1 ; RUN: llc < %s -mcpu=corei7-avx -mattr=+avx | FileCheck %s 2 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" 3 target triple = "x86_64-apple-darwin11" 4 5 ; This test would create a vpand %ymm instruction that is only legal in AVX2. 6 ; CHECK-NOT: vpand %ymm 7 8 declare i32 @llvm.x86.avx.movmsk.ps.256(<8 x float>) nounwind readnone 9 10 define void @ShadeTile() nounwind { 11 allocas: 12 br i1 undef, label %if_then, label %if_else 13 14 if_then: ; preds = %allocas 15 unreachable 16 17 if_else: ; preds = %allocas 18 br i1 undef, label %for_loop156.lr.ph, label %if_exit 19 20 for_loop156.lr.ph: ; preds = %if_else 21 %val_6.i21244 = load i16, i16* undef, align 2 22 %0 = insertelement <8 x i16> undef, i16 %val_6.i21244, i32 6 23 %val_7.i21248 = load i16, i16* undef, align 2 24 %1 = insertelement <8 x i16> %0, i16 %val_7.i21248, i32 7 25 %uint2uint32.i20206 = zext <8 x i16> %1 to <8 x i32> 26 %bitop5.i20208 = and <8 x i32> %uint2uint32.i20206, <i32 31744, i32 31744, i32 31744, i32 31744, i32 31744, i32 31744, i32 31744, i32 31744> 27 %bitop8.i20209 = and <8 x i32> %uint2uint32.i20206, <i32 1023, i32 1023, i32 1023, i32 1023, i32 1023, i32 1023, i32 1023, i32 1023> 28 %bitop12.i20211 = lshr <8 x i32> %bitop5.i20208, <i32 10, i32 10, i32 10, i32 10, i32 10, i32 10, i32 10, i32 10> 29 %binop13.i20212 = add <8 x i32> %bitop12.i20211, <i32 112, i32 112, i32 112, i32 112, i32 112, i32 112, i32 112, i32 112> 30 %bitop15.i20213 = shl <8 x i32> %binop13.i20212, <i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23> 31 %bitop17.i20214 = shl <8 x i32> %bitop8.i20209, <i32 13, i32 13, i32 13, i32 13, i32 13, i32 13, i32 13, i32 13> 32 %bitop20.i20215 = or <8 x i32> undef, %bitop15.i20213 33 %bitop22.i20216 = or <8 x i32> %bitop20.i20215, %bitop17.i20214 34 %int_to_float_bitcast.i.i.i20217 = bitcast <8 x i32> %bitop22.i20216 to <8 x float> 35 %binop401 = fmul <8 x float> undef, <float 4.000000e+00, float 4.000000e+00, float 4.000000e+00, float 4.000000e+00, float 4.000000e+00, float 4.000000e+00, float 4.000000e+00, float 4.000000e+00> 36 %binop402 = fadd <8 x float> %binop401, <float -2.000000e+00, float -2.000000e+00, float -2.000000e+00, float -2.000000e+00, float -2.000000e+00, float -2.000000e+00, float -2.000000e+00, float -2.000000e+00> 37 %binop403 = fmul <8 x float> zeroinitializer, %binop402 38 %binop406 = fmul <8 x float> %int_to_float_bitcast.i.i.i20217, <float 4.000000e+00, float 4.000000e+00, float 4.000000e+00, float 4.000000e+00, float 4.000000e+00, float 4.000000e+00, float 4.000000e+00, float 4.000000e+00> 39 %binop407 = fadd <8 x float> %binop406, <float -2.000000e+00, float -2.000000e+00, float -2.000000e+00, float -2.000000e+00, float -2.000000e+00, float -2.000000e+00, float -2.000000e+00, float -2.000000e+00> 40 %binop408 = fmul <8 x float> zeroinitializer, %binop407 41 %binop411 = fsub <8 x float> <float 3.000000e+00, float 3.000000e+00, float 3.000000e+00, float 3.000000e+00, float 3.000000e+00, float 3.000000e+00, float 3.000000e+00, float 3.000000e+00>, undef 42 %val_4.i21290 = load i16, i16* undef, align 2 43 %2 = insertelement <8 x i16> undef, i16 %val_4.i21290, i32 4 44 %val_5.i21294 = load i16, i16* undef, align 2 45 %3 = insertelement <8 x i16> %2, i16 %val_5.i21294, i32 5 46 %val_6.i21298 = load i16, i16* undef, align 2 47 %4 = insertelement <8 x i16> %3, i16 %val_6.i21298, i32 6 48 %ptr_7.i21301 = inttoptr i64 undef to i16* 49 %val_7.i21302 = load i16, i16* %ptr_7.i21301, align 2 50 %5 = insertelement <8 x i16> %4, i16 %val_7.i21302, i32 7 51 %uint2uint32.i20218 = zext <8 x i16> %5 to <8 x i32> 52 %structelement561 = load i8*, i8** undef, align 8 53 %ptr2int563 = ptrtoint i8* %structelement561 to i64 54 %smear.ptr_smear7571 = insertelement <8 x i64> undef, i64 %ptr2int563, i32 7 55 %new_ptr582 = add <8 x i64> %smear.ptr_smear7571, zeroinitializer 56 %val_5.i21509 = load i8, i8* null, align 1 57 %6 = insertelement <8 x i8> undef, i8 %val_5.i21509, i32 5 58 %7 = insertelement <8 x i8> %6, i8 undef, i32 6 59 %iptr_7.i21515 = extractelement <8 x i64> %new_ptr582, i32 7 60 %ptr_7.i21516 = inttoptr i64 %iptr_7.i21515 to i8* 61 %val_7.i21517 = load i8, i8* %ptr_7.i21516, align 1 62 %8 = insertelement <8 x i8> %7, i8 %val_7.i21517, i32 7 63 %uint2float.i20245 = uitofp <8 x i8> %8 to <8 x float> 64 %binop.i20246 = fmul <8 x float> %uint2float.i20245, <float 0x3F70101020000000, float 0x3F70101020000000, float 0x3F70101020000000, float 0x3F70101020000000, float 0x3F70101020000000, float 0x3F70101020000000, float 0x3F70101020000000, float 0x3F70101020000000> 65 br i1 undef, label %for_loop594.lr.ph, label %for_exit595 66 67 if_exit: ; preds = %if_else 68 ret void 69 70 for_loop594.lr.ph: ; preds = %for_loop156.lr.ph 71 %bitop8.i20221 = and <8 x i32> %uint2uint32.i20218, <i32 1023, i32 1023, i32 1023, i32 1023, i32 1023, i32 1023, i32 1023, i32 1023> 72 br i1 undef, label %cif_test_all730, label %cif_mask_mixed1552 73 74 for_exit595: ; preds = %for_loop156.lr.ph 75 unreachable 76 77 cif_test_all730: ; preds = %for_loop594.lr.ph 78 %binop11.i20545 = fmul <8 x float> %binop408, zeroinitializer 79 %binop12.i20546 = fadd <8 x float> undef, %binop11.i20545 80 %binop15.i20547 = fmul <8 x float> %binop411, undef 81 %binop16.i20548 = fadd <8 x float> %binop12.i20546, %binop15.i20547 82 %bincmp774 = fcmp ogt <8 x float> %binop16.i20548, zeroinitializer 83 %val_to_boolvec32775 = sext <8 x i1> %bincmp774 to <8 x i32> 84 %floatmask.i20549 = bitcast <8 x i32> %val_to_boolvec32775 to <8 x float> 85 %v.i20550 = tail call i32 @llvm.x86.avx.movmsk.ps.256(<8 x float> %floatmask.i20549) nounwind readnone 86 %cond = icmp eq i32 %v.i20550, 255 87 br i1 %cond, label %cif_test_all794, label %cif_test_mixed 88 89 cif_test_all794: ; preds = %cif_test_all730 90 %binop.i20572 = fmul <8 x float> %binop403, undef 91 unreachable 92 93 cif_test_mixed: ; preds = %cif_test_all730 94 %binop1207 = fmul <8 x float> %binop.i20246, undef 95 unreachable 96 97 cif_mask_mixed1552: ; preds = %for_loop594.lr.ph 98 unreachable 99 } 100
