Give SIToFPInst preference over UIToFPInst because it is faster on platforms that are widely used.

llvm-svn: 59476

new CycleBound value. Instead, just update CycleBound on each call.
Also, make ReleasePred and ReleaseSucc methods more consistent accross
the various schedulers.

This also happens to make ScheduleDAGRRList's CycleBound computation
somewhat more interesting, though it still doesn't have any noticeable
effect, because no current targets that use the register-pressure
reduction scheduler provide pipeline models.

llvm-svn: 59475

llvm-svn: 59471

llvm-svn: 59466

functions in these two schedulers.

llvm-svn: 59465

patch by Richard Osborne, tweaked and extended by your humble
servant.

llvm-svn: 59464

llvm-svn: 59461

a side-car datastructure

llvm-svn: 59458

optimizations) into this wishlist.

llvm-svn: 59455

llvm-svn: 59454

for the target. This fixes PR3080.

llvm-svn: 59450

llvm-svn: 59445

llvm-svn: 59432

This fixed PR3078.

llvm-svn: 59416

llvm-svn: 59414

llvm-svn: 59409

llvm-svn: 59406

llvm-svn: 59403

llvm-svn: 59402

llvm-svn: 59401

llvm-svn: 59400

can extract using different indexes for two vectors. Added a few tests
for vector shuffles.

llvm-svn: 59399

llvm-svn: 59398

a bit.  Next step is to factor out into their own helper functions.

llvm-svn: 59397

llvm-svn: 59395

llvm-svn: 59392

llvm-svn: 59390

folded into the select.

llvm-svn: 59389

Eliminate the cases checking for cond?0:-1, since that is already
handled by commutative checking.

llvm-svn: 59388

If this patch causes a performance regression for anyone, please let me know,
and it can be fixed in a different way with much more effort.

llvm-svn: 59384

llvm-svn: 59381

llvm-svn: 59380

well as 2 files that use "Registrator"s. These are to be used by the
MSVC builds, as the Win32 linker does not include libs that are
otherwise unreferenced, even if global constructors in the lib have
side-effects.

Patch by Scott Graham!

llvm-svn: 59378

llvm-svn: 59370

use SHUFFLE_VECTOR instead.  If not practical, fall back
to the old scheme of building the split result by hand
using a BUILD_VECTOR.

llvm-svn: 59361

and when loading/storing a widen vector, make sure that they are loaded
and stored in consecutive order.

llvm-svn: 59357

llvm-svn: 59344

llvm-svn: 59341

llvm-svn: 59340

Strip debug information before stripping symbol names. 

llvm-svn: 59328