with more than two nodes.

llvm-svn: 52617

llvm-svn: 52616

ConstantExprs.

llvm-svn: 52615

llvm-svn: 52614

llvm-svn: 52613

llvm-svn: 52612

structures.  Its default threshold is to promote things that are
smaller than 128 bytes, which is sane.  However, it is not sane
to do this for things that turn into 128 *registers*.  Add a cap
on the number of registers introduced, defaulting to 128/4=32.

llvm-svn: 52611

fixes PR2476; patch by Richard Osborne.  The same
problem exists for a bunch of other operators, but
I'm ignoring this because they will be automagically
fixed when the new LegalizeTypes infrastructure lands,
since it already solves this problem centrally.

llvm-svn: 52610

15).

See also PR1800, which is about the signed case.

llvm-svn: 52608

llvm-svn: 52606

llvm-svn: 52605

llvm-svn: 52604

check this with an assert.

llvm-svn: 52603

llvm-svn: 52602

llvm-svn: 52601

llvm-svn: 52600

llvm-svn: 52594

llvm-svn: 52592

llvm-svn: 52590

llvm-svn: 52589

llvm-svn: 52588

llvm-svn: 52587

llvm-svn: 52585

field, which is otherwise unused after instruction selection, as an index
into the SUnit array.

llvm-svn: 52583

and provides fairly efficient removal of arbitrary elements. Switch
ScheduleDAGRRList from std::set to this new priority queue.

llvm-svn: 52582

don't know if a truncating store is possible here,
but added support for it anyway.

llvm-svn: 52577

to DenseMap<SDNode*, SUnit*>, and adjust the way cloned SUnit nodes are
handled so that only the original node needs to be in the map.
This speeds up llc on 447.dealII.llvm.bc by about 2%.

llvm-svn: 52576

llvm-svn: 52574

llvm-svn: 52572

llvm-svn: 52571

llvm-svn: 52570

Enhanced heuristic to determine the *best* register to spill. Instead of picking the register with the lowest spill weight. Consider (up to) 2 additional registers with spill weights that are close to the lowest spill weight. The one with fewest defs and uses that conflicts with the current interval (weighted by loop depth) is the spill candidate.

This is not always a win, but there are much more wins than loses and wins tend to be more noticeable.

llvm-svn: 52554

float expansion (and sometimes vector splitting too).

llvm-svn: 52548

integer of the same type.  Before it was "promotion",
but this is confusing because it is quite different
to promotion of integers.  Call it "softening" instead,
inspired by "soft float".

llvm-svn: 52546

llvm-svn: 52545

llvm-svn: 52544

getInsertedValueOperand accessors. Thanks Matthijs!

llvm-svn: 52543

instructions. Thanks to Matthijs Kooijman for pointing this out!

llvm-svn: 52542

llvm-svn: 52539

type wasn't changed.

llvm-svn: 52538