentially suspect thing is making GCStrategy recompute dominators if
it used a custom lowering strategy.
llvm-svn: 123064
them into the loop preheader, eliminating silly instructions like
"icmp i32 0, 100" in fixed tripcount loops. This also better exposes the
bigger problem with loop rotate that I'd like to fix: once this has been
folded, the duplicated conditional branch *often* turns into an uncond branch.
Not aggressively handling this is pessimizing later loop optimizations
somethin' fierce by making "dominates all exit blocks" checks fail.
llvm-svn: 123060
1. Take a flags argument instead of a bool. This makes
it more clear to the reader what it is used for.
2. Add a flag that says that "remapping a value not in the
map is ok".
3. Reimplement MapValue to share a bunch of code and be a lot
more efficient. For lookup failures, don't drop null values
into the map.
4. Using the new flag a bunch of code can vaporize in LinkModules
and LoopUnswitch, kill it.
No functionality change.
llvm-svn: 123058
map from ValueMapper.h (giving us access to its utilities)
and add a fastpath in the loop rotation code, avoiding expensive
ssa updator manipulation for values with nothing to update.
llvm-svn: 123057
X = sext x; x >s c ? X : C+1 --> X = sext x; X <s C+1 ? C+1 : X
X = sext x; x <s c ? X : C-1 --> X = sext x; X >s C-1 ? C-1 : X
X = zext x; x >u c ? X : C+1 --> X = zext x; X <u C+1 ? C+1 : X
X = zext x; x <u c ? X : C-1 --> X = zext x; X >u C-1 ? C-1 : X
X = sext x; x >u c ? X : C+1 --> X = sext x; X <u C+1 ? C+1 : X
X = sext x; x <u c ? X : C-1 --> X = sext x; X >u C-1 ? C-1 : X
Instead of calculating this with mixed types promote all to the
larger type. This enables scalar evolution to analyze this
expression. PR8866
llvm-svn: 123034
skipping them, but it should probably use a worklist and only revisit those
instructions in subloops that have actually changed. It should probably also
use a worklist after the first iteration like instsimplify now does. Regardless,
it's only 0.3% of opt -O2 time on 403.gcc if it replaces the instcombine placed
in the middle of the loop passes.
llvm-svn: 122868
case where a static caller is itself inlined everywhere else, and
thus may go away if it doesn't get too big due to inlining other
things into it. If there are references to the caller other than
calls, it will not be removed; account for this.
This results in same-day completion of the case in PR8853.
llvm-svn: 122821
when safe.
The testcase is basically this nested loop:
void foo(char *X) {
for (int i = 0; i != 100; ++i)
for (int j = 0; j != 100; ++j)
X[j+i*100] = 0;
}
which gets turned into a single memset now. clang -O3 doesn't optimize
this yet though due to a phase ordering issue I haven't analyzed yet.
llvm-svn: 122806
instruction *after* the store. The store will always be deleted
if the transformation kicks in, so we'd do an N^2 scan of every
loop block. Whoops.
llvm-svn: 122805
FunctionPass. It probably doesn't have a reason to be a LoopPass, as it will
probably drop the simple fixed point and either use RPO iteration or Duncan's
approach in instsimplify of only revisiting instructions that have changed.
The next step is to preserve LoopSimplify. This looks like it won't be too hard,
although the pass manager doesn't actually seem to respect when non-loop passes
claim to preserve LCSSA or LoopSimplify. This will have to be fixed.
llvm-svn: 122791
that are allowed to have metadata operands are intrinsic calls,
and the only ones that take metadata currently return void.
Just reject all void instructions, which should not be value
numbered anyway. To future proof things, add an assert to the
getHashValue impl for calls to check that metadata operands
aren't present.
llvm-svn: 122759
nested values, so they can change and drop to null, which can
change the hash and cause havok.
It turns out that it isn't a good idea to value number stuff
with metadata operands anyway, so... don't.
llvm-svn: 122758
capacity on the Visited SmallPtrSet. On 403.gcc, this is about a 4.5% speedup of
CodeGenPrepare time (which itself is 10% of time spent in the backend).
This is progress towards PR8889.
llvm-svn: 122741
of instcombine that is currently in the middle of the loop pass pipeline. This
commit only checks in the pass; it will hopefully be enabled by default later.
llvm-svn: 122719
sure that the loop we're promoting into a memcpy doesn't mutate the input
of the memcpy. Before we were just checking that the dest of the memcpy
wasn't mod/ref'd by the loop.
llvm-svn: 122712
isExitBlockDominatedByBlockInLoop is a relic of the days when domtree was
*just* a tree and didn't have DFS numbers. Checking DFS numbers is faster
and easier than "limiting the search of the tree".
llvm-svn: 122702
in the PR, the pass could break LCSSA form when inserting preheaders. It probably
would be easy enough to fix this, but since currently we always go into LCSSA form
after running this pass, doing so is not urgent.
llvm-svn: 122695
header for now for memset/memcpy opportunities. It turns out that loop-rotate
is successfully rotating loops, but *DOESN'T MERGE THE BLOCKS*, turning "for
loops" into 2 basic block loops that loop-idiom was ignoring.
With this fix, we form many *many* more memcpy and memsets than before, including
on the "history" loops in the viterbi benchmark, which look like this:
for (j=0; j<MAX_history; ++j) {
history_new[i][j+1] = history[2*i][j];
}
Transforming these loops into memcpy's speeds up the viterbi benchmark from
11.98s to 3.55s on my machine. Woo.
llvm-svn: 122685
maintains the guarantee that the DenseSet expects two elements it contains to
not go from inequal to equal under its nose.
As a side-effect, this also lets us switch from iterating to a fixed-point to
actually maintaining a work queue of functions to look at again, and we don't
add thunks to our work queue so we don't need to detect and ignore them.
llvm-svn: 122677
pipeline to be caught by instcombine, and it's not feasible to catch them in SimplifyCFG because the
use-lists are in an inconsistent state at the point where it could know that it need to simplify them.
Instead, have CodeGenPrepare look for trivially redundant PHIs as part of its general cleanup effort.
llvm-svn: 122516
if both A op B and A op C simplify. This fires fairly often but doesn't
make that much difference. On gcc-as-one-file it removes two "and"s and
turns one branch into a select.
llvm-svn: 122399
I still think that LVI should be handling this, but that capability is some ways off in the future,
and this matters for some significant benchmarks.
llvm-svn: 122378
visit instructions before their uses, since InstructionSimplify does a
better job in that case. All this prompted by Frits van Bommel.
llvm-svn: 122343
it could only be tested indirectly, via instcombine, gvn or some other
pass that makes use of InstructionSimplify, which means that testcases
had to be carefully contrived to dance around any other transformations
that that pass did.
llvm-svn: 122264
argument. The generated alloca has to have at least the alignment of the
byval, if not, the client may be making assumptions that the new alloca won't
satisfy.
llvm-svn: 122234
This resolves a README entry and technically resolves PR4916,
but we still get poor code for the testcase in that PR because
GVN isn't CSE'ing uadd with add, filed as PR8817.
Previously we got:
_test7: ## @test7
addq %rsi, %rdi
cmpq %rdi, %rsi
movl $42, %eax
cmovaq %rsi, %rax
ret
Now we get:
_test7: ## @test7
addq %rsi, %rdi
movl $42, %eax
cmovbq %rsi, %rax
ret
llvm-svn: 122182
the old thing end up on the instcombine worklist. Not doing this
can cause an extra top-level iteration of instcombine, burning
compile time.
llvm-svn: 122179
sadd formed is half the size of the original type. We can
now compile this into a sadd.i8:
unsigned char X(char a, char b) {
int res = a+b;
if ((unsigned )(res+128) > 255U)
abort();
return res;
}
llvm-svn: 122178
checking to see if the high bits of the original add result were dead.
Inserting a smaller add and zexting back to that size is not good enough.
This is likely to be the fix for 8816.
llvm-svn: 122177
which have trapping constant exprs in them due to PHI nodes.
Eliminating them can cause the constant expr to be evalutated
on new paths if the input edges are critical.
llvm-svn: 122164
on the DragonEgg self-host bot. Unfortunately, the testcase is pretty messy and doesn't reduce well due to
interactions with other parts of InstCombine.
llvm-svn: 122072
a null endptr argument, because they may write to errno.
This fixes a seflhost miscompile observed on Linux targets when TBAA
was enabled.
llvm-svn: 122014
dragonegg self-host buildbot. Original commit message:
Add an InstCombine transform to recognize instances of manual overflow-safe addition
(performing the addition in a wider type and explicitly checking for overflow), and
fold them down to intrinsics. This currently only supports signed-addition, but could
be generalized if someone works out the magic constant formulas for other operations.
llvm-svn: 121965
(performing the addition in a wider type and explicitly checking for overflow), and
fold them down to intrinsics. This currently only supports signed-addition, but could
be generalized if someone works out the magic constant formulas for other operations.
Fixes <rdar://problem/8558713>.
llvm-svn: 121905
When it sees a promising select it now tries to figure out whether the condition of the select is known in any of the predecessors and if so it maps the operands appropriately.
llvm-svn: 121859
which is simpler than finding a place to insert in BB.
- Don't perform the 'if condition hoisting' xform on certain
i1 PHIs, as it interferes with switch formation.
This re-fixes "example 7", without breaking the world hopefully.
llvm-svn: 121764
first, it can kick in on blocks whose conditions have been
folded to a constant, even though one of the edges will be
trivially folded.
second, it doesn't clean up the "if diamond" that it just
eliminated away. This is a problem because other simplifycfg
xforms kick in depending on the order of block visitation,
causing pointless work.
llvm-svn: 121762
when simplifying, allowing them to be eagerly turned into switches. This
is the last step required to get "Example 7" from this blog post:
http://blog.regehr.org/archives/320
On X86, we now generate this machine code, which (to my eye) seems better
than the ICC generated code:
_crud: ## @crud
## BB#0: ## %entry
cmpb $33, %dil
jb LBB0_4
## BB#1: ## %switch.early.test
addb $-34, %dil
cmpb $58, %dil
ja LBB0_3
## BB#2: ## %switch.early.test
movzbl %dil, %eax
movabsq $288230376537592865, %rcx ## imm = 0x400000017001421
btq %rax, %rcx
jb LBB0_4
LBB0_3: ## %lor.rhs
xorl %eax, %eax
ret
LBB0_4: ## %lor.end
movl $1, %eax
ret
llvm-svn: 121690
location in simplifycfg. In the old days, SimplifyCFG was never run on
the entry block, so we had to scan over all preds of the BB passed into
simplifycfg to do this xform, now we can just check blocks ending with
a condbranch. This avoids a scan over all preds of every simplified
block, which should be a significant compile-time perf win on functions
with lots of edges. No functionality change.
llvm-svn: 121668
(x & 2^n) ? 2^m+C : C
we can offset both arms by C to get the "(x & 2^n) ? 2^m : 0" form, optimize the
select to a shift and apply the offset afterwards.
llvm-svn: 121609
zextOrTrunc(), and APSInt methods extend(), extOrTrunc() and new method
trunc(), to be const and to return a new value instead of modifying the
object in place.
llvm-svn: 121120
(if available) as we go so that we get simple constantexprs not insane ones.
This fixes the failure of clang/test/CodeGenCXX/virtual-base-ctor.cpp
that the previous iteration of this patch had.
llvm-svn: 121111
optimization.
Consider:
static void foo() {
A = alloca
...
}
static void bar() {
B = alloca
...
call foo();
}
void main() {
bar()
}
The inliner proceeds bottom up, but lets pretend it decides not to inline foo
into bar. When it gets to main, it inlines bar into main(), and says "hey, I
just inlined an alloca "B" into main, lets remember that. Then it keeps going
and finds that it now contains a call to foo. It decides to inline foo into
main, and says "hey, foo has an alloca A, and I have an alloca B from another
inlined call site, lets reuse it". The problem with this of course, is that
the lifetime of A and B are nested, not disjoint.
Unfortunately I can't create a reasonable testcase for this: the one in the
PR is both huge and extremely sensitive, because you minor tweaks end up
causing foo to get inlined into bar too early. We already have tests for the
basic alloca merging optimization and this does not break them.
llvm-svn: 120995
memcpy's like:
memcpy(A, B)
memcpy(A, C)
we cannot delete the first memcpy as dead if A and C might be aliases.
If so, we actually get:
memcpy(A, B)
memcpy(A, A)
which is not correct to transform into:
memcpy(A, A)
This patch was heavily influenced by Jakub Staszak's patch in PR8728, thanks
Jakub!
llvm-svn: 120974
Should have no functional change other than the order of two transformations that are mutually-exclusive and the exact formatting of debug output.
Internally, it now stores the ConstantInt*s as Constant*s, and actual undef values instead of nulls.
llvm-svn: 120946
1. if the underlying pointer passed in can be resolved
to any argument or alloca, then we don't need to scan.
Previously we would only avoid the scan if the alloca
or byval was actually considered dead.
2. The dead store processing code is itself completely
dead and didn't handle volatile stores right anyway,
so delete it. This allows simplifying the interface
to RemoveAccessedObjects.
llvm-svn: 120467
made sense to me. We now have a set of dead stack objects, and
they become live when loaded. Fix a theoretical problem where
we'd pass in the wrong pointer to the alias query.
llvm-svn: 120465
If the call might read all the allocas, stop scanning early.
Convert a vector to smallvector, shrink SmallPtrSet to 16 instead
of 64 to avoid crazy linear scans.
llvm-svn: 120463
about pairs of AA::Location's instead of looking for MemDep's
"Def" predicate. This is more powerful and general, handling
memset/memcpy/store all uniformly, and implementing PR8701 and
probably obsoleting parts of memcpyoptimizer.
This also fixes an obscure bug with init.trampoline and i8
stores, but I'm not surprised it hasn't been hit yet. Enhancing
init.trampoline to carry the size that it stores would allow
DSE to be much more aggressive about optimizing them.
llvm-svn: 120406
contains "ref".
Enhance DSE to use a modref query instead of a store-specific hack
to generalize the "ignore may-alias stores" optimization to handle
memset and memcpy.
llvm-svn: 120368
1. Don't bother trying to optimize:
lifetime.end(ptr)
store(ptr)
as it is undefined, and therefore shouldn't exist.
2. Move the 'storing a loaded pointer' xform up, simplifying
the may-aliased store code.
llvm-svn: 120359
by my recent GVN improvement. Looking through a single layer of
PHI nodes when attempting to sink GEPs, we need to iteratively
look through arbitrary PHI nests.
llvm-svn: 120202
fairly systematic way in instcombine. Some of these cases were already dealt
with, in which case I removed the existing code. The case of Add has a bunch of
funky logic which covers some of this plus a few variants (considers shifts to be
a form of multiplication), which I didn't touch. The simplification performed is:
A*B+A*C -> A*(B+C). The improvement is to do this in cases that were not already
handled [such as A*B-A*C -> A*(B-C), which was reported on the mailing list], and
also to do it more often by not checking for "only one use" if "B+C" simplifies.
llvm-svn: 120024
allowing the memcpy to be eliminated.
Unfortunately, the requirements on byval's without explicit
alignment are really weak and impossible to predict in the
mid-level optimizer, so this doesn't kick in much with current
frontends. The fix is to change clang to set alignment on all
byval arguments.
llvm-svn: 119916
preserves LCSSA form out of ScalarEvolution and into the LoopInfo
class. Use it to check that SimplifyInstruction simplifications
are not breaking LCSSA form. Fixes PR8622.
llvm-svn: 119727
this was a tree of hashtables, and a query recursed into the table for the immediate dominator ad infinitum
if the initial lookup failed. This led to really bad performance on tall, narrow CFGs.
We can instead replace it with what is conceptually a multimap of value numbers to leaders (actually
represented by a hashtable with a list of Value*'s as the value type), and then
determine which leader from that set to use very cheaply thanks to the DFS numberings maintained by
DominatorTree. Because there are typically few duplicates of a given value, this scan tends to be
quite fast. Additionally, we use a custom linked list and BumpPtr allocation to avoid any unnecessary
allocation in representing the value-side of the multimap.
This change brings with it a 15% (!) improvement in the total running time of GVN on 403.gcc, which I
think is pretty good considering that includes all the "real work" being done by MemDep as well.
The one downside to this approach is that we can no longer use GVN to perform simple conditional progation,
but that seems like an acceptable loss since we now have LVI and CorrelatedValuePropagation to pick up
the slack. If you see conditional propagation that's not happening, please file bugs against LVI or CVP.
llvm-svn: 119714
refusing to optimize two memcpy's like this:
copy A <- B
copy C <- A
if it couldn't prove that noalias(B,C). We can eliminate
the copy by producing a memmove instead of memcpy.
llvm-svn: 119694
if it is passed as a byval argument. The byval argument will just be a
read, so it is safe to read from the original global instead. This allows
us to promote away the %agg.tmp alloca in PR8582
llvm-svn: 119686
instructions out of InstCombine and into InstructionSimplify. While
there, introduce an m_AllOnes pattern to simplify matching with integers
and vectors with all bits equal to one.
llvm-svn: 119536
hasConstantValue. I was leery of using SimplifyInstruction
while the IR was still in a half-baked state, which is the
reason for delaying the simplification until the IR is fully
cooked.
llvm-svn: 119494
systematically, CollapsePhi will always return null here. Note
that CollapsePhi did an extra check, isSafeReplacement, which
the SimplifyInstruction logic does not do. I think that check
was bogus - I guess we will soon find out! (It was originally
added in commit 41998 without a testcase).
llvm-svn: 119456
offload the work to hasConstantValue rather than do something more
complicated (such handling mutually recursive phis) because (1) it is
not clear it is worth it; and (2) if it is worth it, maybe such logic
would be better placed in hasConstantValue. Adjust some GVN tests
which are now cleaned up much further (eg: all phi nodes are removed).
llvm-svn: 119043
SimplifyAssociativeOrCommutative) "(A op C1) op C2" -> "A op (C1 op C2)",
which previously was only done if C1 and C2 were constants, to occur whenever
"C1 op C2" simplifies (a la InstructionSimplify). Since the simplifying operand
combination can no longer be assumed to be the right-hand terms, consider all of
the possible permutations. When compiling "gcc as one big file", transform 2
(i.e. using right-hand operands) fires about 4000 times but it has to be said
that most of the time the simplifying operands are both constants. Transforms
3, 4 and 5 each fired once. Transform 6, which is an existing transform that
I didn't change, never fired. With this change, the testcase is now optimized
perfectly with one run of instcombine (previously it required instcombine +
reassociate + instcombine, and it may just have been luck that this worked).
llvm-svn: 119002
"%z = %x and %y". If GVN can prove that %y equals %x, then it turns
this into "%z = %x and %x". With the new code, %z will be replaced
with %x everywhere (and then deleted). Previously %z would be value
numbered too, which is a waste of time. Also, while a clever value
numbering algorithm would give %z the same value number as %x, our
current one doesn't do so (at least I don't think it does). The new
logic has an essentially equivalent effect to what you would get if
%z was given the same value number as %x, i.e. it should make value
numbering smarter. While there, get hold of target data once at the
start rather than a gazillion times all over the place.
llvm-svn: 118923
testing for dereferenceable pointers into a helper function,
isDereferenceablePointer. Teach it how to reason about GEPs
with simple non-zero indices.
Also eliminate ArgumentPromtion's IsAlwaysValidPointer,
which didn't check for weak externals or out of range gep
indices.
llvm-svn: 118840
references. For example, this allows gvn to eliminate the load in
this example:
void foo(int n, int* p, int *q) {
p[0] = 0;
p[1] = 1;
if (n) {
*q = p[0];
}
}
llvm-svn: 118714
to optionally look for constant or local (alloca) memory.
Teach BasicAliasAnalysis::pointsToConstantMemory to look through Select
and Phi nodes, and to support looking for local memory.
Remove FunctionAttrs' PointsToLocalOrConstantMemory function, now that
AliasAnalysis knows all the tricks that it knew.
llvm-svn: 118412
consider it to be readonly. In fact, don't even consider it to be
readonly if it does a volatile load from an AllocaInst either (it
is debatable as to whether readonly would be correct or not in this
case; play safe for the moment). This fixes PR8279.
llvm-svn: 117783
This code had previously used 2*N, where N is the mask length, to represent
undef. That is not safe because the shufflevector operands may have more
than N elements -- they don't have to match the result type.
llvm-svn: 117721
Allow splats even if they don't match either of the original shuffles,
possibly due to undef entries in the shuffles masks. Radar 8597790.
Also fix some 80-column violations.
llvm-svn: 117719
needs to be guaranteed never to be run on an unreachable block. However, earlier block simplifications may have
changed the CFG to make block that were reachable when we began our iteration unreachable by the time we try to
simplify them. (Note that this also means that our depth-first iterators were potentially being invalidated).
This should not have a large impact on code quality, since later runs of instcombine should pick up these simplifications.
Fixes PR8506.
llvm-svn: 117709
it isn't unreachable and should not be zapped. The check for the entry block
was missing in one case: a block containing a unwind instruction. While there,
do some small cleanups: "M" is not a great name for a Function* (it would be
more appropriate for a Module*), change it to "Fn"; use Fn in more places.
llvm-svn: 117224
must be called in the pass's constructor. This function uses static dependency declarations to recursively initialize
the pass's dependencies.
Clients that only create passes through the createFooPass() APIs will require no changes. Clients that want to use the
CommandLine options for passes will need to manually call the appropriate initialization functions in PassInitialization.h
before parsing commandline arguments.
I have tested this with all standard configurations of clang and llvm-gcc on Darwin. It is possible that there are problems
with the static dependencies that will only be visible with non-standard options. If you encounter any crash in pass
registration/creation, please send the testcase to me directly.
llvm-svn: 116820
perform initialization without static constructors AND without explicit initialization
by the client. For the moment, passes are required to initialize both their
(potential) dependencies and any passes they preserve. I hope to be able to relax
the latter requirement in the future.
llvm-svn: 116334
formulae which become illegal as a result of the offset updating don't
escape.
This is for rdar://8529692. No testcase yet, because the given cases
hit use-list ordering differences.
llvm-svn: 116093
This doesn't usually matter, because the other heuristics usually
succeed regardless, but it's good to keep the register use
bookkeeping consistent.
llvm-svn: 116005
Anyone interested in more general PRE would be better served by implementing it separately, to get real
anticipation calculation, etc.
llvm-svn: 115337
The x86_mmx type is used for MMX intrinsics, parameters and
return values where these use MMX registers, and is also
supported in load, store, and bitcast.
Only the above operations generate MMX instructions, and optimizations
do not operate on or produce MMX intrinsics.
MMX-sized vectors <2 x i32> etc. are lowered to XMM or split into
smaller pieces. Optimizations may occur on these forms and the
result casted back to x86_mmx, provided the result feeds into a
previous existing x86_mmx operation.
The point of all this is prevent optimizations from introducing
MMX operations, which is unsafe due to the EMMS problem.
llvm-svn: 115243
Because of this, we cannot use the Simplify* APIs, as they can assert-fail on unreachable code. Since it's not easy to determine
if a given threading will cause a block to become unreachable, simply defer simplifying simplification to later InstCombine and/or
DCE passes.
llvm-svn: 115082
register pressure and thus excess spills, which we don't currently recover from well. This should
be re-evaluated in the future if our ability to generate good spills/splits improves.
Partial fix for <rdar://problem/7635585>.
llvm-svn: 114919
This reverts revision 114633. It was breaking llvm-gcc-i386-linux-selfhost.
It seems there is a downstream bug that is exposed by
-cgp-critical-edge-splitting=0. When that bug is fixed, this patch can go back
in.
Note that the changes to tailcallfp2.ll are not reverted. They were good are
required.
llvm-svn: 114859
Splitting critical edges at the merge point only addressed part of the issue; it is also possible for non-post-domination
to occur when the path from the load to the merge has branches in it. Unfortunately, full anticipation analysis is
time-consuming, so for now approximate it. This is strictly more conservative than real anticipation, so we will miss
some cases that real PRE would allow, but we also no longer insert loads into paths where they didn't exist before. :-)
This is a very slight net positive on SPEC for me (0.5% on average). Most of the benchmarks are largely unaffected, but
when it pays off it pays off decently: 181.mcf improves by 4.5% on my machine.
llvm-svn: 114785
"external" even when doing lazy bitcode loading. This was broken because
a function that is not materialized fails the !isDeclaration() test.
llvm-svn: 114666
truncates are free only in the case where the extended type is legal but the
load type is not. If both types are illegal, such as when they are too big,
the load may not be legalized into an extended load.
llvm-svn: 114568
load when the type of the load is not legal, even if truncates are not free.
The load is going to be legalized to an extending load anyway.
llvm-svn: 114488
walking the asm arguments once and stashing their Values. This is
wrong because the same memory location can be in the list twice, and
if the first one has a sunkaddr substituted, the stashed value for the
second one will be wrong (use-after-free). PR 8154.
llvm-svn: 114104
to expose greater opportunities for store narrowing in codegen. This patch fixes a potential
infinite loop in instcombine caused by one of the introduced transforms being overly aggressive.
llvm-svn: 113763
This can result in increased opportunities for store narrowing in code generation. Update a number of
tests for this change. This fixes <rdar://problem/8285027>.
Additionally, because this inverts the order of ors and ands, some patterns for optimizing or-of-and-of-or
no longer fire in instances where they did originally. Add a simple transform which recaptures most of these
opportunities: if we have an or-of-constant-or and have failed to fold away the inner or, commute the order
of the two ors, to give the non-constant or a chance for simplification instead.
llvm-svn: 113679
not unrolling loops that contain calls that would be better off getting inlined. This mostly
comes up when an interleaved devirtualization pass has devirtualized a call which the inliner
will inline on a future pass. Thus, rather than blocking all loops containing calls, add
a metric for "inline candidate calls" and block loops containing those instead.
llvm-svn: 113535
unrolling threshold to the optimize-for-size threshold. Basically, for loops containing calls, unrolling
can still be profitable as long as the loop is REALLY small.
llvm-svn: 113439
The threshold value of 50 is arbitrary, and I chose it simply by analogy to the inlining thresholds, where
the baseline unrolling threshold is slightly smaller than the baseline inlining threshold. This could
undoubtedly use some tuning.
llvm-svn: 113306
turning (fptrunc (sqrt (fpext x))) -> (sqrtf x) is great, but we have
to delete the original sqrt as well. Not doing so causes us to do
two sqrt's when building with -fmath-errno (the default on linux).
llvm-svn: 113260
Switch from isWeakForLinker to mayBeOverridden which is more accurate.
Add more statistics and debugging info. Add comments. Move static function
outside anonymous namespace.
llvm-svn: 113190
in the duplicated block instead of duplicating them.
Duplicating them into the end of the loop and the preheader
means that we got a phi node in the header of the loop,
which prevented LICM from hoisting them. GVN would
usually come around later and merge the duplicated
instructions so we'd get reasonable output... except that
anything dependent on the shoulda-been-hoisted value can't
be hoisted. In PR5319 (which this fixes), a memory value
didn't get promoted.
llvm-svn: 113134
Loop::hasLoopInvariantOperands method. Remove
a useless and confusing Loop::isLoopInvariant(Instruction)
method, which didn't do what you thought it did.
No functionality change.
llvm-svn: 113133
location is being re-stored to the memory location. We would get
a dangling pointer from the SSAUpdate data structure and miss a
use. This fixes PR8068
llvm-svn: 113042
I'm sure it is harmless. Original commit message:
If PrototypeValue is erased in the middle of using the SSAUpdator
then the SSAUpdator may access freed memory. Instead, simply pass
in the type and name explicitly, which is all that was used anyway.
llvm-svn: 112810
on llvmdev: SRoA is introducing MMX datatypes like <1 x i64>,
which then cause random problems because the X86 backend is
producing mmx stuff without inserting proper emms calls.
In the short term, force off MMX datatypes. In the long term,
the X86 backend should not select generic vector types to MMX
registers. This is being worked on, but won't be done in time
for 2.8. rdar://8380055
llvm-svn: 112696
two are weak, we make them thunks to a new strong function) so don't iterate
through the function list as we're modifying it.
Also add back the outermost loop which got removed during the cleanups.
llvm-svn: 112595
This actually exposed an infinite recursion bug in ComputeValueKnownInPredecessors which theoretically already existed (in JumpThreading's
handling of and/or of i1's), but never manifested before. This patch adds a tracking set to prevent this case.
llvm-svn: 112589
instead of PromoteMemToReg. This allows it to stop using DF and DT,
eliminating a computation of DT and DF from clang -O3. Clang is now
down to 2 runs of DomFrontier.
llvm-svn: 112457
assertingvh so we get a violent explosion if the pointer dangles.
2) Fix AliasSetTracker::deleteValue to remove call sites with
by-pointer comparisons instead of by-alias queries. Using
findAliasSetForCallSite can cause alias sets to get merged
when they shouldn't, and can also miss alias sets when the
call is readonly.
#2 fixes PR6889, which only repros with a .c file :(
llvm-svn: 112452
LICM correctly. When sinking an instruction, it should not add
entries for the sunk instruction to the AST, it should remove
the entry for the sunk instruction. The blocks being sunk to
are not in the loop, so their instructions shouldn't be in the
AST (yet)!
llvm-svn: 112447
keeping them around until the pass is destroyed, keep them
around a) just when useful (not for outer loops) and b) destroy
them right after we use them. This should reduce memory use
and fixes potential bugs where a loop is deleted and another
loop gets allocated to the same address.
llvm-svn: 112446
LSRInstance data structures up to date. This fixes some
pessimizations caused by stale data which will be exposed
in an upcoming change.
llvm-svn: 112440
A = shl x, 42
...
B = lshr ..., 38
which can be transformed into:
A = shl x, 4
...
iff we can prove that the would-be-shifted-in bits
are already zero. This eliminates two shifts in the testcase
and allows eliminate of the whole i128 chain in the real example.
llvm-svn: 112314
framework, which is good at ripping through bitfield
operations. This generalize a bunch of the existing
xforms that instcombine does, such as
(x << c) >> c -> and
to handle intermediate logical nodes. This is useful for
ripping up the "promote to large integer" code produced by
SRoA.
llvm-svn: 112304
by the SRoA "promote to large integer" code, eliminating
some type conversions like this:
%94 = zext i16 %93 to i32 ; <i32> [#uses=2]
%96 = lshr i32 %94, 8 ; <i32> [#uses=1]
%101 = trunc i32 %96 to i8 ; <i8> [#uses=1]
This also unblocks other xforms from happening, now clang is able to compile:
struct S { float A, B, C, D; };
float foo(struct S A) { return A.A + A.B+A.C+A.D; }
into:
_foo: ## @foo
## BB#0: ## %entry
pshufd $1, %xmm0, %xmm2
addss %xmm0, %xmm2
movdqa %xmm1, %xmm3
addss %xmm2, %xmm3
pshufd $1, %xmm1, %xmm0
addss %xmm3, %xmm0
ret
on x86-64, instead of:
_foo: ## @foo
## BB#0: ## %entry
movd %xmm0, %rax
shrq $32, %rax
movd %eax, %xmm2
addss %xmm0, %xmm2
movapd %xmm1, %xmm3
addss %xmm2, %xmm3
movd %xmm1, %rax
shrq $32, %rax
movd %eax, %xmm0
addss %xmm3, %xmm0
ret
This seems pretty close to optimal to me, at least without
using horizontal adds. This also triggers in lots of other
code, including SPEC.
llvm-svn: 112278
fix: add a flag to MapValue and friends which indicates whether
any module-level mappings are being made. In the common case of
inlining, no module-level mappings are needed, so MapValue doesn't
need to examine non-function-local metadata, which can be very
expensive in the case of a large module with really deep metadata
(e.g. a large C++ program compiled with -g).
This flag is a little awkward; perhaps eventually it can be moved
into the ClonedCodeInfo class.
llvm-svn: 112190
which does the same thing. This eliminates redundant code and
handles MDNodes better. MDNode linking still doesn't fully
work yet though.
llvm-svn: 111941
that it avoids a lot of unnecessary cloning by avoiding remapping
MDNode cycles when none of the nodes in the cycle actually need to
be remapped. Also it uses the new temporary MDNode mechanism.
llvm-svn: 111922
from the LHS should disable reconsidering that pred on the
RHS. However, knowing something about the pred on the RHS
shouldn't disable subsequent additions on the RHS from
happening.
llvm-svn: 111349
uninteresting, just put all the operands on one list and make
GenerateReassociations make the decision about what's interesting.
This is simpler, and it avoids an extra ScalarEvolution::getAddExpr call.
llvm-svn: 111133
- Eliminate redundant successors.
- Convert an indirectbr with one successor into a direct branch.
Also, generalize SimplifyCFG to be able to be run on a function entry block.
It knows quite a few simplifications which are applicable to the entry
block, and it only needs a few checks to avoid trouble with the entry block.
llvm-svn: 111060
patterns generated by clang for transpose of a matrix in generic vectors. This is made
of two parts:
1) Propagating vector extracts of hi/lo half into their users
2) Recognizing an insertion of even elements followed by the odd elements as an unpack.
Testcase to come, but this shrinks the # of shuffle instructions generated on x86 from ~40 to the minimal 8.
llvm-svn: 110734
Further clean up the comparison function by removing overly generalized
"domains".
Remove all understanding of ELF aliases and simplify folding code and comments.
llvm-svn: 110434
eliminate several const_casts.
Make CallSite implicitly convertible to ImmutableCallSite.
Rename the getModRefBehavior for intrinsic IDs to
getIntrinsicModRefBehavior to avoid overload ambiguity with CallSite,
which happens to be implicitly convertible to bool.
llvm-svn: 110155
Start cleaning up MergeFunctions to look more like the rest of LLVM. The
primary change here is to move the methods responsible for comparison into the
new FunctionComparator object. Some comments added. There's more to do.
llvm-svn: 110021
exactly what bugpoint expected it to do.
There was also only one user of
BlockExtractorPass(const std::vector<BasicBlock*> &B), so just remove it and
make BlockExtractorPass read BlockFile.
This fixes bugpoint's block extraction.
Nick, please review.
llvm-svn: 109936
alloca instructions (constrained by their internal encoding),
and add error checking for it. Fix an instcombine bug which
generated huge alignment values (null is infinitely aligned).
This fixes undefined behavior noticed by John Regehr.
llvm-svn: 109643
dependence on DominanceFrontier. Instead, add an explicit DominanceFrontier
pass in StandardPasses.h to ensure that it gets scheduled at the right
time.
Declare that loop unrolling preserves ScalarEvolution, and shuffle some
getAnalysisUsages.
This eliminates one LoopSimplify and one LCCSA run in the standard
compile opts sequence.
llvm-svn: 109413
different widths. In a use with a narrower fixup, formulae
may be wider than the fixup, in which case the high bits
aren't necessarily meaningful, so it isn't safe to reuse
them for uses with wider fixups.
This fixes PR7618, though the testcase is too large for a
reasonable regression test, since it heavily dependes on
hitting LSR's heuristics in a certain way.
llvm-svn: 108455
the corresponding or-icmp-and pattern. This has the added benefit of doing
the matching earlier, and thus being less susceptible to being confused by
earlier transforms.
llvm-svn: 108429
it *changing* the things it replaces, not just causing them
to drop to null. There is no functionality change yet, but
this is required for a subsequent patch.
llvm-svn: 108414
"bonus" instruction to be speculatively executed. Add a heuristic to
ensure we're not tripping up out-of-order execution by checking that this bonus
instruction only uses values that were already guaranteed to be available.
This allows us to eliminate the short circuit in (x&1)&&(x&2).
llvm-svn: 108351
by a return that returns a constant, while elsewhere in the function
another return instruction returns a different constant. This is a
special case of accumulator recursion, so just generalize the existing
logic a bit.
llvm-svn: 108241
operation, but the way it's implemented requires the operation to also be
commutative. So add a check for commutativity (and tweak the corresponding
comments). This makes no difference in practice since every associative
LLVM instruction is also commutative! Here's an example to show the need
for commutativity: the accum_recursion.ll testcase calculates the factorial
function. Before the transformation the result of a call is
((((1*1)*2)*3)...)*x
while afterwards it is
(((1*x)*(x-1))...*2)*1
which clearly requires both associativity and commutativity of * to be equal
to the original.
llvm-svn: 108056
(X >s -1) ? C1 : C2 and (X <s 0) ? C2 : C1
into ((X >>s 31) & (C2 - C1)) + C1, avoiding the conditional.
This optimization could be extended to take non-const C1 and C2 but we better
stay conservative to avoid code size bloat for now.
for
int sel(int n) {
return n >= 0 ? 60 : 100;
}
we now generate
sarl $31, %edi
andl $40, %edi
leal 60(%rdi), %eax
instead of
testl %edi, %edi
movl $60, %ecx
movl $100, %eax
cmovnsl %ecx, %eax
llvm-svn: 107866
builds to "Release". The default build is unchanged (optimization on,
assertions on), however it is now called Release+Asserts. The intent
is that future LLVM releases released via llvm.org will be Release builds
in the new sense, i.e. will have assertions disabled (currently they have
assertions enabled, for a more than 20% slowdown). This will bring them
in line with MacOS releases, which ship with assertions disabled. It also
means that "Release" now means the same things in make and cmake builds:
cmake already disables assertions for "Release" builds AFAICS.
llvm-svn: 107758
have any effect, and second, deleting stores can potentially invalidate
an AliasAnalysis, and there's currently no notification for this.
llvm-svn: 107496
Objective-C metadata types which should be marked as "weak", but which the
linker will remove upon final linkage. However, this linkage isn't specific to
Objective-C.
For example, the "objc_msgSend_fixup_alloc" symbol is defined like this:
.globl l_objc_msgSend_fixup_alloc
.weak_definition l_objc_msgSend_fixup_alloc
.section __DATA, __objc_msgrefs, coalesced
.align 3
l_objc_msgSend_fixup_alloc:
.quad _objc_msgSend_fixup
.quad L_OBJC_METH_VAR_NAME_1
This is different from the "linker_private" linkage type, because it can't have
the metadata defined with ".weak_definition".
Currently only supported on Darwin platforms.
llvm-svn: 107433
such a way that debug info for symbols preserved even if symbols are
optimized away by the optimizer.
Add new special pass to remove debug info for such symbols.
llvm-svn: 107416
metadata types which should be marked as "weak", but which the linker will
remove upon final linkage. For example, the "objc_msgSend_fixup_alloc" symbol is
defined like this:
.globl l_objc_msgSend_fixup_alloc
.weak_definition l_objc_msgSend_fixup_alloc
.section __DATA, __objc_msgrefs, coalesced
.align 3
l_objc_msgSend_fixup_alloc:
.quad _objc_msgSend_fixup
.quad L_OBJC_METH_VAR_NAME_1
This is different from the "linker_private" linkage type, because it can't have
the metadata defined with ".weak_definition".
llvm-svn: 107205
is stripped off. Currently set unconditionally, since the API
does not provide a way of working out if anything was actually
stripped off.
llvm-svn: 107142
large integers, the first inserted value would always create
an 'or X, 0'. Even though this is trivially zapped by
instcombine, don't bother creating this pointless instruction.
llvm-svn: 106979
the returned value after the tail call if it differs from other return
values. The optimal thing to do would be to introduce a phi node for
the return value, but for the moment just fix the miscompile.
llvm-svn: 106947
SCEVUnknown values which are loop-variant, as LSR can't do anything
interesting with these values in any case. This fixes very slow compile
times on loops which have large numbers of such values.
llvm-svn: 106897
for an "i" constraint should get lowered; PR 6309. While
this argument was passed around a lot, this is the only
place it was used, so it goes away from a lot of other
places.
llvm-svn: 106893
Failure to seed metdata in such cases causes troubles when in a cloned module, metadata from a new module refers to values in old module. Usually this results in mysterious bugpoint crashes. For example,
Checking to see if we can delete global inits: Unknown constant!
UNREACHABLE executed at /d/g/llvm/lib/Bitcode/Writer/BitcodeWriter.cpp:904!
llvm-svn: 106592
use sharing map. The reconcileNewOffset logic already forces a
separate use if the kinds differ, so incorporating the kind in the
key means we can track more sharing opportunities.
More sharing means fewer total uses to track, which means smaller
problem sizes, which means the conservative throttles don't kick
in as often.
llvm-svn: 106396
Changed directly instead of using a return value.
Rename FilterOutUndesirableDedicatedRegisters's Changed variable to
distinguish it from LSRInstance's Changed member.
llvm-svn: 104269
operand on the left, the interesting operand is on the right. This
fixes a bug where LSR was failing to recognize ICmpZero uses,
which led it to be unable to reverse the induction variable in the
attached testcase.
Delete test/CodeGen/X86/stack-color-with-reg-2.ll, because its test
is extremely fragile and hard to meaningfully update.
llvm-svn: 104262
vector<>::push_back() in:
int foo(vector<int> &a, vector<unsigned> &b) {
a.push_back(10);
b.push_back(11);
}
to two calls to the same push_back function, or fold away the two copies of
push_back() in:
struct T { int; };
struct S { char; };
vector<T*> t;
vector<S*> s;
void f(T *x) { t.push_back(x); }
void g(S *x) { s.push_back(x); }
but leave f() and g() separate, since they refer to two different global
variables.
llvm-svn: 103698
on RAUW of functions, this is a correctness issue instead of a mere memory
usage problem.
No testcase until the new MergeFunctions can land.
llvm-svn: 103653
when it detects undefined behavior. llvm.trap generally codegens into some
thing really small (e.g. a 2 byte ud2 instruction on x86) and debugging this
sort of thing is "nontrivial". For example, we now compile:
void foo() { *(int*)0 = 42; }
into:
_foo:
pushl %ebp
movl %esp, %ebp
ud2
Some may even claim that this is a security hole, though that seems dubious
to me. This addresses rdar://7958343 - Optimizing away null dereference
potentially allows arbitrary code execution
llvm-svn: 103356
with a vector input and output into a shuffle vector. This sort of
sequence happens when the input code stores with one type and reloads
with another type and then SROA promotes to i96 integers, which make
everyone sad.
This fixes rdar://7896024
llvm-svn: 103354
LSRUse's Regs set after all pruning is done, rather than trying
to do it on the fly, which can produce an incomplete result.
This fixes a case where heuristic pruning was stripping all
formulae from a use, which led the solver to enter an infinite
loop.
Also, add a few asserts to diagnose this kind of situation.
llvm-svn: 103328
indirect branches in all the predecessors. This avoids unnecessarily
splitting edges in cases where load PRE is not possible anyway.
Thanks to Jakub Staszak for pointing this out.
llvm-svn: 103034
halting analysis, it is illegal to delete a call to a read-only function.
The correct solution is almost certainly to add a "must halt" attribute and
only allow deletions in its presence.
XFAIL the relevant testcase for now.
llvm-svn: 102831
that can have a big effect :). The first is to enable the
iterative SCC passmanager juice that kicks in when the
scc passmgr detects that a function pass has devirtualized
a call. In this case, it will rerun all the passes it
manages on the SCC, up to the iteration count limit (4). This
is useful because a function pass may devirualize a call, and
we want the inliner to inline it, or pruneeh to infer stuff
about it, etc.
The second patch is to add *all* call sites to the
DevirtualizedCalls list the inliner uses. This list is
about to get renamed, but the jist of this is that the
inliner now reconsiders *all* inlined call sites as candidates
for further inlining. The intuition is this that in cases
like this:
f() { g(1); } g(int x) { h(x); }
We analyze this bottom up, and may decide that it isn't
profitable to inline H into G. Next step, we decide that it is
profitable to inline G into F, and do so, which means that F
now calls H. Even though the call from G -> H may not have been
profitable to inline, the call from F -> H may be (in this case
because a constant allows folding etc).
In my spot checks, this doesn't have a big impact on code. For
example, the LLC output for 252.eon grew from 0.02% (from
317252 to 317308) and 176.gcc actually shrunk by .3% (from 1525612
to 1520964 bytes). 252.eon never iterated in the SCC Passmgr,
176.gcc iterated at most 1 time.
llvm-svn: 102823
that appear due to inlining a callee as candidates for
futher inlining, but a recent patch made it do this if
those call sites were indirect and became direct.
Unfortunately, in bizarre cases (see testcase) doing this
can cause us to infinitely inline mutually recursive
functions into callers not in the cycle. Fix this by
keeping track of the inline history from which callsite
inline candidates got inlined from.
This shouldn't affect any "real world" code, but is required
for a follow on patch that is coming up next.
llvm-svn: 102822
add a version of createLowerInvokePass that allows the client
to specify whether it wants "expensive" or "cheap" lowering.
Patch by Alex Mac!
llvm-svn: 102402
This fixes a bug where calls inlined into an invoke would get
changed into an invoke but the array would keep pointing to
the (now dead) call. The improved inliner behavior is still
disabled for now.
llvm-svn: 102196
that appear in the SCC as a result of inlining as candidates
for inlining. Change this so that it *does* consider call
sites that change from being indirect to being direct as a
result of inlining. This allows it to completely
"devirtualize" the testcase.
llvm-svn: 102146
arguments are handled with a new InlineFunctionInfo class. This
makes it easier to extend InlineFunction to return more info in the
future.
llvm-svn: 102137
define void @f3(void (i8*)* %__f) ssp {
entry:
call void %__f(i8* undef)
unreachable
}
define void @f4(i8* %this) ssp align 2 {
entry:
call void @f3(void (i8*)* @f2) ssp
ret void
}
The inliner is turning the indirect call to %__f into a direct
call to F2. Make the call graph more precise when this happens.
The inliner doesn't revisit call sites introduced by inlining,
so there isn't an easy way to test for this, but a more precise
callgraph is a good thing.
llvm-svn: 102131
condition we're unswitching on. In this case, don't try to
simplify the second copy of the loop which may be dead or not,
but is probably a constant now. This fixes PR6879
llvm-svn: 101870
Arg promotion was deleting call graph nodes that still had references
from the 'indirect' CGN. Like the inliner, it should only delete the
function if all references are gone.
llvm-svn: 101845
just ask ScalarEvolution for it on demand. This helps IVUsers be more robust
in the case of expressions changing underneath it. This fixes PR6862.
llvm-svn: 101819
to determine where to place PHIs by iteratively comparing reaching definitions
at each block. That was just plain wrong. This version now computes the
dominator tree within the subset of the CFG where PHIs may need to be placed,
and then places the PHIs in the iterated dominance frontier of each definition.
The rest of the patch is mostly the same, with a few more performance
improvements added in.
llvm-svn: 101612
to CallGraphSCCPass's instead of passing around a
std::vector<CallGraphNode*>. No functionality change,
but now we have a much tidier interface.
llvm-svn: 101558
with a fix for self-hosting
rotate CallInst operands, i.e. move callee to the back
of the operand array
the motivation for this patch are laid out in my mail to llvm-commits:
more efficient access to operands and callee, faster callgraph-construction,
smaller compiler binary
llvm-svn: 101465
with a fix
rotate CallInst operands, i.e. move callee to the back
of the operand array
the motivation for this patch are laid out in my mail to llvm-commits:
more efficient access to operands and callee, faster callgraph-construction,
smaller compiler binary
llvm-svn: 101397
of the operand array
the motivation for this patch are laid out in my mail to llvm-commits:
more efficient access to operands and callee, faster callgraph-construction,
smaller compiler binary
llvm-svn: 101364
The commit "Adding IPSCCP and Internalize passes to the C-bindings" introduced
new dependencies for IPO. Add these to the CMAKE build as otherwise the
BUILD_SHARED_LIBS=1 build fails.
llvm-svn: 101313
- TryToOptimizeStoreOfMallocToGlobal should check if TargetData is available and bail out if it is not. The transformations being done requires TD.
llvm-svn: 101285
it can check whether the visible direct callers are passing in parameters to
dead arguments and replace those with undef.
This reinstates r94322 with bugs fixed.
llvm-svn: 101213
numerator is an induction variable. For example, with code like this:
for (i=0;i<n;++i)
x[i%n] = 0;
IndVarSimplify will now recognize that i is always less than n inside
the loop, and eliminate the remainder.
llvm-svn: 101113
expression is a UDiv and it doesn't appear that the UDiv came from
the user's source.
ScalarEvolution has recently figured out how to compute a tripcount
expression for the inner loop in
SingleSource/Benchmarks/Shootout/sieve.c, using a udiv. Emitting a
udiv instruction dramatically slows down the enclosing loop.
llvm-svn: 101068
a ScalarEvolution bug with overflow handling is fixed, the normal analysis
code will automatically decline to operate on the icmp instructions which
are responsible for the loop exit.
llvm-svn: 101032
instead of deleting just the user. This makes it more consistent with
other code in IndVarSimplify, and theoretically can eliminate more users
earlier.
llvm-svn: 101027
the loop exit test. This usually doesn't come up for a variety of
reasons, but it isn't impossible, so make IndVarSimplify handle it
conservatively.
llvm-svn: 101008
variables. For example, with code like this:
for (i=0;i<n;++i)
if (i<n)
x[i] = 0;
IndVarSimplify will now recognize that i is always less than n inside
the loop, and eliminate the if.
llvm-svn: 101000
parameters in the CBE by implicitly adding a fixed argument.
This allows eliminating a work-around from DAE. Patch by
Sylvere Teissier!
llvm-svn: 100944
into adjacent loops. Also, ensure that the insert position is
dominated by the loop latch of any loop in the post-inc set which
has a latch.
llvm-svn: 100906
forced constant is changed to a constant, we would end
up adding the instruction to the wrong worklist,
preventing it from being properly revisited. This fixes
rdar://7832370
llvm-svn: 100837
explicitly split into stride-and-offset pairs. Also, add the
ability to track multiple post-increment loops on the same expression.
This refines the concept of "normalizing" SCEV expressions used for
to post-increment uses, and introduces a dedicated utility routine for
normalizing and denormalizing expressions.
This fixes the expansion of expressions which are post-increment users
of more than one loop at a time. More broadly, this takes LSR another
step closer to being able to reason about more than one loop at a time.
llvm-svn: 100699
undefs in branches/switches, we have two cases: a branch on a literal
undef or a branch on a symbolic value which is undef. If we have a
literal undef, the code was correct: forcing it to a constant is the
right thing to do.
If we have a branch on a symbolic value that is undef, we should force
the symbolic value to a constant, which then makes the successor block
live. Forcing the condition of the branch to being a constant isn't
safe if later paths become live and the value becomes overdefined. This
is the case that 'forcedconstant' is designed to handle, so just use it.
This fixes rdar://7765019 but there is no good testcase for this, the
one I have is too insane to be useful in the future.
llvm-svn: 100478
Added support for address spaces and added a isVolatile field to memcpy, memmove, and memset,
e.g., llvm.memcpy.i32(i8*, i8*, i32, i32) -> llvm.memcpy.p0i8.p0i8.i32(i8*, i8*, i32, i32, i1)
llvm-svn: 100304
exits the loop. With this information we can guarantee
the iteration count of the loop is bounded by the
compare. I think this xforms is finally safe now.
llvm-svn: 100285
checker. Amusingly, we already had tests that we should
have rejects because they would be miscompiled in the
testsuite.
The remaining issue with this is that we don't check that
the branch causes us to exit the loop if it fails, so we
don't actually know if we remain in bounds.
llvm-svn: 100284
to a signed vs unsigned value depending on the sign of the
constant fp means that we can't distinguish between a
truly negative number and a positive number so large the
32nd bit is set. So, do don't this!
llvm-svn: 100283
this cleans up a bunch of code and also fixes several crashes and
miscompiles. More to come unfortunately, this optimization
is quite broken.
llvm-svn: 100270
(what was I thinking?) and there's also a problem with LCSSA. I'll try again
later with fixes.
--- Reverse-merging r100263 into '.':
U lib/Transforms/Utils/SSAUpdater.cpp
--- Reverse-merging r100177 into '.':
G lib/Transforms/Utils/SSAUpdater.cpp
--- Reverse-merging r100148 into '.':
G lib/Transforms/Utils/SSAUpdater.cpp
--- Reverse-merging r100147 into '.':
U include/llvm/Transforms/Utils/SSAUpdater.h
G lib/Transforms/Utils/SSAUpdater.cpp
--- Reverse-merging r100131 into '.':
G include/llvm/Transforms/Utils/SSAUpdater.h
G lib/Transforms/Utils/SSAUpdater.cpp
--- Reverse-merging r100130 into '.':
G lib/Transforms/Utils/SSAUpdater.cpp
--- Reverse-merging r100126 into '.':
G include/llvm/Transforms/Utils/SSAUpdater.h
G lib/Transforms/Utils/SSAUpdater.cpp
--- Reverse-merging r100050 into '.':
D test/Transforms/GVN/2010-03-31-RedundantPHIs.ll
--- Reverse-merging r100047 into '.':
G include/llvm/Transforms/Utils/SSAUpdater.h
G lib/Transforms/Utils/SSAUpdater.cpp
llvm-svn: 100264
Added support for address spaces and added a isVolatile field to memcpy, memmove, and memset,
e.g., llvm.memcpy.i32(i8*, i8*, i32, i32) -> llvm.memcpy.p0i8.p0i8.i32(i8*, i8*, i32, i32, i1)
llvm-svn: 100191
is necessary. Inherits from new templated baseclass CallSiteBase<>
which is highly customizable. Base CallSite on it too, in a configuration
that allows full mutation.
Adapt some call sites in analyses to employ ImmutableCallSite.
llvm-svn: 100100
PHIs. The previous algorithm was unable to reliably detect when existing
PHIs in a cycle can be reused. I'm still working on reducing a testcase.
Radar 7711900.
llvm-svn: 100047
generate wrong code pretty much anywhere AFAICT.
A case that hits the bug reproducibly is impossible,
but the situation was like this:
Addr = ...
Store -> Addr
Addr2 = GEP , 0, 0
Store -> Addr2
Handling the first store, the code changed replaced Addr
with a sunkaddr and deleted Addr, but not its table
entry. Code in OptimizedBlock replaced Addr2 with a
bitcast; if that happened to reuse the memory of Addr,
the old table entry was erroneously found when handling
the second store.
llvm-svn: 100044
e.g., llvm.memcpy.i32(i8*, i8*, i32, i32) -> llvm.memcpy.p0i8.p0i8.i32(i8*, i8*, i32, i32, i1)
A update of langref will occur in a subsequent checkin.
llvm-svn: 99928
I have audited all getOperandNo calls now, fixing
hidden assumptions. CallSite related uglyness will
be eliminated successively.
Note this patch has a long and griveous history,
for all the back-and-forths have a look at
CallSite.h's log.
llvm-svn: 99399
for the noinline attribute, and make the inliner refuse to
inline a call site when the call site is marked noinline even
if the callee isn't. This fixes PR6682.
llvm-svn: 99341
so that the SCEVExpander doesn't retain a dangling pointer as its
insert position. The dangling pointer in this case wasn't ever used
to insert new instructions, but it was causing trouble with
SCEVExpander's code for automatically advancing its insert position
past debug intrinsics.
This fixes use-after-free errors that valgrind noticed in
test/Transforms/IndVarSimplify/2007-06-06-DeleteDanglesPtr.ll and
test/Transforms/IndVarSimplify/exit_value_tests.ll.
llvm-svn: 99036
This time I did a self-hosted bootstrap on Linux x86-64,
with no problems. Let's see how darwin 64-bit self-hosting
goes. At the first sign of failure I'll back this out.
Maybe the valgrind bots give me a hint of what may be wrong
(it at all).
llvm-svn: 98957
The Caller cost info would be reset everytime a callee was inlined. If the
caller has lots of calls and there is some mutual recursion going on, the
caller cost info could be calculated many times.
This patch reduces inliner runtime from 240s to 0.5s for a function with 20000
small function calls.
This is a more conservative version of r98089 that doesn't break the clang
test CodeGenCXX/temp-order.cpp. That test relies on rather extreme inlining
for constant folding.
llvm-svn: 98099
The Caller cost info would be reset everytime a callee was inlined. If the
caller has lots of calls and there is some mutual recursion going on, the
caller cost info could be calculated many times.
This patch reduces inliner runtime from 240s to 0.5s for a function with 20000
small function calls.
llvm-svn: 98089
out the remainder of the calls that we should lower in some way and
move the tests to the new correct directory. Fix up tests that are now
optimized more than they were before by -instcombine.
llvm-svn: 97875
Log:
Transform @llvm.objectsize to integer if the argument is a result of malloc of known size.
Modified:
llvm/trunk/lib/Transforms/InstCombine/InstCombineCalls.cpp
llvm/trunk/test/Transforms/InstCombine/objsize.ll
It appears to be causing swb and nightly test failures.
llvm-svn: 97866
can be used in more places. Add an argument for the TargetData that
most of them need. Update for the getInt8PtrTy() change. Should be
no functionality change.
llvm-svn: 97844
parts of the cmp|cmp and cmp&cmp folding logic wasn't prepared for vectors
(unrelated to the bug but noticed while in the code) and the code was
*definitely* not safe to use by the (cast icmp)|(cast icmp) handling logic
that I added in r95855. Fix all this up by changing the various routines
to more consistently use IRBuilder and not pass in the I which had the wrong
type.
llvm-svn: 97801
long test(long x) { return (x & 123124) | 3; }
Currently compiles to:
_test:
orl $3, %edi
movq %rdi, %rax
andq $123127, %rax
ret
This is because instruction and DAG combiners canonicalize
(or (and x, C), D) -> (and (or, D), (C | D))
However, this is only profitable if (C & D) != 0. It gets in the way of the
3-addressification because the input bits are known to be zero.
llvm-svn: 97616
predecessors before returning. Otherwise, if multiple predecessor edges need
splitting, we only get one of them per iteration. This makes a small but
measurable compile time improvement with -enable-full-load-pre.
llvm-svn: 97521
confusing the old MAT variable with the new GlobalType one. This caused
us to promote the @disp global pointer into:
@disp.body = internal global double*** undef
instead of:
@disp.body = internal global [3 x double**] undef
llvm-svn: 97285
which branch on undef to branch on a boolean constant for the edge
exiting the loop. This helps ScalarEvolution compute trip counts for
loops.
Teach ScalarEvolution to recognize single-value PHIs, when safe, and
ForgetSymbolicName to forget such single-value PHI nodes as apprpriate
in ForgetSymbolicName.
llvm-svn: 97126
argument is non-null, pass it along to PHITranslateSubExpr so that it can
prefer using existing values that dominate the PredBB, instead of just
blindly picking the first equivalent value that it finds on a uselist.
Also when the DominatorTree is specified, have PHITranslateValue filter
out any result that does not dominate the PredBB. This is basically just
refactoring the check that used to be in GetAvailablePHITranslatedSubExpr
and also in GVN.
Despite my initial expectations, this change does not affect the results
of GVN for any testcases that I could find, but it should help compile time.
Before this change, if PHITranslateSubExpr picked a value that does not
dominate, PHITranslateWithInsertion would then insert a new value, which GVN
would later determine to be redundant and would replace. By picking a good
value to begin with, we save GVN the extra work of inserting and then
replacing a new value.
llvm-svn: 97010
induction variable value and a loop-variant value, don't force the
insert position to be at the post-increment position, because it may
not be dominated by the loop-variant value. This fixes a
use-before-def problem noticed on PPC.
llvm-svn: 96774
strides in foreign loops. This helps locate reuse opportunities
with existing induction variables in foreign loops and reduces
the need for inserting new ones. This fixes rdar://7657764.
llvm-svn: 96629
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.
llvm-svn: 96626
with multiplication by constants distributed through, occasionally
those subexpressions can include both x and -x. For now, if this
condition is discovered within LSR, just prune such cases away,
as they won't be profitable. This fixes a "zero allocated in a
base register" assertion failure.
llvm-svn: 96177
and add a doxygen comment.
Cache the phi entry to avoid doing tons of
PHINode::getBasicBlockIndex calls in the common case.
On my insane testcase from re2c, this speeds up CGP from
617.4s to 7.9s (78x).
llvm-svn: 96083
to a PHI, avoid it in the common case where the BB occurs
in the same index for multiple phis. This speeds up CGP on
an insane testcase from 8.35 to 3.58s.
llvm-svn: 96080
Functions explicitly marked inline will get an inlining threshold slightly
more aggressive than the default for -O3. This means than -O3 builds are
mostly unaffected while -Os builds will be a bit bigger and faster.
The difference depends entirely on how many 'inline's are sprinkled on the
source.
In the CINT2006 suite, only these tests are significantly affected under -Os:
Size Time
471.omnetpp +1.63% -1.85%
473.astar +4.01% -6.02%
483.xalancbmk +4.60% 0.00%
Note that 483.xalancbmk runs too quickly to give useful timing results.
llvm-svn: 96066
2. don't bother trying to merge globals in non-default sections,
doing so is quite dubious at best anyway.
3. fix a bug reported by Arnaud de Grandmaison where we'd try to
merge two globals in different address spaces.
llvm-svn: 95995
bug fixes, and with improved heuristics for analyzing foreign-loop
addrecs.
This change also flattens IVUsers, eliminating the stride-oriented
groupings, which makes it easier to work with.
llvm-svn: 95975
what it does. Enhance it to return false to optimizing vector
sign extensions from vector comparisions, which is the idiom used
to get a splatted vector for a vector comparison.
Doing this breaks vector-casts.ll, add some compensating
transformations to handle the important case they cover without
depending on this canonicalization.
This fixes rdar://7434900 a serious pessimization of vector compares.
llvm-svn: 95855
block. Other blocks may have pointer cycles that will crash
basicaa and other alias analyses. In any case, there is no
point wasting cycles optimizing dead blocks. This fixes
rdar://7635088
llvm-svn: 95852
Initial skeleton and SCEVUnknown lowering implemented,
the rest should come relatively quickly. Move testcase
to new directory.
Move pass to right before SimplifyLibCalls - which is
moved down a bit so we can take advantage of a few opts.
llvm-svn: 95628
This time it's for real! I am going to hook this up in the frontends as well.
The inliner has some experimental heuristics for dealing with the inline hint.
When given a -respect-inlinehint option, functions marked with the inline
keyword are given a threshold just above the default for -O3.
We need some experiments to determine if that is the right thing to do.
llvm-svn: 95466
xform it is checking to actually pass. There is no need to match
m_SelectCst<0, -1> since instcombine canonicalizes that into not(sext).
Add matches for sext(not(x)) in addition to not(sext(x)).
llvm-svn: 95420
short-circuited conditions to AND/OR expressions, and those expressions
are often converted back to a short-circuited form in code gen. The
original source order may have been optimized to take advantage of the
expected values, and if we reassociate them, we change the order and
subvert that optimization. Radar 7497329.
llvm-svn: 95333
This makes the inliner about as agressive as it was before my changes to the
inliner cost calculations. These levels give the same performance and slightly
smaller code than before.
llvm-svn: 95320
Fix bugs where we would compute out of bounds as in bounds, and where
we couldn't know that the linker could override the size of an array.
Add a few new testcases, change existing testcase to use a private
global array instead of extern.
llvm-svn: 95283
The SRThreshold value makes perfect sense for checking if an entire aggregate
should be promoted to a scalar integer, but it is not so good for splitting
an aggregate into its separate elements. A struct may contain a large embedded
array along with some scalar fields that would benefit from being split apart
by SROA. Even if the total aggregate size is large, it may still be good to
perform SROA. Thus, the most important piece of this patch is simply moving
the aggregate size comparison vs. SRThreshold so that it guards only the
aggregate promotion.
We have also been checking the number of elements to decide if an aggregate
should be split up. The limit of "SRThreshold/4" seemed rather arbitrary,
and I don't think it's very useful to derive this limit from SRThreshold
anyway. I've collected some data showing that the current default limit of
32 (since SRThreshold defaults to 128) is a reasonable cutoff for struct
types. One thing suggested by the data is that distinguishing between structs
and arrays might be useful. There are (obviously) a lot more large arrays
than large structs (as measured by the number of elements and not the total
size -- a large array inside a struct still counts as a single element given
the way we do SROA right now). Out of 8377 arrays where we successfully
performed SROA while compiling a large set of benchmarks, only 16 of them had
more than 8 elements. And, for those 16 arrays, it's not at all clear that
SROA was actually beneficial. So, to offset the compile time cost of
investigating more large structs for SROA, the patch lowers the limit on array
elements to 8.
This fixes Apple Radar 7563690.
llvm-svn: 95224
disabled by default. This divides the existing load PRE code into 2 phases:
first it checks that it is safe to move the load to each of the predecessors
where it is unavailable, and then if it is safe, the code is changed to move
the load. Radar 7571861.
llvm-svn: 95007
of objc message send was getting marked arm_apcscc, but the prototype
isn't. This is fine at runtime because objcmsgsend is implemented in
assembly. Only turn a mismatched caller and callee into 'unreachable'
if the callee is a definition.
llvm-svn: 94986
case, instcombine can't zap the invoke for fear of changing the CFG.
However, we have to do something to prevent the next iteration of
instcombine from inserting another store -> undef before the invoke
thereby getting into infinite iteration between dead store elim and
store insertion.
Just zap the callee to null, which will prevent the next iteration
from doing anything.
llvm-svn: 94985
This bug was exposed by my inliner cost changes in r94615, and caused failures
of lencod on most architectures when building with LTO.
This patch fixes lencod and 464.h264ref on x86-64 (and likely others).
llvm-svn: 94858
create a testcase where this matters. The select+load transformation only
occurs when isSafeToLoadUnconditionally is true, and in those situations,
instcombine also changes the underlying objects to be aligned. This seems
like a good idea regardless, and I've verified that it doesn't pessimize
the subsequent realignment.
llvm-svn: 94850
(via APInt &RHSKnownZero = KnownZero, etc) seems dangerous and confusing to me: it
is easy not to notice this, and then wonder why KnownZero/RHSKnownZero changed
underneath you when you modified RHSKnownZero/KnownZero etc. So get rid of this.
No intended functionality change (tested with "make check" + llvm-gcc bootstrap).
llvm-svn: 94802
This was already being done in SSAUpdater::GetValueAtEndOfBlock so I've
just changed SSAUpdater to check for existing PHIs in both places.
llvm-svn: 94690
Modules and ModuleProviders. Because the "ModuleProvider" simply materializes
GlobalValues now, and doesn't provide modules, it's renamed to
"GVMaterializer". Code that used to need a ModuleProvider to materialize
Functions can now materialize the Functions directly. Functions no longer use a
magic linkage to record that they're materializable; they simply ask the
GVMaterializer.
Because the C ABI must never change, we can't remove LLVMModuleProviderRef or
the functions that refer to it. Instead, because Module now exposes the same
functionality ModuleProvider used to, we store a Module* in any
LLVMModuleProviderRef and translate in the wrapper methods. The bindings to
other languages still use the ModuleProvider concept. It would probably be
worth some time to update them to follow the C++ more closely, but I don't
intend to do it.
Fixes http://llvm.org/PR5737 and http://llvm.org/PR5735.
llvm-svn: 94686
parameter with a default value, instead of just hardcoding it in the
implementation. The limit of MaxLookup = 6 was introduced in r69151 to fix
a performance problem with O(n^2) behavior in instcombine, but the scalarrepl
pass is relying on getUnderlyingObject to go all the way back to an AllocaInst.
Making the limit part of the method signature makes it clear that by default
the result is limited and should help avoid similar problems in the future.
This fixes pr6126.
llvm-svn: 94433
"sext cond" instead of a select. This simplifies some instcombine
code, matches the policy for zext (cond ? 1 : 0 -> zext), and allows
us to generate better code for a testcase on ppc.
llvm-svn: 94339
for arbitrary terminators in predecessors, don't assume
it is a conditional or uncond branch. The testcase shows
an example where they can happen with switches.
llvm-svn: 94323
handle the case when we can infer an input to the xor
from all inputs that agree, instead of going into an
infinite loop. Another part of PR6199
llvm-svn: 94321
Chris Lattner