Commit Graph

8807 Commits

Author SHA1 Message Date
Chandler Carruth
464bda3a16 Teach the reassociate pass to fold chains of multiplies with repeated
elements to minimize the number of multiplies required to compute the
final result. This uses a heuristic to attempt to form near-optimal
binary exponentiation-style multiply chains. While there are some cases
it misses, it seems to at least a decent job on a very diverse range of
inputs.

Initial benchmarks show no interesting regressions, and an 8%
improvement on SPASS. Let me know if any other interesting results (in
either direction) crop up!

Credit to Richard Smith for the core algorithm, and helping code the
patch itself.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155616 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-26 05:30:30 +00:00
Jakob Stoklund Olesen
165324cd7d Print IV chain numbers while collecting them.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155567 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-25 18:01:32 +00:00
Lang Hames
87aac6a877 Reverting r155468. Chris and Chandler have convinced me that it's dangerous and
in poor taste.

Talking through some alternate solutions with Chandler.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155530 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-25 02:16:54 +00:00
Dan Gohman
50ade65982 Simplify the known retain count tracking; use a boolean state instead
of a precise count. Also, move RRInfo's Partial field into PtrState,
now that it won't increase the size.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155513 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-25 00:50:46 +00:00
Dan Gohman
eeeb7752a8 Build custom predecessor and successor lists for each basic block.
These lists exclude invoke unwind edges and loop backedges which
are being ignored. This makes it easier to ignore them
consistently.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155500 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-24 22:53:18 +00:00
Lang Hames
1d9e68dab1 Add support for llvm.arm.neon.vmull* intrinsics to InstCombine. This fixes
<rdar://problem/11291436>.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155468 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-24 18:58:36 +00:00
Jakob Stoklund Olesen
72847f3057 Reapply r155136 after fixing PR12599.
Original commit message:

Defer some shl transforms to DAGCombine.

The shl instruction is used to represent multiplication by a constant
power of two as well as bitwise left shifts. Some InstCombine
transformations would turn an shl instruction into a bit mask operation,
making it difficult for later analysis passes to recognize the
constsnt multiplication.

Disable those shl transformations, deferring them to DAGCombine time.
An 'shl X, C' instruction is now treated mostly the same was as 'mul X, C'.

These transformations are deferred:

  (X >>? C) << C   --> X & (-1 << C)  (When X >> C has multiple uses)
  (X >>? C1) << C2 --> X << (C2-C1) & (-1 << C2)   (When C2 > C1)
  (X >>? C1) << C2 --> X >>? (C1-C2) & (-1 << C2)  (When C1 > C2)

The corresponding exact transformations are preserved, just like
div-exact + mul:

  (X >>?,exact C) << C   --> X
  (X >>?,exact C1) << C2 --> X << (C2-C1)
  (X >>?,exact C1) << C2 --> X >>?,exact (C1-C2)

The disabled transformations could also prevent the instruction selector
from recognizing rotate patterns in hash functions and cryptographic
primitives. I have a test case for that, but it is too fragile.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155362 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-23 17:39:52 +00:00
Alexander Potapenko
55cabae685 Fix issue 67 by checking that the interface functions weren't redefined in the compiled source file.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155346 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-23 10:47:31 +00:00
Kostya Serebryany
2d5fdf81dc [tsan] use llvm/ADT/Statistic.h for tsan stats
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155341 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-23 08:44:59 +00:00
Jakob Stoklund Olesen
eece9dc81c Revert r155136 "Defer some shl transforms to DAGCombine."
While the patch was perfect and defect free, it exposed a really nasty
bug in X86 SelectionDAG that caused an llc crash when compiling lencod.

I'll put the patch back in after fixing the SelectionDAG problem.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155181 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-20 00:38:45 +00:00
Bill Wendling
dc21604d4a Put this expensive check below the less expensive ones.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155166 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-19 23:31:07 +00:00
Dan Gohman
8b74e5afda Avoid a bug in the path count computation, preventing an infinite
loop repeatedlt making the same change. This is for rdar://11256239.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155160 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-19 21:50:46 +00:00
Jakob Stoklund Olesen
0d5fcae6cd Defer some shl transforms to DAGCombine.
The shl instruction is used to represent multiplication by a constant
power of two as well as bitwise left shifts. Some InstCombine
transformations would turn an shl instruction into a bit mask operation,
making it difficult for later analysis passes to recognize the
constsnt multiplication.

Disable those shl transformations, deferring them to DAGCombine time.
An 'shl X, C' instruction is now treated mostly the same was as 'mul X, C'.

These transformations are deferred:

  (X >>? C) << C   --> X & (-1 << C)  (When X >> C has multiple uses)
  (X >>? C1) << C2 --> X << (C2-C1) & (-1 << C2)   (When C2 > C1)
  (X >>? C1) << C2 --> X >>? (C1-C2) & (-1 << C2)  (When C1 > C2)

The corresponding exact transformations are preserved, just like
div-exact + mul:

  (X >>?,exact C) << C   --> X
  (X >>?,exact C1) << C2 --> X << (C2-C1)
  (X >>?,exact C1) << C2 --> X >>?,exact (C1-C2)

The disabled transformations could also prevent the instruction selector
from recognizing rotate patterns in hash functions and cryptographic
primitives. I have a test case for that, but it is too fragile.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155136 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-19 16:46:26 +00:00
Dan Gohman
3b5b2a22ca Don't crash on code where the user put __attribute__((constructor)) on
a function with arguments. This fixes rdar://11265785.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155073 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-18 22:24:33 +00:00
Bill Wendling
1356041296 Use a heavy hammer to fix PR12573.
If the loop contains invoke instructions, whose unwind edge escapes the loop,
then don't try to unswitch the loop. Doing so may cause the unwind edge to be
split, which not only is non-trivial but doesn't preserve loop simplify
information.

Fixes PR12573


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154987 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-18 06:00:09 +00:00
Andrew Trick
b512263596 loop-reduce: Add an early bailout to catch extremely large loops.
This introduces a threshold of 200 IV Users, which is very
conservative but should be sufficient to avoid serious compile time
sink or stack overflow. The llvm test-suite with LTO never exceeds 190
users per loop.

The bug doesn't relate to a specific type of loop. Checking in an
arbitrary giant loop as a unit test would be silly.

Fixes rdar://11262507.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154983 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-18 04:00:10 +00:00
Joe Groff
d5bda5ec66 fix pr12559: mark unavailable win32 math libcalls
also fix SimplifyLibCalls to use TLI rather than compile-time conditionals to enable optimizations on floor, ceil, round, rint, and nearbyint

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154960 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-17 23:05:54 +00:00
Hal Finkel
05bc5087a2 Fix style violation in BBVectorize (pointed out by Bill Wendling)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154810 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-16 12:39:17 +00:00
Bill Wendling
ab3a9193b1 Add a Fixme.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154793 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-16 04:23:52 +00:00
Hal Finkel
831a6e86d5 Simplify checking for pointer types in BBVectorize (this change was suggested by Duncan).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154787 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-16 03:49:42 +00:00
Hal Finkel
bba23ed672 Fix an error in BBVectorize important for vectorizing pointer types.
When vectorizing pointer types it is important to realize that potential
pairs cannot be connected via the address pointer argument of a load or store.
This is because even after vectorization, the address is still a scalar because
the address of the higher half of the pair is implicit from the address of the
lower half (it need not be, and should not be, explicitly computed).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154735 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-14 07:32:50 +00:00
Hal Finkel
f3f5a1e6f7 Enhance BBVectorize to more-properly handle pointer values and vectorize GEPs.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154734 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-14 07:32:43 +00:00
Hal Finkel
fc3665c875 Add support to BBVectorize for vectorizing selects.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154700 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-13 20:45:45 +00:00
Dan Gohman
d6bf201fa2 Add some comments, and fix a few places that missed setting Changed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154687 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-13 18:57:48 +00:00
Dan Gohman
4423477548 Consider ObjC runtime calls objc_storeWeak and others which make a copy of
their argument as "escape" points for objc_retainBlock optimization.
This fixes rdar://11229925.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154682 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-13 18:28:58 +00:00
Hal Finkel
064551e94c By default, use Early-CSE instead of GVN for vectorization cleanup.
As has been suggested by Duncan and others, Early-CSE and GVN should
do similar redundancy elimination, but Early-CSE is much less expensive.
Most of my autovectorization benchmarks show a performance regresion, but
all of these are < 0.1%, and so I think that it is still worth using
the less expensive pass.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154673 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-13 17:15:33 +00:00
Dan Gohman
6c189ecbe6 Use the new Use-aware dominates method to apply the objc runtime
library return value optimization for phi uses. Even when the
phi itself is not dominated, the specific use may be dominated.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154647 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-13 01:08:28 +00:00
Bill Wendling
aab3c0cb07 Code-gen may inject code into the IR before it emits the ASM. The linker
obviously cannot know that this code is present, let alone used. So prevent the
internalize pass from internalizing those global values which code-gen may
insert.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154645 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-13 01:06:27 +00:00
Dan Gohman
511568dd1f Don't move objc_autorelease calls past autorelease pool boundaries when
optimizing autorelease calls on phi nodes with null operands.
This fixes rdar://11207070.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154642 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-13 00:59:57 +00:00
Chad Rosier
c9f27ee842 Typo.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154522 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-11 19:21:58 +00:00
Chandler Carruth
d6fc26217e Add two statistics to help track how we are computing the inline cost.
Yea, 'NumCallerCallersAnalyzed' isn't a great name, suggestions welcome.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154492 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-11 10:15:10 +00:00
Kostya Serebryany
cff60c1409 [tsan] two more compile-time optimizations:
- don't isntrument reads from constant globals.
Saves ~1.5% of instrumented instructions on CPU2006
(counting static instructions, not their execution).
- don't insrument reads from vtable (which is a global constant too).
Saves ~5%.

I did not measure the run-time impact of this,
but it is certainly non-negative.




git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154444 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-10 22:29:17 +00:00
Kostya Serebryany
2076af0184 [tsan] compile-time instrumentation: do not instrument a read if
a write to the same temp follows in the same BB.
Also add stats printing.

On Spec CPU2006 this optimization saves roughly 4% of instrumented reads
(which is 3% of all instrumented accesses):
Writes            : 161216
Reads             : 446458
Reads-before-write: 18295



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154418 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-10 18:18:56 +00:00
Andrew Trick
d9fc1ce809 Fix 12513: Loop unrolling breaks with indirect branches.
Take this opportunity to generalize the indirectbr bailout logic for
loop transformations. CFG transformations will never get indirectbr
right, and there's no point trying.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154386 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-10 05:14:42 +00:00
Andrew Trick
64c0748eb3 whitespace
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154385 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-10 05:14:37 +00:00
Chandler Carruth
2450eca960 Teach InstCombine to nuke a common alloca pattern -- an alloca which has
GEPs, bit casts, and stores reaching it but no other instructions. These
often show up during the iterative processing of the inliner, SROA, and
DCE. Once we hit this point, we can completely remove the alloca. These
were actually showing up in the final, fully optimized code in a bunch
of inliner tests I've been working on, and notably they show up after
LLVM finishes optimizing away all function calls involved in
hash_combine(a, b).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154285 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-08 14:36:56 +00:00
Hongbin Zheng
86312cc15f Refactor: Use positive field names in VectorizeConfig.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154249 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-07 03:56:23 +00:00
Chandler Carruth
9ceebb7e92 Sink the collection of return instructions until after *all*
simplification has been performed. This is a bit less efficient
(requires another ilist walk of the basic blocks) but shouldn't matter
in practice. More importantly, it's just too much work to keep track of
all the various ways the return instructions can be mutated while
simplifying them. This fixes yet another crasher, reported by Daniel
Dunbar.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154179 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-06 17:21:31 +00:00
Duncan Sands
a28bd85aa9 Make GVN's propagateEquality non-recursive. No intended functionality change.
The modifications are a lot more trivial than they appear to be in the diff!


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154174 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-06 15:31:09 +00:00
Chandler Carruth
6bbab86af9 Sink the return instruction collection until after we're done deleting
dead code, including dead return instructions in some cases. Otherwise,
we end up having a bogus poniter to a return instruction that blows up
much further down the road.

It turns out that this pattern is both simpler to code, easier to update
in the face of enhancements to the inliner cleanup, and likely cheaper
given that it won't add dead instructions to the list.

Thanks to John Regehr's numerous test cases for teasing this out.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154157 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-06 01:11:52 +00:00
Dan Gohman
036ebfd874 Fix accidentally inverted logic from r152803, and make the
testcase slightly less trivial. This fixes rdar://11171718.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154118 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-05 20:27:21 +00:00
Hongbin Zheng
940371bc65 BBVectorize: Add the const modifier to the VectorizeConfig because we won't
modify it.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154098 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-05 16:07:49 +00:00
Hongbin Zheng
bef377b7d7 Introduce the VectorizeConfig class, with which we can control the behavior
of the BBVectorizePass without using command line option. As pointed out
  by Hal, we can ask the TargetLoweringInfo for the architecture specific
  VectorizeConfig to perform vectorizing with architecture specific
  information.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154096 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-05 15:46:55 +00:00
Hongbin Zheng
87825e7970 Add the function "vectorizeBasicBlock" which allow users vectorize a
BasicBlock in other passes, e.g. we can call vectorizeBasicBlock in the
 loop unroll pass right after the loop is unrolled.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154089 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-05 08:05:16 +00:00
Jakob Stoklund Olesen
9243c4f7c5 Pass the right sign to TLI->isLegalICmpImmediate.
LSR can fold three addressing modes into its ICmpZero node:

  ICmpZero BaseReg + Offset      => ICmp BaseReg, -Offset
  ICmpZero -1*ScaleReg + Offset  => ICmp ScaleReg, Offset
  ICmpZero BaseReg + -1*ScaleReg => ICmp BaseReg, ScaleReg

The first two cases are only used if TLI->isLegalICmpImmediate() likes
the offset.

Make sure the right Offset sign is passed to this method in the second
case. The ARM version is not symmetric.

<rdar://problem/11184260>

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154079 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-05 03:10:56 +00:00
Rafael Espindola
26c8dcc692 Always compute all the bits in ComputeMaskedBits.
This allows us to keep passing reduced masks to SimplifyDemandedBits, but
know about all the bits if SimplifyDemandedBits fails. This allows instcombine
to simplify cases like the one in the included testcase.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154011 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-04 12:51:34 +00:00
Hongbin Zheng
00b73a5e44 LoopUnrollPass: Use variable "Threshold" instead of "CurrentThreshold" when
reducing unroll count, otherwise the reduced unroll count is not taking
  the "OptimizeForSize" attribute into account.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154007 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-04 11:44:08 +00:00
Bill Wendling
3197b4453d Add an option to turn off the expensive GVN load PRE part of GVN.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@153902 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-02 22:16:50 +00:00
Stepan Dyatkovskiy
aad9c3f17a Fast fix for PR12343:
http://llvm.org/bugs/show_bug.cgi?id=12343

We have not trivial way for splitting edges that are goes from indirect branch. We can do it with some tricks, but it should be additionally discussed. And it is still dangerous due to difficulty of indirect branches controlling.

Fix forbids this case for unswitching.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@153879 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-02 17:16:45 +00:00
Chandler Carruth
dafe48e230 Belatedly address some code review from Chris.
As a side note, I really dislike array_pod_sort... Do we really still
care about any STL implementations that get this so wrong? Does libc++?

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@153834 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-01 10:41:24 +00:00