1 ; RUN: llc -march=x86-64 -mattr=+sse4.1 -mcpu=penryn < %s | FileCheck %s -check-prefix=CHECK-W-SSE4 2 ; RUN: llc -march=x86-64 -mattr=-sse4.1 -mcpu=penryn < %s | FileCheck %s -check-prefix=CHECK-WO-SSE4 3 ; Test case for r146671 4 target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128" 5 target triple = "x86_64-apple-macosx10.7" 6 7 define <16 x i8> @shift(<16 x i8> %a, <16 x i8> %b) nounwind { 8 ; Make sure operands to pblend are in the right order. 9 ; CHECK-W-SSE4: psllw $4, [[REG1:%xmm.]] 10 ; CHECK-W-SSE4: pblendvb [[REG1]],{{ %xmm.}} 11 ; CHECK-W-SSE4: psllw $2 12 13 ; Make sure we're masking and pcmp'ing the VSELECT conditon vector. 14 ; CHECK-WO-SSE4: psllw $5, [[REG1:%xmm.]] 15 ; CHECK-WO-SSE4: pand [[REG1]], [[REG2:%xmm.]] 16 ; CHECK-WO-SSE4: pcmpeqb {{%xmm., }}[[REG2]] 17 %1 = shl <16 x i8> %a, %b 18 ret <16 x i8> %1 19 } 20
