llvm-svn: 30805

This allows us to use the multiclass for other packs.

llvm-svn: 30804

llvm-svn: 30800

llvm-svn: 30797

llvm-svn: 30796

llvm-svn: 30795

llvm-svn: 30794

llvm-svn: 30793

llvm-svn: 30792

llvm-svn: 30791

llvm-svn: 30790

llvm-svn: 30789

llvm-svn: 30788

llvm-svn: 30787

llvm-svn: 30786

llvm-svn: 30785

Introduce SS_IntUnary, a multiclass to replace SS_Int[rm].

llvm-svn: 30784

llvm-svn: 30783

llvm-svn: 30694

operand of a conditional branch to allow load folding into CMP / TEST
instructions.

llvm-svn: 30241

llvm-svn: 30112

llvm-svn: 29275

(load x+8), (load x+12), <0, 1, 2, 3> to a single 128-bit load (aligned and
unaligned).

e.g.

__m128 test(float a, float b, float c, float d) {
  return _mm_set_ps(d, c, b, a);
}

_test:
        movups 4(%esp), %xmm0
        ret

llvm-svn: 29042

llvm-svn: 28984

llvm-svn: 28979

llvm-svn: 28864

llvm-svn: 28862

llvm-svn: 28860

llvm-svn: 28796

llvm-svn: 28594

llvm-svn: 28582

llvm-svn: 28578

llvm-svn: 28577

llvm-svn: 28565

llvm-svn: 28324

llvm-svn: 28131

movlps cannot be used (e.g. when load from m64 has multiple uses).

llvm-svn: 28089

llvm-svn: 27968

result). Otherwise tblgen will pick the default (v16i8 for 128-bit vector).

llvm-svn: 27965

a temporary workaround for the 2-wide vector_shuffle problem (i.e. its mask
would have type v2i32 which is not legal).

llvm-svn: 27964