Lines Matching refs:VPERM
444 // We can custom expand all VECTOR_SHUFFLEs to VPERM, others we can handle
614 case PPCISD::VPERM: return "PPCISD::VPERM";
5407 /// lowered into a vperm.
5438 // VPERM.
5480 // discrete instructions, or whether we should use a vperm.
5489 // Determining when to avoid vperm is tricky. Many things affect the cost
5490 // of vperm, particularly how many times the perm mask needs to be computed.
5493 // mask?) the vperm has a cost of 1. OTOH, hoisting the permute mask out of
5498 // available, if this block is within a loop, we should avoid using vperm
5504 // Lower this to a VPERM(V1, V2, V3) expression, where V3 is a constant
5508 // The SHUFFLE_VECTOR mask is almost exactly what we want for vperm, except
5524 return DAG.getNode(PPCISD::VPERM, dl, V1.getValueType(), V1, V2, VPermMask);
