llvm-svn: 122360

the folding it can.

llvm-svn: 122359

are not actually relaxed. For example, a section with only alignments will never
needs relaxation.

llvm-svn: 122356

llvm-svn: 122355

count operand.  These should be the same but apparently are
not always, and this is cleaner anyway.  This improves the
code in an existing test.

llvm-svn: 122354

llvm-svn: 122353

llvm-svn: 122349

llvm-svn: 122345

visit instructions before their uses, since InstructionSimplify does a
better job in that case.  All this prompted by Frits van Bommel.

llvm-svn: 122343

llvm-svn: 122342

not very important since the pass is only used for testing, but it does make
it more realistic.  Suggested by Frits van Bommel.

llvm-svn: 122336

type.

llvm-svn: 122334

not assume this (for example in case more transforms get added below
it).  Suggested by Frits van Bommel.

llvm-svn: 122332

llvm-svn: 122331

plenty left though!), in particular for multiplication.

llvm-svn: 122330

quite often, but don't make much difference in practice presumably because
instcombine also knows them and more.

llvm-svn: 122328

No functionality change.

llvm-svn: 122327

a couple of existing transforms.  This fires surprisingly often, for
example when compiling gcc "(X+(-1))+1->X" fires quite a lot as well
as various "and" simplifications (usually with a phi node operand).
Most of the time this doesn't make a real difference since the same
thing would have been done elsewhere anyway, eg: by instcombine, but
there are a few places where this results in simplifications that we
were not doing before.

llvm-svn: 122326

of the problems with my last attempt were in the updating of LiveIntervals
rather than the coalescing itself. Therefore, I decided to get that right first
by essentially reimplementing the existing PHIElimination using LiveIntervals.

It works correctly, with only a few tests failing (which may not be legitimate
failures) and no new verifier failures (at least as far as I can tell, I didn't
count the number per file).

llvm-svn: 122321

Type legalization splits up i64 values into pairs of i32 values, which leads
to poor quality code when inserting or extracting i64 vector elements.
If the vector element is loaded or stored, it can be treated as an f64 value
and loaded or stored directly from a VPR register.  Use the pre-legalization
DAG combiner to cast those vector elements to f64 types so that the type
legalizer won't mess them up.  Radar 8755338.

llvm-svn: 122319

section.

This helps because in practice sections form a dag with debug sections pointing
to text sections. Finishing up the text sections first makes the debug section
relaxation trivial.

llvm-svn: 122314

Fixes rdar://8782223

llvm-svn: 122313

something that just glues two nodes together, even if it is
sometimes used for flags.

llvm-svn: 122310

to be the one we want to use. bugpoint reduced testcase is a little large,
I'll see if I can simplify it down more.

Fixes part of rdar://8782207

llvm-svn: 122307

they are just as useless, but at least a bit more gramatical

llvm-svn: 122305

tPseudoInst class, its size was changed from "special" to "2 bytes". This is
incorrect because the jump table will no longer be taken into account when
calculating branch offsets.
<rdar://problem/8782216>

llvm-svn: 122303

llvm-svn: 122302

Edge bundles is an annotation on the CFG that turns it into a bipartite directed
graph where each basic block is connected to an outgoing and an ingoing bundle.
These bundles are useful for identifying regions of the CFG for live range
splitting.

llvm-svn: 122301

llvm-svn: 122296

This implementation already exists as ConnectedVNInfoEqClasses in
LiveInterval.cpp, and it seems to be generally useful to have a light-weight way
of forming equivalence classes of small integers.

IntEqClasses doesn't allow enumeration of the elements in a class.

llvm-svn: 122293

Speculatively revert the use of DenseMap in LazyValueInfo, which may be causing Linux self-host failures.

llvm-svn: 122291

llvm-svn: 122288

Revert r122114 (CallbackVH observing use-list changes) because it caused severe slowdowns on the Linux self-host configuration.

llvm-svn: 122279

Implement feedback from Bruno on making pblendvb an x86-specific ISD node in addition to being an intrinsic, and convert
lowering to use it.  Hopefully the pattern fragment is doing the right thing with XMM0, looks correct in testing.

llvm-svn: 122277

Convert one of LVI's primary maps to a DenseMap, now that we know are more assured of iterator stability.

llvm-svn: 122273

llvm-svn: 122269

llvm-svn: 122268

llvm-svn: 122265

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

llvm-svn: 122261