1 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
2 ; RUN: llc < %s -mtriple=i686-unknown -mattr=+sse4.1 | FileCheck %s --check-prefix=X86
3 ; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+sse4.1 | FileCheck %s --check-prefix=X64
5 define <2 x i64> @shl1(<4 x i32> %r, <4 x i32> %a) nounwind readnone ssp {
7 ; X86: # %bb.0: # %entry
8 ; X86-NEXT: pslld $23, %xmm1
9 ; X86-NEXT: paddd {{\.?LCPI[0-9]+_[0-9]+}}, %xmm1
10 ; X86-NEXT: cvttps2dq %xmm1, %xmm1
11 ; X86-NEXT: pmulld %xmm1, %xmm0
15 ; X64: # %bb.0: # %entry
16 ; X64-NEXT: pslld $23, %xmm1
17 ; X64-NEXT: paddd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1
18 ; X64-NEXT: cvttps2dq %xmm1, %xmm1
19 ; X64-NEXT: pmulld %xmm1, %xmm0
22 %shl = shl <4 x i32> %r, %a ; <<4 x i32>> [#uses=1]
23 %tmp2 = bitcast <4 x i32> %shl to <2 x i64> ; <<2 x i64>> [#uses=1]
27 define <2 x i64> @shl2(<16 x i8> %r, <16 x i8> %a) nounwind readnone ssp {
29 ; X86: # %bb.0: # %entry
30 ; X86-NEXT: movdqa %xmm1, %xmm2
31 ; X86-NEXT: movdqa %xmm0, %xmm1
32 ; X86-NEXT: psllw $5, %xmm2
33 ; X86-NEXT: movdqa %xmm0, %xmm3
34 ; X86-NEXT: psllw $4, %xmm3
35 ; X86-NEXT: pand {{\.?LCPI[0-9]+_[0-9]+}}, %xmm3
36 ; X86-NEXT: movdqa %xmm2, %xmm0
37 ; X86-NEXT: pblendvb %xmm0, %xmm3, %xmm1
38 ; X86-NEXT: movdqa %xmm1, %xmm3
39 ; X86-NEXT: psllw $2, %xmm3
40 ; X86-NEXT: pand {{\.?LCPI[0-9]+_[0-9]+}}, %xmm3
41 ; X86-NEXT: paddb %xmm2, %xmm2
42 ; X86-NEXT: movdqa %xmm2, %xmm0
43 ; X86-NEXT: pblendvb %xmm0, %xmm3, %xmm1
44 ; X86-NEXT: movdqa %xmm1, %xmm3
45 ; X86-NEXT: paddb %xmm1, %xmm3
46 ; X86-NEXT: paddb %xmm2, %xmm2
47 ; X86-NEXT: movdqa %xmm2, %xmm0
48 ; X86-NEXT: pblendvb %xmm0, %xmm3, %xmm1
49 ; X86-NEXT: movdqa %xmm1, %xmm0
53 ; X64: # %bb.0: # %entry
54 ; X64-NEXT: movdqa %xmm0, %xmm2
55 ; X64-NEXT: psllw $5, %xmm1
56 ; X64-NEXT: movdqa %xmm0, %xmm3
57 ; X64-NEXT: psllw $4, %xmm3
58 ; X64-NEXT: pand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm3
59 ; X64-NEXT: movdqa %xmm1, %xmm0
60 ; X64-NEXT: pblendvb %xmm0, %xmm3, %xmm2
61 ; X64-NEXT: movdqa %xmm2, %xmm3
62 ; X64-NEXT: psllw $2, %xmm3
63 ; X64-NEXT: pand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm3
64 ; X64-NEXT: paddb %xmm1, %xmm1
65 ; X64-NEXT: movdqa %xmm1, %xmm0
66 ; X64-NEXT: pblendvb %xmm0, %xmm3, %xmm2
67 ; X64-NEXT: movdqa %xmm2, %xmm3
68 ; X64-NEXT: paddb %xmm2, %xmm3
69 ; X64-NEXT: paddb %xmm1, %xmm1
70 ; X64-NEXT: movdqa %xmm1, %xmm0
71 ; X64-NEXT: pblendvb %xmm0, %xmm3, %xmm2
72 ; X64-NEXT: movdqa %xmm2, %xmm0
75 %shl = shl <16 x i8> %r, %a ; <<16 x i8>> [#uses=1]
76 %tmp2 = bitcast <16 x i8> %shl to <2 x i64> ; <<2 x i64>> [#uses=1]