clang and llc infinite loop when combining AVX-512 code

[davidgmbb]

This is the maximum I could reduce the IR. I might have used it wrong because I have hardly ever used it, but with llvm-reduce the IR ended up being empty.

Godbolt llc link: https://godbolt.org/z/163x34bcb
Godbolt clang link (C): https://godbolt.org/z/1nv9e7rcq (here you can observe how it compiles in debug mode)

; ModuleID = 'faulty.ll'
source_filename = "bootstrap/bloat-buster/bb.c"
target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-i128:128-f80:128-n8:16:32:64-S128"
target triple = "x86_64-pc-linux-gnu"

; Function Attrs: noreturn nounwind sspstrong uwtable
define dso_local noundef i32 @encode(ptr noalias nocapture noundef writeonly %0, ptr noalias nocapture noundef readonly %1) local_unnamed_addr #0 {
  %3 = alloca [64 x i8], align 16
  %4 = alloca [64 x i8], align 16
  %5 = alloca [64 x i8], align 16
  %6 = alloca [64 x i8], align 16
  %7 = alloca [64 x i8], align 16
  %8 = alloca [64 x i8], align 16
  %9 = alloca [64 x i8], align 16
  %10 = alloca [64 x i8], align 16
  %11 = alloca [64 x i8], align 16
  %12 = alloca [64 x i8], align 16
  %13 = alloca [64 x i8], align 16
  %14 = alloca [64 x i8], align 16
  %15 = alloca [64 x i8], align 16
  %16 = alloca [64 x i8], align 16
  %17 = alloca [64 x [16 x i8]], align 16
  %18 = alloca [64 x i8], align 16
  %19 = load <64 x i1>, ptr %1, align 8
  %20 = select <64 x i1> %19, <64 x i8> <i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16>, <64 x i8> zeroinitializer
  %21 = getelementptr i8, ptr %1, i64 8
  %22 = load <64 x i1>, ptr %21, align 8
  %23 = getelementptr i8, ptr %1, i64 16
  %24 = load <64 x i1>, ptr %23, align 8
  %25 = select <64 x i1> %24, <64 x i8> <i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13>, <64 x i8> zeroinitializer
  %26 = getelementptr i8, ptr %1, i64 24
  %27 = load <64 x i1>, ptr %26, align 8
  %28 = select <64 x i1> %27, <64 x i8> <i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46>, <64 x i8> zeroinitializer
  %29 = getelementptr i8, ptr %1, i64 32
  %30 = load <64 x i1>, ptr %29, align 8
  %31 = select <64 x i1> %30, <64 x i8> <i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54>, <64 x i8> zeroinitializer
  %32 = getelementptr i8, ptr %1, i64 40
  %33 = load <64 x i1>, ptr %32, align 8
  %34 = select <64 x i1> %33, <64 x i8> <i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62>, <64 x i8> zeroinitializer
  %35 = getelementptr i8, ptr %1, i64 48
  %36 = load <64 x i1>, ptr %35, align 8
  %37 = select <64 x i1> %36, <64 x i8> <i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38>, <64 x i8> zeroinitializer
  %38 = getelementptr i8, ptr %1, i64 56
  %39 = load <64 x i1>, ptr %38, align 8
  %40 = select <64 x i1> %39, <64 x i8> <i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100>, <64 x i8> zeroinitializer
  %41 = getelementptr i8, ptr %1, i64 64
  %42 = load <64 x i1>, ptr %41, align 8
  %43 = select <64 x i1> %42, <64 x i8> <i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101>, <64 x i8> zeroinitializer
  %44 = getelementptr i8, ptr %1, i64 72
  %45 = load <64 x i1>, ptr %44, align 8
  %46 = select <64 x i1> %45, <64 x i8> <i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102>, <64 x i8> zeroinitializer
  %47 = getelementptr i8, ptr %1, i64 80
  %48 = load <64 x i1>, ptr %47, align 8
  %49 = or <64 x i8> %46, %43
  %50 = select <64 x i1> %48, <64 x i8> <i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103>, <64 x i8> %49
  %51 = select <64 x i1> %22, <64 x i8> <i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14>, <64 x i8> %20
  %52 = or <64 x i8> %51, %25
  %53 = or <64 x i8> %52, %28
  %54 = or <64 x i8> %53, %31
  %55 = or <64 x i8> %54, %34
  %56 = or <64 x i8> %55, %37
  %57 = or <64 x i8> %56, %40
  %58 = or <64 x i8> %57, %50
  %59 = icmp ne <64 x i8> %58, zeroinitializer
  %60 = select <64 x i1> %59, <64 x i8> zeroinitializer, <64 x i8> <i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15>
  call void @llvm.lifetime.start.p0(i64 64, ptr nonnull %3) #3
  call void @llvm.lifetime.start.p0(i64 64, ptr nonnull %4) #3
  store <64 x i8> %58, ptr %3, align 16
  store <64 x i8> %60, ptr %4, align 16
  %61 = getelementptr inbounds i8, ptr %1, i64 392
  %62 = load <64 x i1>, ptr %61, align 8
  %63 = getelementptr inbounds i8, ptr %1, i64 400
  %64 = load <64 x i1>, ptr %63, align 8
  %65 = getelementptr inbounds i8, ptr %1, i64 120
  %66 = freeze <4 x i64> poison
  %67 = getelementptr inbounds i8, ptr %1, i64 104
  %68 = getelementptr inbounds i8, ptr %1, i64 88
  %69 = load i64, ptr %67, align 8
  %70 = getelementptr inbounds i8, ptr %1, i64 112
  %71 = load i64, ptr %70, align 8
  %72 = load i64, ptr %68, align 8
  %73 = getelementptr inbounds i8, ptr %1, i64 96
  %74 = load i64, ptr %73, align 8
  %75 = load <4 x i64>, ptr %65, align 1
  %76 = and <4 x i64> %75, <i64 1085102592571150095, i64 poison, i64 1085102592571150095, i64 poison>
  %77 = lshr <4 x i64> %75, <i64 4, i64 4, i64 4, i64 4>
  %78 = and <4 x i64> %77, <i64 1085102592571150095, i64 poison, i64 1085102592571150095, i64 poison>
  %79 = shufflevector <4 x i64> %76, <4 x i64> %66, <8 x i32> <i32 0, i32 poison, i32 2, i32 poison, i32 4, i32 poison, i32 6, i32 poison>
  %80 = shufflevector <4 x i64> %78, <4 x i64> %66, <8 x i32> <i32 0, i32 poison, i32 2, i32 poison, i32 4, i32 poison, i32 6, i32 poison>
  %81 = bitcast <8 x i64> %79 to <64 x i8>
  %82 = bitcast <8 x i64> %80 to <64 x i8>
  %83 = shufflevector <64 x i8> %81, <64 x i8> %82, <64 x i32> <i32 0, i32 64, i32 1, i32 65, i32 2, i32 66, i32 3, i32 67, i32 4, i32 68, i32 5, i32 69, i32 6, i32 70, i32 7, i32 71, i32 16, i32 80, i32 17, i32 81, i32 18, i32 82, i32 19, i32 83, i32 20, i32 84, i32 21, i32 85, i32 22, i32 86, i32 23, i32 87, i32 32, i32 96, i32 33, i32 97, i32 34, i32 98, i32 35, i32 99, i32 36, i32 100, i32 37, i32 101, i32 38, i32 102, i32 39, i32 103, i32 48, i32 112, i32 49, i32 113, i32 50, i32 114, i32 51, i32 115, i32 52, i32 116, i32 53, i32 117, i32 54, i32 118, i32 55, i32 119>
  %84 = getelementptr i8, ptr %1, i64 152
  %85 = load <4 x i64>, ptr %84, align 1
  %86 = and <4 x i64> %85, <i64 1085102592571150095, i64 poison, i64 1085102592571150095, i64 poison>
  %87 = lshr <4 x i64> %85, <i64 4, i64 4, i64 4, i64 4>
  %88 = and <4 x i64> %87, <i64 1085102592571150095, i64 poison, i64 1085102592571150095, i64 poison>
  %89 = shufflevector <4 x i64> %86, <4 x i64> %66, <8 x i32> <i32 0, i32 poison, i32 2, i32 poison, i32 4, i32 poison, i32 6, i32 poison>
  %90 = shufflevector <4 x i64> %88, <4 x i64> %66, <8 x i32> <i32 0, i32 poison, i32 2, i32 poison, i32 4, i32 poison, i32 6, i32 poison>
  %91 = bitcast <8 x i64> %89 to <64 x i8>
  %92 = bitcast <8 x i64> %90 to <64 x i8>
  %93 = shufflevector <64 x i8> %91, <64 x i8> %92, <64 x i32> <i32 0, i32 64, i32 1, i32 65, i32 2, i32 66, i32 3, i32 67, i32 4, i32 68, i32 5, i32 69, i32 6, i32 70, i32 7, i32 71, i32 16, i32 80, i32 17, i32 81, i32 18, i32 82, i32 19, i32 83, i32 20, i32 84, i32 21, i32 85, i32 22, i32 86, i32 23, i32 87, i32 32, i32 96, i32 33, i32 97, i32 34, i32 98, i32 35, i32 99, i32 36, i32 100, i32 37, i32 101, i32 38, i32 102, i32 39, i32 103, i32 48, i32 112, i32 49, i32 113, i32 50, i32 114, i32 51, i32 115, i32 52, i32 116, i32 53, i32 117, i32 54, i32 118, i32 55, i32 119>
  %94 = zext <64 x i1> %59 to <64 x i8>
  %95 = bitcast i64 %74 to <64 x i1>
  %96 = select <64 x i1> %95, <64 x i8> %93, <64 x i8> zeroinitializer
  %97 = bitcast i64 %72 to <64 x i1>
  %98 = select <64 x i1> %97, <64 x i8> %83, <64 x i8> %96
  %99 = bitcast <64 x i8> %98 to <8 x i64>
  %100 = or i64 %72, %74
  %101 = bitcast i64 %100 to <64 x i1>
  %102 = xor <64 x i1> %101, <i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true>
  %103 = select <64 x i1> %102, <64 x i1> <i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true>, <64 x i1> %97
  %104 = select <64 x i1> %103, <64 x i8> %83, <64 x i8> %93
  %105 = bitcast <64 x i8> %104 to <8 x i64>
  %106 = select <64 x i1> %103, <64 x i8> %93, <64 x i8> %83
  %107 = lshr <64 x i8> %104, <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>
  %108 = bitcast <64 x i8> %107 to <8 x i64>
  %109 = and <8 x i64> %108, <i64 72340172838076673, i64 72340172838076673, i64 72340172838076673, i64 72340172838076673, i64 72340172838076673, i64 72340172838076673, i64 72340172838076673, i64 72340172838076673>
  %110 = lshr <64 x i8> %106, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
  %111 = bitcast <64 x i8> %110 to <8 x i64>
  %112 = and <8 x i64> %111, <i64 289360691352306692, i64 289360691352306692, i64 289360691352306692, i64 289360691352306692, i64 289360691352306692, i64 289360691352306692, i64 289360691352306692, i64 289360691352306692>
  %113 = or disjoint <8 x i64> %109, %112
  %114 = or disjoint <8 x i64> %113, <i64 274877907008, i64 274877907008, i64 274877907008, i64 274877907008, i64 274877907008, i64 274877907008, i64 274877907008, i64 274877907008>
  %115 = bitcast <8 x i64> %113 to <64 x i8>
  %116 = icmp eq <64 x i8> %115, zeroinitializer
  %117 = select <64 x i1> %116, <64 x i8> <i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15>, <64 x i8> %94
  %118 = select <64 x i1> %59, <64 x i8> <i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2>, <64 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
  %119 = select <64 x i1> %116, <64 x i8> zeroinitializer, <64 x i8> %118
  call void @llvm.lifetime.start.p0(i64 64, ptr nonnull %5) #3
  call void @llvm.lifetime.start.p0(i64 64, ptr nonnull %6) #3
  store <8 x i64> %114, ptr %5, align 16
  store <64 x i8> %117, ptr %6, align 16
  %120 = getelementptr inbounds i8, ptr %1, i64 184
  %121 = getelementptr inbounds i8, ptr %1, i64 376
  %122 = load <2 x i64>, ptr %121, align 1
  %123 = and <2 x i64> %122, <i64 217020518514230019, i64 poison>
  %124 = lshr <2 x i64> %122, <i64 2, i64 2>
  %125 = and <2 x i64> %124, <i64 217020518514230019, i64 poison>
  %126 = lshr <2 x i64> %122, <i64 4, i64 4>
  %127 = and <2 x i64> %126, <i64 217020518514230019, i64 poison>
  %128 = lshr <2 x i64> %122, <i64 6, i64 6>
  %129 = and <2 x i64> %128, <i64 217020518514230019, i64 poison>
  %130 = freeze <2 x i64> poison
  %131 = shufflevector <2 x i64> %123, <2 x i64> %130, <4 x i32> <i32 0, i32 poison, i32 2, i32 poison>
  %132 = shufflevector <2 x i64> %125, <2 x i64> %130, <4 x i32> <i32 0, i32 poison, i32 2, i32 poison>
  %133 = bitcast <4 x i64> %131 to <32 x i8>
  %134 = bitcast <4 x i64> %132 to <32 x i8>
  %135 = shufflevector <32 x i8> %133, <32 x i8> %134, <32 x i32> <i32 0, i32 32, i32 1, i32 33, i32 2, i32 34, i32 3, i32 35, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 16, i32 48, i32 17, i32 49, i32 18, i32 50, i32 19, i32 51, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
  %136 = shufflevector <2 x i64> %127, <2 x i64> %130, <4 x i32> <i32 0, i32 poison, i32 2, i32 poison>
  %137 = shufflevector <2 x i64> %129, <2 x i64> %130, <4 x i32> <i32 0, i32 poison, i32 2, i32 poison>
  %138 = bitcast <4 x i64> %136 to <32 x i8>
  %139 = bitcast <4 x i64> %137 to <32 x i8>
  %140 = shufflevector <32 x i8> %138, <32 x i8> %139, <32 x i32> <i32 0, i32 32, i32 1, i32 33, i32 2, i32 34, i32 3, i32 35, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 16, i32 48, i32 17, i32 49, i32 18, i32 50, i32 19, i32 51, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
  %141 = bitcast <4 x i64> %66 to <32 x i8>
  %142 = shufflevector <32 x i8> %135, <32 x i8> %141, <64 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
  %143 = bitcast <4 x i64> %66 to <32 x i8>
  %144 = shufflevector <32 x i8> %140, <32 x i8> %143, <64 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
  %145 = shufflevector <64 x i8> %142, <64 x i8> %144, <64 x i32> <i32 0, i32 64, i32 1, i32 65, i32 2, i32 66, i32 3, i32 67, i32 4, i32 68, i32 5, i32 69, i32 6, i32 70, i32 7, i32 71, i32 16, i32 80, i32 17, i32 81, i32 18, i32 82, i32 19, i32 83, i32 20, i32 84, i32 21, i32 85, i32 22, i32 86, i32 23, i32 87, i32 32, i32 96, i32 33, i32 97, i32 34, i32 98, i32 35, i32 99, i32 36, i32 100, i32 37, i32 101, i32 38, i32 102, i32 39, i32 103, i32 48, i32 112, i32 49, i32 113, i32 50, i32 114, i32 51, i32 115, i32 52, i32 116, i32 53, i32 117, i32 54, i32 118, i32 55, i32 119>
  %146 = load <8 x i64>, ptr %120, align 1
  %147 = add nuw nsw <64 x i8> %119, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
  call void @llvm.lifetime.start.p0(i64 64, ptr nonnull %7) #3
  call void @llvm.lifetime.start.p0(i64 64, ptr nonnull %8) #3
  store <8 x i64> %146, ptr %7, align 16
  store <64 x i8> %119, ptr %8, align 16
  %148 = getelementptr i8, ptr %1, i64 248
  %149 = load <8 x i64>, ptr %148, align 1
  %150 = and <64 x i8> %145, <i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2>
  %151 = icmp eq <64 x i8> %150, zeroinitializer
  %152 = select <64 x i1> %151, <64 x i8> <i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15>, <64 x i8> %147
  %153 = add nuw nsw <64 x i8> %119, <i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2>
  %154 = add nuw nsw <64 x i8> %119, <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>
  call void @llvm.lifetime.start.p0(i64 64, ptr nonnull %9) #3
  call void @llvm.lifetime.start.p0(i64 64, ptr nonnull %10) #3
  store <8 x i64> %149, ptr %9, align 16
  store <64 x i8> %152, ptr %10, align 16
  %155 = getelementptr i8, ptr %1, i64 312
  %156 = load <8 x i64>, ptr %155, align 1
  %157 = icmp eq <64 x i8> %145, <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>
  %158 = select <64 x i1> %157, <64 x i8> %153, <64 x i8> <i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15>
  %159 = select <64 x i1> %157, <64 x i8> %154, <64 x i8> zeroinitializer
  call void @llvm.lifetime.start.p0(i64 64, ptr nonnull %11) #3
  call void @llvm.lifetime.start.p0(i64 64, ptr nonnull %12) #3
  store <8 x i64> %156, ptr %11, align 16
  store <64 x i8> %158, ptr %12, align 16
  %160 = or i64 %71, %69
  %161 = zext <64 x i1> %102 to <64 x i8>
  %162 = bitcast <64 x i8> %161 to <8 x i64>
  %163 = zext <64 x i1> %64 to <64 x i8>
  %164 = bitcast <64 x i8> %163 to <16 x i32>
  %165 = shl nuw nsw <16 x i32> %164, <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1>
  %166 = bitcast <16 x i32> %165 to <8 x i64>
  %167 = zext <64 x i1> %62 to <64 x i8>
  %168 = bitcast <64 x i8> %167 to <8 x i64>
  %169 = bitcast <64 x i8> %161 to <16 x i32>
  %170 = or disjoint <8 x i64> %166, %168
  %171 = or <8 x i64> %170, %162
  %172 = and <8 x i64> %105, <i64 506381209866536711, i64 506381209866536711, i64 506381209866536711, i64 506381209866536711, i64 506381209866536711, i64 506381209866536711, i64 506381209866536711, i64 506381209866536711>
  %173 = bitcast <64 x i8> %106 to <16 x i32>
  %174 = shl <16 x i32> %173, <i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3>
  %175 = bitcast <16 x i32> %174 to <8 x i64>
  %176 = and <8 x i64> %175, <i64 4051049678932293688, i64 4051049678932293688, i64 4051049678932293688, i64 4051049678932293688, i64 4051049678932293688, i64 4051049678932293688, i64 4051049678932293688, i64 4051049678932293688>
  %177 = or disjoint <8 x i64> %176, %172
  %178 = bitcast <8 x i64> %171 to <16 x i32>
  %179 = shl nuw <16 x i32> %169, <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
  %180 = shl <16 x i32> %178, <i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6>
  %181 = or <16 x i32> %179, %180
  %182 = bitcast <16 x i32> %181 to <8 x i64>
  %183 = or <8 x i64> %177, %182
  %184 = bitcast i64 %160 to <64 x i1>
  %185 = select <64 x i1> %184, <64 x i8> %159, <64 x i8> <i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15>
  %186 = add nuw nsw <64 x i8> %159, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
  %187 = select <64 x i1> %184, <64 x i8> %186, <64 x i8> zeroinitializer
  call void @llvm.lifetime.start.p0(i64 64, ptr nonnull %13) #3
  call void @llvm.lifetime.start.p0(i64 64, ptr nonnull %14) #3
  store <8 x i64> %183, ptr %13, align 16
  store <64 x i8> %185, ptr %14, align 16
  %188 = and <8 x i64> %99, <i64 506381209866536711, i64 506381209866536711, i64 506381209866536711, i64 506381209866536711, i64 506381209866536711, i64 506381209866536711, i64 506381209866536711, i64 506381209866536711>
  %189 = bitcast <8 x i64> %188 to <64 x i8>
  %190 = icmp eq <64 x i8> %189, <i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4>
  %191 = select <64 x i1> %190, <64 x i8> <i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32>, <64 x i8> zeroinitializer
  %192 = bitcast <64 x i8> %191 to <8 x i64>
  %193 = or disjoint <8 x i64> %188, %192
  %194 = select <64 x i1> %190, <64 x i8> %187, <64 x i8> <i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15>
  %195 = add nuw nsw <64 x i8> %187, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
  %196 = select <64 x i1> %190, <64 x i8> %195, <64 x i8> zeroinitializer
  call void @llvm.lifetime.start.p0(i64 64, ptr nonnull %15) #3
  call void @llvm.lifetime.start.p0(i64 64, ptr nonnull %16) #3
  store <8 x i64> %193, ptr %15, align 16
  store <64 x i8> %194, ptr %16, align 16
  call void @llvm.lifetime.start.p0(i64 1024, ptr nonnull %17) #3
  call void @llvm.lifetime.start.p0(i64 64, ptr nonnull %18) #3
  store <64 x i8> %196, ptr %18, align 16
  %197 = extractelement <64 x i8> %196, i64 0
  br label %201

198:                                              ; preds = %201
  %199 = add i8 %197, -1
  %200 = icmp ult i8 %199, 15
  br i1 %200, label %290, label %307

201:                                              ; preds = %201, %2
  %202 = phi i64 [ 0, %2 ], [ %288, %201 ]
  %203 = getelementptr [64 x i8], ptr %3, i64 0, i64 %202
  %204 = load i8, ptr %203, align 2
  %205 = getelementptr [64 x i8], ptr %4, i64 0, i64 %202
  %206 = load i8, ptr %205, align 2
  %207 = zext i8 %206 to i64
  %208 = getelementptr [64 x [16 x i8]], ptr %17, i64 0, i64 %202, i64 %207
  store i8 %204, ptr %208, align 1
  %209 = getelementptr [64 x i8], ptr %5, i64 0, i64 %202
  %210 = load i8, ptr %209, align 2
  %211 = getelementptr [64 x i8], ptr %6, i64 0, i64 %202
  %212 = load i8, ptr %211, align 2
  %213 = zext i8 %212 to i64
  %214 = getelementptr [64 x [16 x i8]], ptr %17, i64 0, i64 %202, i64 %213
  store i8 %210, ptr %214, align 1
  %215 = getelementptr [64 x i8], ptr %7, i64 0, i64 %202
  %216 = load i8, ptr %215, align 2
  %217 = getelementptr [64 x i8], ptr %8, i64 0, i64 %202
  %218 = load i8, ptr %217, align 2
  %219 = zext i8 %218 to i64
  %220 = getelementptr [64 x [16 x i8]], ptr %17, i64 0, i64 %202, i64 %219
  store i8 %216, ptr %220, align 1
  %221 = getelementptr [64 x i8], ptr %9, i64 0, i64 %202
  %222 = load i8, ptr %221, align 2
  %223 = getelementptr [64 x i8], ptr %10, i64 0, i64 %202
  %224 = load i8, ptr %223, align 2
  %225 = zext i8 %224 to i64
  %226 = getelementptr [64 x [16 x i8]], ptr %17, i64 0, i64 %202, i64 %225
  store i8 %222, ptr %226, align 1
  %227 = getelementptr [64 x i8], ptr %11, i64 0, i64 %202
  %228 = load i8, ptr %227, align 2
  %229 = getelementptr [64 x i8], ptr %12, i64 0, i64 %202
  %230 = load i8, ptr %229, align 2
  %231 = zext i8 %230 to i64
  %232 = getelementptr [64 x [16 x i8]], ptr %17, i64 0, i64 %202, i64 %231
  store i8 %228, ptr %232, align 1
  %233 = getelementptr [64 x i8], ptr %13, i64 0, i64 %202
  %234 = load i8, ptr %233, align 2
  %235 = getelementptr [64 x i8], ptr %14, i64 0, i64 %202
  %236 = load i8, ptr %235, align 2
  %237 = zext i8 %236 to i64
  %238 = getelementptr [64 x [16 x i8]], ptr %17, i64 0, i64 %202, i64 %237
  store i8 %234, ptr %238, align 1
  %239 = getelementptr [64 x i8], ptr %15, i64 0, i64 %202
  %240 = load i8, ptr %239, align 2
  %241 = getelementptr [64 x i8], ptr %16, i64 0, i64 %202
  %242 = load i8, ptr %241, align 2
  %243 = zext i8 %242 to i64
  %244 = getelementptr [64 x [16 x i8]], ptr %17, i64 0, i64 %202, i64 %243
  store i8 %240, ptr %244, align 1
  %245 = or disjoint i64 %202, 1
  %246 = getelementptr [64 x i8], ptr %3, i64 0, i64 %245
  %247 = load i8, ptr %246, align 1
  %248 = getelementptr [64 x i8], ptr %4, i64 0, i64 %245
  %249 = load i8, ptr %248, align 1
  %250 = zext i8 %249 to i64
  %251 = getelementptr [64 x [16 x i8]], ptr %17, i64 0, i64 %245, i64 %250
  store i8 %247, ptr %251, align 1
  %252 = getelementptr [64 x i8], ptr %5, i64 0, i64 %245
  %253 = load i8, ptr %252, align 1
  %254 = getelementptr [64 x i8], ptr %6, i64 0, i64 %245
  %255 = load i8, ptr %254, align 1
  %256 = zext i8 %255 to i64
  %257 = getelementptr [64 x [16 x i8]], ptr %17, i64 0, i64 %245, i64 %256
  store i8 %253, ptr %257, align 1
  %258 = getelementptr [64 x i8], ptr %7, i64 0, i64 %245
  %259 = load i8, ptr %258, align 1
  %260 = getelementptr [64 x i8], ptr %8, i64 0, i64 %245
  %261 = load i8, ptr %260, align 1
  %262 = zext i8 %261 to i64
  %263 = getelementptr [64 x [16 x i8]], ptr %17, i64 0, i64 %245, i64 %262
  store i8 %259, ptr %263, align 1
  %264 = getelementptr [64 x i8], ptr %9, i64 0, i64 %245
  %265 = load i8, ptr %264, align 1
  %266 = getelementptr [64 x i8], ptr %10, i64 0, i64 %245
  %267 = load i8, ptr %266, align 1
  %268 = zext i8 %267 to i64
  %269 = getelementptr [64 x [16 x i8]], ptr %17, i64 0, i64 %245, i64 %268
  store i8 %265, ptr %269, align 1
  %270 = getelementptr [64 x i8], ptr %11, i64 0, i64 %245
  %271 = load i8, ptr %270, align 1
  %272 = getelementptr [64 x i8], ptr %12, i64 0, i64 %245
  %273 = load i8, ptr %272, align 1
  %274 = zext i8 %273 to i64
  %275 = getelementptr [64 x [16 x i8]], ptr %17, i64 0, i64 %245, i64 %274
  store i8 %271, ptr %275, align 1
  %276 = getelementptr [64 x i8], ptr %13, i64 0, i64 %245
  %277 = load i8, ptr %276, align 1
  %278 = getelementptr [64 x i8], ptr %14, i64 0, i64 %245
  %279 = load i8, ptr %278, align 1
  %280 = zext i8 %279 to i64
  %281 = getelementptr [64 x [16 x i8]], ptr %17, i64 0, i64 %245, i64 %280
  store i8 %277, ptr %281, align 1
  %282 = getelementptr [64 x i8], ptr %15, i64 0, i64 %245
  %283 = load i8, ptr %282, align 1
  %284 = getelementptr [64 x i8], ptr %16, i64 0, i64 %245
  %285 = load i8, ptr %284, align 1
  %286 = zext i8 %285 to i64
  %287 = getelementptr [64 x [16 x i8]], ptr %17, i64 0, i64 %245, i64 %286
  store i8 %283, ptr %287, align 1
  %288 = add nuw nsw i64 %202, 2
  %289 = icmp eq i64 %288, 64
  br i1 %289, label %198, label %201, !llvm.loop !5

290:                                              ; preds = %290, %198
  %291 = phi i8 [ %304, %290 ], [ %197, %198 ]
  %292 = phi i64 [ %302, %290 ], [ 0, %198 ]
  %293 = phi i32 [ %301, %290 ], [ 0, %198 ]
  %294 = phi i32 [ %300, %290 ], [ 0, %198 ]
  %295 = zext nneg i8 %291 to i32
  %296 = zext i32 %294 to i64
  %297 = getelementptr i8, ptr %0, i64 %296
  %298 = getelementptr [64 x [16 x i8]], ptr %17, i64 0, i64 %292
  %299 = zext nneg i8 %291 to i64
  call void @llvm.memcpy.p0.p0.i64(ptr align 1 %297, ptr align 16 %298, i64 %299, i1 false)
  %300 = add i32 %294, %295
  %301 = add i32 %293, 1
  %302 = zext i32 %301 to i64
  %303 = getelementptr [64 x i8], ptr %18, i64 0, i64 %302
  %304 = load i8, ptr %303, align 1
  %305 = add i8 %304, -1
  %306 = icmp ult i8 %305, 15
  br i1 %306, label %290, label %307

307:                                              ; preds = %290, %198
  unreachable
}

; Function Attrs: nocallback nofree nosync nounwind willreturn memory(argmem: readwrite)
declare void @llvm.lifetime.start.p0(i64 immarg, ptr nocapture) #1

; Function Attrs: nocallback nofree nounwind willreturn memory(argmem: readwrite)
declare void @llvm.memcpy.p0.p0.i64(ptr noalias nocapture writeonly, ptr noalias nocapture readonly, i64, i1 immarg) #2

attributes #0 = { noreturn nounwind sspstrong uwtable "min-legal-vector-width"="512" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-cpu"="znver4" "target-features"="+64bit,+adx,+aes,+avx,+avx2,+avx512bf16,+avx512bitalg,+avx512bw,+avx512cd,+avx512dq,+avx512f,+avx512ifma,+avx512vbmi,+avx512vbmi2,+avx512vl,+avx512vnni,+avx512vpopcntdq,+bmi,+bmi2,+clflushopt,+clwb,+clzero,+cmov,+crc32,+cx16,+cx8,+evex512,+f16c,+fma,+fsgsbase,+fxsr,+gfni,+invpcid,+lzcnt,+mmx,+movbe,+mwaitx,+pclmul,+pku,+popcnt,+prfchw,+rdpid,+rdpru,+rdrnd,+rdseed,+sahf,+sha,+shstk,+sse,+sse2,+sse3,+sse4.1,+sse4.2,+sse4a,+ssse3,+vaes,+vpclmulqdq,+wbnoinvd,+x87,+xsave,+xsavec,+xsaveopt,+xsaves,-amx-bf16,-amx-complex,-amx-fp16,-amx-int8,-amx-tile,-avx10.1-256,-avx10.1-512,-avx512fp16,-avx512vp2intersect,-avxifma,-avxneconvert,-avxvnni,-avxvnniint16,-avxvnniint8,-ccmp,-cf,-cldemote,-cmpccxadd,-egpr,-enqcmd,-fma4,-hreset,-kl,-lwp,-movdir64b,-movdiri,-ndd,-pconfig,-ppx,-prefetchi,-ptwrite,-push2pop2,-raoint,-rtm,-serialize,-sgx,-sha512,-sm3,-sm4,-tbm,-tsxldtrk,-uintr,-usermsr,-waitpkg,-widekl,-xop" }
attributes #1 = { nocallback nofree nosync nounwind willreturn memory(argmem: readwrite) }
attributes #2 = { nocallback nofree nounwind willreturn memory(argmem: readwrite) }
attributes #3 = { nounwind }

!llvm.module.flags = !{!0, !1, !2, !3}
!llvm.ident = !{!4}

!0 = !{i32 1, !"wchar_size", i32 4}
!1 = !{i32 8, !"PIC Level", i32 2}
!2 = !{i32 7, !"PIE Level", i32 2}
!3 = !{i32 7, !"uwtable", i32 2}
!4 = !{!"clang version 19.1.6"}
!5 = distinct !{!5, !6}
!6 = !{!"llvm.loop.mustprogress"}

Original C function: https://github.com/birth-software/bloat-buster/blob/515617c05b21ed3b526e15dc9cebd5815543f463/bootstrap/bloat-buster/bb.c#L314

I attach the stripped log of llc -debug faulty.ll (you can see it repeats at the end)

log_faulty.log

[davidgmbb]

I have added two Godbolt links (both llc with optimized IR and clang C code), a link to the original source code and a log of the llc -debug output. Hopefully this is useful for you to find the bug. Thanks a lot!

[davidgmbb]

I have also verified with Godbolt that it used to work with Clang 18

[llvmbot]

@llvm/issue-subscribers-backend-x86

Author: David (david4r4)

This is the maximum I could reduce the IR. I might have used it wrong because I have hardly ever used it, but with llvm-reduce the IR ended up being empty.

Godbolt llc link: https://godbolt.org/z/163x34bcb
Godbolt clang link (C): https://godbolt.org/z/1nv9e7rcq (here you can observe how it compiles in debug mode)

; ModuleID = 'faulty.ll'
source_filename = "bootstrap/bloat-buster/bb.c"
target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-i128:128-f80:128-n8:16:32:64-S128"
target triple = "x86_64-pc-linux-gnu"

; Function Attrs: noreturn nounwind sspstrong uwtable
define dso_local noundef i32 @<!-- -->encode(ptr noalias nocapture noundef writeonly %0, ptr noalias nocapture noundef readonly %1) local_unnamed_addr #<!-- -->0 {
  %3 = alloca [64 x i8], align 16
  %4 = alloca [64 x i8], align 16
  %5 = alloca [64 x i8], align 16
  %6 = alloca [64 x i8], align 16
  %7 = alloca [64 x i8], align 16
  %8 = alloca [64 x i8], align 16
  %9 = alloca [64 x i8], align 16
  %10 = alloca [64 x i8], align 16
  %11 = alloca [64 x i8], align 16
  %12 = alloca [64 x i8], align 16
  %13 = alloca [64 x i8], align 16
  %14 = alloca [64 x i8], align 16
  %15 = alloca [64 x i8], align 16
  %16 = alloca [64 x i8], align 16
  %17 = alloca [64 x [16 x i8]], align 16
  %18 = alloca [64 x i8], align 16
  %19 = load &lt;64 x i1&gt;, ptr %1, align 8
  %20 = select &lt;64 x i1&gt; %19, &lt;64 x i8&gt; &lt;i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16&gt;, &lt;64 x i8&gt; zeroinitializer
  %21 = getelementptr i8, ptr %1, i64 8
  %22 = load &lt;64 x i1&gt;, ptr %21, align 8
  %23 = getelementptr i8, ptr %1, i64 16
  %24 = load &lt;64 x i1&gt;, ptr %23, align 8
  %25 = select &lt;64 x i1&gt; %24, &lt;64 x i8&gt; &lt;i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13, i8 -13&gt;, &lt;64 x i8&gt; zeroinitializer
  %26 = getelementptr i8, ptr %1, i64 24
  %27 = load &lt;64 x i1&gt;, ptr %26, align 8
  %28 = select &lt;64 x i1&gt; %27, &lt;64 x i8&gt; &lt;i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46, i8 46&gt;, &lt;64 x i8&gt; zeroinitializer
  %29 = getelementptr i8, ptr %1, i64 32
  %30 = load &lt;64 x i1&gt;, ptr %29, align 8
  %31 = select &lt;64 x i1&gt; %30, &lt;64 x i8&gt; &lt;i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54, i8 54&gt;, &lt;64 x i8&gt; zeroinitializer
  %32 = getelementptr i8, ptr %1, i64 40
  %33 = load &lt;64 x i1&gt;, ptr %32, align 8
  %34 = select &lt;64 x i1&gt; %33, &lt;64 x i8&gt; &lt;i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62, i8 62&gt;, &lt;64 x i8&gt; zeroinitializer
  %35 = getelementptr i8, ptr %1, i64 48
  %36 = load &lt;64 x i1&gt;, ptr %35, align 8
  %37 = select &lt;64 x i1&gt; %36, &lt;64 x i8&gt; &lt;i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38, i8 38&gt;, &lt;64 x i8&gt; zeroinitializer
  %38 = getelementptr i8, ptr %1, i64 56
  %39 = load &lt;64 x i1&gt;, ptr %38, align 8
  %40 = select &lt;64 x i1&gt; %39, &lt;64 x i8&gt; &lt;i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100, i8 100&gt;, &lt;64 x i8&gt; zeroinitializer
  %41 = getelementptr i8, ptr %1, i64 64
  %42 = load &lt;64 x i1&gt;, ptr %41, align 8
  %43 = select &lt;64 x i1&gt; %42, &lt;64 x i8&gt; &lt;i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101, i8 101&gt;, &lt;64 x i8&gt; zeroinitializer
  %44 = getelementptr i8, ptr %1, i64 72
  %45 = load &lt;64 x i1&gt;, ptr %44, align 8
  %46 = select &lt;64 x i1&gt; %45, &lt;64 x i8&gt; &lt;i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102, i8 102&gt;, &lt;64 x i8&gt; zeroinitializer
  %47 = getelementptr i8, ptr %1, i64 80
  %48 = load &lt;64 x i1&gt;, ptr %47, align 8
  %49 = or &lt;64 x i8&gt; %46, %43
  %50 = select &lt;64 x i1&gt; %48, &lt;64 x i8&gt; &lt;i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103, i8 103&gt;, &lt;64 x i8&gt; %49
  %51 = select &lt;64 x i1&gt; %22, &lt;64 x i8&gt; &lt;i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14, i8 -14&gt;, &lt;64 x i8&gt; %20
  %52 = or &lt;64 x i8&gt; %51, %25
  %53 = or &lt;64 x i8&gt; %52, %28
  %54 = or &lt;64 x i8&gt; %53, %31
  %55 = or &lt;64 x i8&gt; %54, %34
  %56 = or &lt;64 x i8&gt; %55, %37
  %57 = or &lt;64 x i8&gt; %56, %40
  %58 = or &lt;64 x i8&gt; %57, %50
  %59 = icmp ne &lt;64 x i8&gt; %58, zeroinitializer
  %60 = select &lt;64 x i1&gt; %59, &lt;64 x i8&gt; zeroinitializer, &lt;64 x i8&gt; &lt;i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15&gt;
  call void @<!-- -->llvm.lifetime.start.p0(i64 64, ptr nonnull %3) #<!-- -->3
  call void @<!-- -->llvm.lifetime.start.p0(i64 64, ptr nonnull %4) #<!-- -->3
  store &lt;64 x i8&gt; %58, ptr %3, align 16
  store &lt;64 x i8&gt; %60, ptr %4, align 16
  %61 = getelementptr inbounds i8, ptr %1, i64 392
  %62 = load &lt;64 x i1&gt;, ptr %61, align 8
  %63 = getelementptr inbounds i8, ptr %1, i64 400
  %64 = load &lt;64 x i1&gt;, ptr %63, align 8
  %65 = getelementptr inbounds i8, ptr %1, i64 120
  %66 = freeze &lt;4 x i64&gt; poison
  %67 = getelementptr inbounds i8, ptr %1, i64 104
  %68 = getelementptr inbounds i8, ptr %1, i64 88
  %69 = load i64, ptr %67, align 8
  %70 = getelementptr inbounds i8, ptr %1, i64 112
  %71 = load i64, ptr %70, align 8
  %72 = load i64, ptr %68, align 8
  %73 = getelementptr inbounds i8, ptr %1, i64 96
  %74 = load i64, ptr %73, align 8
  %75 = load &lt;4 x i64&gt;, ptr %65, align 1
  %76 = and &lt;4 x i64&gt; %75, &lt;i64 1085102592571150095, i64 poison, i64 1085102592571150095, i64 poison&gt;
  %77 = lshr &lt;4 x i64&gt; %75, &lt;i64 4, i64 4, i64 4, i64 4&gt;
  %78 = and &lt;4 x i64&gt; %77, &lt;i64 1085102592571150095, i64 poison, i64 1085102592571150095, i64 poison&gt;
  %79 = shufflevector &lt;4 x i64&gt; %76, &lt;4 x i64&gt; %66, &lt;8 x i32&gt; &lt;i32 0, i32 poison, i32 2, i32 poison, i32 4, i32 poison, i32 6, i32 poison&gt;
  %80 = shufflevector &lt;4 x i64&gt; %78, &lt;4 x i64&gt; %66, &lt;8 x i32&gt; &lt;i32 0, i32 poison, i32 2, i32 poison, i32 4, i32 poison, i32 6, i32 poison&gt;
  %81 = bitcast &lt;8 x i64&gt; %79 to &lt;64 x i8&gt;
  %82 = bitcast &lt;8 x i64&gt; %80 to &lt;64 x i8&gt;
  %83 = shufflevector &lt;64 x i8&gt; %81, &lt;64 x i8&gt; %82, &lt;64 x i32&gt; &lt;i32 0, i32 64, i32 1, i32 65, i32 2, i32 66, i32 3, i32 67, i32 4, i32 68, i32 5, i32 69, i32 6, i32 70, i32 7, i32 71, i32 16, i32 80, i32 17, i32 81, i32 18, i32 82, i32 19, i32 83, i32 20, i32 84, i32 21, i32 85, i32 22, i32 86, i32 23, i32 87, i32 32, i32 96, i32 33, i32 97, i32 34, i32 98, i32 35, i32 99, i32 36, i32 100, i32 37, i32 101, i32 38, i32 102, i32 39, i32 103, i32 48, i32 112, i32 49, i32 113, i32 50, i32 114, i32 51, i32 115, i32 52, i32 116, i32 53, i32 117, i32 54, i32 118, i32 55, i32 119&gt;
  %84 = getelementptr i8, ptr %1, i64 152
  %85 = load &lt;4 x i64&gt;, ptr %84, align 1
  %86 = and &lt;4 x i64&gt; %85, &lt;i64 1085102592571150095, i64 poison, i64 1085102592571150095, i64 poison&gt;
  %87 = lshr &lt;4 x i64&gt; %85, &lt;i64 4, i64 4, i64 4, i64 4&gt;
  %88 = and &lt;4 x i64&gt; %87, &lt;i64 1085102592571150095, i64 poison, i64 1085102592571150095, i64 poison&gt;
  %89 = shufflevector &lt;4 x i64&gt; %86, &lt;4 x i64&gt; %66, &lt;8 x i32&gt; &lt;i32 0, i32 poison, i32 2, i32 poison, i32 4, i32 poison, i32 6, i32 poison&gt;
  %90 = shufflevector &lt;4 x i64&gt; %88, &lt;4 x i64&gt; %66, &lt;8 x i32&gt; &lt;i32 0, i32 poison, i32 2, i32 poison, i32 4, i32 poison, i32 6, i32 poison&gt;
  %91 = bitcast &lt;8 x i64&gt; %89 to &lt;64 x i8&gt;
  %92 = bitcast &lt;8 x i64&gt; %90 to &lt;64 x i8&gt;
  %93 = shufflevector &lt;64 x i8&gt; %91, &lt;64 x i8&gt; %92, &lt;64 x i32&gt; &lt;i32 0, i32 64, i32 1, i32 65, i32 2, i32 66, i32 3, i32 67, i32 4, i32 68, i32 5, i32 69, i32 6, i32 70, i32 7, i32 71, i32 16, i32 80, i32 17, i32 81, i32 18, i32 82, i32 19, i32 83, i32 20, i32 84, i32 21, i32 85, i32 22, i32 86, i32 23, i32 87, i32 32, i32 96, i32 33, i32 97, i32 34, i32 98, i32 35, i32 99, i32 36, i32 100, i32 37, i32 101, i32 38, i32 102, i32 39, i32 103, i32 48, i32 112, i32 49, i32 113, i32 50, i32 114, i32 51, i32 115, i32 52, i32 116, i32 53, i32 117, i32 54, i32 118, i32 55, i32 119&gt;
  %94 = zext &lt;64 x i1&gt; %59 to &lt;64 x i8&gt;
  %95 = bitcast i64 %74 to &lt;64 x i1&gt;
  %96 = select &lt;64 x i1&gt; %95, &lt;64 x i8&gt; %93, &lt;64 x i8&gt; zeroinitializer
  %97 = bitcast i64 %72 to &lt;64 x i1&gt;
  %98 = select &lt;64 x i1&gt; %97, &lt;64 x i8&gt; %83, &lt;64 x i8&gt; %96
  %99 = bitcast &lt;64 x i8&gt; %98 to &lt;8 x i64&gt;
  %100 = or i64 %72, %74
  %101 = bitcast i64 %100 to &lt;64 x i1&gt;
  %102 = xor &lt;64 x i1&gt; %101, &lt;i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true&gt;
  %103 = select &lt;64 x i1&gt; %102, &lt;64 x i1&gt; &lt;i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true&gt;, &lt;64 x i1&gt; %97
  %104 = select &lt;64 x i1&gt; %103, &lt;64 x i8&gt; %83, &lt;64 x i8&gt; %93
  %105 = bitcast &lt;64 x i8&gt; %104 to &lt;8 x i64&gt;
  %106 = select &lt;64 x i1&gt; %103, &lt;64 x i8&gt; %93, &lt;64 x i8&gt; %83
  %107 = lshr &lt;64 x i8&gt; %104, &lt;i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3&gt;
  %108 = bitcast &lt;64 x i8&gt; %107 to &lt;8 x i64&gt;
  %109 = and &lt;8 x i64&gt; %108, &lt;i64 72340172838076673, i64 72340172838076673, i64 72340172838076673, i64 72340172838076673, i64 72340172838076673, i64 72340172838076673, i64 72340172838076673, i64 72340172838076673&gt;
  %110 = lshr &lt;64 x i8&gt; %106, &lt;i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1&gt;
  %111 = bitcast &lt;64 x i8&gt; %110 to &lt;8 x i64&gt;
  %112 = and &lt;8 x i64&gt; %111, &lt;i64 289360691352306692, i64 289360691352306692, i64 289360691352306692, i64 289360691352306692, i64 289360691352306692, i64 289360691352306692, i64 289360691352306692, i64 289360691352306692&gt;
  %113 = or disjoint &lt;8 x i64&gt; %109, %112
  %114 = or disjoint &lt;8 x i64&gt; %113, &lt;i64 274877907008, i64 274877907008, i64 274877907008, i64 274877907008, i64 274877907008, i64 274877907008, i64 274877907008, i64 274877907008&gt;
  %115 = bitcast &lt;8 x i64&gt; %113 to &lt;64 x i8&gt;
  %116 = icmp eq &lt;64 x i8&gt; %115, zeroinitializer
  %117 = select &lt;64 x i1&gt; %116, &lt;64 x i8&gt; &lt;i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15&gt;, &lt;64 x i8&gt; %94
  %118 = select &lt;64 x i1&gt; %59, &lt;64 x i8&gt; &lt;i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2&gt;, &lt;64 x i8&gt; &lt;i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1&gt;
  %119 = select &lt;64 x i1&gt; %116, &lt;64 x i8&gt; zeroinitializer, &lt;64 x i8&gt; %118
  call void @<!-- -->llvm.lifetime.start.p0(i64 64, ptr nonnull %5) #<!-- -->3
  call void @<!-- -->llvm.lifetime.start.p0(i64 64, ptr nonnull %6) #<!-- -->3
  store &lt;8 x i64&gt; %114, ptr %5, align 16
  store &lt;64 x i8&gt; %117, ptr %6, align 16
  %120 = getelementptr inbounds i8, ptr %1, i64 184
  %121 = getelementptr inbounds i8, ptr %1, i64 376
  %122 = load &lt;2 x i64&gt;, ptr %121, align 1
  %123 = and &lt;2 x i64&gt; %122, &lt;i64 217020518514230019, i64 poison&gt;
  %124 = lshr &lt;2 x i64&gt; %122, &lt;i64 2, i64 2&gt;
  %125 = and &lt;2 x i64&gt; %124, &lt;i64 217020518514230019, i64 poison&gt;
  %126 = lshr &lt;2 x i64&gt; %122, &lt;i64 4, i64 4&gt;
  %127 = and &lt;2 x i64&gt; %126, &lt;i64 217020518514230019, i64 poison&gt;
  %128 = lshr &lt;2 x i64&gt; %122, &lt;i64 6, i64 6&gt;
  %129 = and &lt;2 x i64&gt; %128, &lt;i64 217020518514230019, i64 poison&gt;
  %130 = freeze &lt;2 x i64&gt; poison
  %131 = shufflevector &lt;2 x i64&gt; %123, &lt;2 x i64&gt; %130, &lt;4 x i32&gt; &lt;i32 0, i32 poison, i32 2, i32 poison&gt;
  %132 = shufflevector &lt;2 x i64&gt; %125, &lt;2 x i64&gt; %130, &lt;4 x i32&gt; &lt;i32 0, i32 poison, i32 2, i32 poison&gt;
  %133 = bitcast &lt;4 x i64&gt; %131 to &lt;32 x i8&gt;
  %134 = bitcast &lt;4 x i64&gt; %132 to &lt;32 x i8&gt;
  %135 = shufflevector &lt;32 x i8&gt; %133, &lt;32 x i8&gt; %134, &lt;32 x i32&gt; &lt;i32 0, i32 32, i32 1, i32 33, i32 2, i32 34, i32 3, i32 35, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 16, i32 48, i32 17, i32 49, i32 18, i32 50, i32 19, i32 51, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison&gt;
  %136 = shufflevector &lt;2 x i64&gt; %127, &lt;2 x i64&gt; %130, &lt;4 x i32&gt; &lt;i32 0, i32 poison, i32 2, i32 poison&gt;
  %137 = shufflevector &lt;2 x i64&gt; %129, &lt;2 x i64&gt; %130, &lt;4 x i32&gt; &lt;i32 0, i32 poison, i32 2, i32 poison&gt;
  %138 = bitcast &lt;4 x i64&gt; %136 to &lt;32 x i8&gt;
  %139 = bitcast &lt;4 x i64&gt; %137 to &lt;32 x i8&gt;
  %140 = shufflevector &lt;32 x i8&gt; %138, &lt;32 x i8&gt; %139, &lt;32 x i32&gt; &lt;i32 0, i32 32, i32 1, i32 33, i32 2, i32 34, i32 3, i32 35, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 16, i32 48, i32 17, i32 49, i32 18, i32 50, i32 19, i32 51, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison&gt;
  %141 = bitcast &lt;4 x i64&gt; %66 to &lt;32 x i8&gt;
  %142 = shufflevector &lt;32 x i8&gt; %135, &lt;32 x i8&gt; %141, &lt;64 x i32&gt; &lt;i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison&gt;
  %143 = bitcast &lt;4 x i64&gt; %66 to &lt;32 x i8&gt;
  %144 = shufflevector &lt;32 x i8&gt; %140, &lt;32 x i8&gt; %143, &lt;64 x i32&gt; &lt;i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison&gt;
  %145 = shufflevector &lt;64 x i8&gt; %142, &lt;64 x i8&gt; %144, &lt;64 x i32&gt; &lt;i32 0, i32 64, i32 1, i32 65, i32 2, i32 66, i32 3, i32 67, i32 4, i32 68, i32 5, i32 69, i32 6, i32 70, i32 7, i32 71, i32 16, i32 80, i32 17, i32 81, i32 18, i32 82, i32 19, i32 83, i32 20, i32 84, i32 21, i32 85, i32 22, i32 86, i32 23, i32 87, i32 32, i32 96, i32 33, i32 97, i32 34, i32 98, i32 35, i32 99, i32 36, i32 100, i32 37, i32 101, i32 38, i32 102, i32 39, i32 103, i32 48, i32 112, i32 49, i32 113, i32 50, i32 114, i32 51, i32 115, i32 52, i32 116, i32 53, i32 117, i32 54, i32 118, i32 55, i32 119&gt;
  %146 = load &lt;8 x i64&gt;, ptr %120, align 1
  %147 = add nuw nsw &lt;64 x i8&gt; %119, &lt;i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1&gt;
  call void @<!-- -->llvm.lifetime.start.p0(i64 64, ptr nonnull %7) #<!-- -->3
  call void @<!-- -->llvm.lifetime.start.p0(i64 64, ptr nonnull %8) #<!-- -->3
  store &lt;8 x i64&gt; %146, ptr %7, align 16
  store &lt;64 x i8&gt; %119, ptr %8, align 16
  %148 = getelementptr i8, ptr %1, i64 248
  %149 = load &lt;8 x i64&gt;, ptr %148, align 1
  %150 = and &lt;64 x i8&gt; %145, &lt;i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2&gt;
  %151 = icmp eq &lt;64 x i8&gt; %150, zeroinitializer
  %152 = select &lt;64 x i1&gt; %151, &lt;64 x i8&gt; &lt;i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15&gt;, &lt;64 x i8&gt; %147
  %153 = add nuw nsw &lt;64 x i8&gt; %119, &lt;i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2&gt;
  %154 = add nuw nsw &lt;64 x i8&gt; %119, &lt;i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3&gt;
  call void @<!-- -->llvm.lifetime.start.p0(i64 64, ptr nonnull %9) #<!-- -->3
  call void @<!-- -->llvm.lifetime.start.p0(i64 64, ptr nonnull %10) #<!-- -->3
  store &lt;8 x i64&gt; %149, ptr %9, align 16
  store &lt;64 x i8&gt; %152, ptr %10, align 16
  %155 = getelementptr i8, ptr %1, i64 312
  %156 = load &lt;8 x i64&gt;, ptr %155, align 1
  %157 = icmp eq &lt;64 x i8&gt; %145, &lt;i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3&gt;
  %158 = select &lt;64 x i1&gt; %157, &lt;64 x i8&gt; %153, &lt;64 x i8&gt; &lt;i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15&gt;
  %159 = select &lt;64 x i1&gt; %157, &lt;64 x i8&gt; %154, &lt;64 x i8&gt; zeroinitializer
  call void @<!-- -->llvm.lifetime.start.p0(i64 64, ptr nonnull %11) #<!-- -->3
  call void @<!-- -->llvm.lifetime.start.p0(i64 64, ptr nonnull %12) #<!-- -->3
  store &lt;8 x i64&gt; %156, ptr %11, align 16
  store &lt;64 x i8&gt; %158, ptr %12, align 16
  %160 = or i64 %71, %69
  %161 = zext &lt;64 x i1&gt; %102 to &lt;64 x i8&gt;
  %162 = bitcast &lt;64 x i8&gt; %161 to &lt;8 x i64&gt;
  %163 = zext &lt;64 x i1&gt; %64 to &lt;64 x i8&gt;
  %164 = bitcast &lt;64 x i8&gt; %163 to &lt;16 x i32&gt;
  %165 = shl nuw nsw &lt;16 x i32&gt; %164, &lt;i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1&gt;
  %166 = bitcast &lt;16 x i32&gt; %165 to &lt;8 x i64&gt;
  %167 = zext &lt;64 x i1&gt; %62 to &lt;64 x i8&gt;
  %168 = bitcast &lt;64 x i8&gt; %167 to &lt;8 x i64&gt;
  %169 = bitcast &lt;64 x i8&gt; %161 to &lt;16 x i32&gt;
  %170 = or disjoint &lt;8 x i64&gt; %166, %168
  %171 = or &lt;8 x i64&gt; %170, %162
  %172 = and &lt;8 x i64&gt; %105, &lt;i64 506381209866536711, i64 506381209866536711, i64 506381209866536711, i64 506381209866536711, i64 506381209866536711, i64 506381209866536711, i64 506381209866536711, i64 506381209866536711&gt;
  %173 = bitcast &lt;64 x i8&gt; %106 to &lt;16 x i32&gt;
  %174 = shl &lt;16 x i32&gt; %173, &lt;i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3&gt;
  %175 = bitcast &lt;16 x i32&gt; %174 to &lt;8 x i64&gt;
  %176 = and &lt;8 x i64&gt; %175, &lt;i64 4051049678932293688, i64 4051049678932293688, i64 4051049678932293688, i64 4051049678932293688, i64 4051049678932293688, i64 4051049678932293688, i64 4051049678932293688, i64 4051049678932293688&gt;
  %177 = or disjoint &lt;8 x i64&gt; %176, %172
  %178 = bitcast &lt;8 x i64&gt; %171 to &lt;16 x i32&gt;
  %179 = shl nuw &lt;16 x i32&gt; %169, &lt;i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7&gt;
  %180 = shl &lt;16 x i32&gt; %178, &lt;i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6&gt;
  %181 = or &lt;16 x i32&gt; %179, %180
  %182 = bitcast &lt;16 x i32&gt; %181 to &lt;8 x i64&gt;
  %183 = or &lt;8 x i64&gt; %177, %182
  %184 = bitcast i64 %160 to &lt;64 x i1&gt;
  %185 = select &lt;64 x i1&gt; %184, &lt;64 x i8&gt; %159, &lt;64 x i8&gt; &lt;i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15&gt;
  %186 = add nuw nsw &lt;64 x i8&gt; %159, &lt;i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1&gt;
  %187 = select &lt;64 x i1&gt; %184, &lt;64 x i8&gt; %186, &lt;64 x i8&gt; zeroinitializer
  call void @<!-- -->llvm.lifetime.start.p0(i64 64, ptr nonnull %13) #<!-- -->3
  call void @<!-- -->llvm.lifetime.start.p0(i64 64, ptr nonnull %14) #<!-- -->3
  store &lt;8 x i64&gt; %183, ptr %13, align 16
  store &lt;64 x i8&gt; %185, ptr %14, align 16
  %188 = and &lt;8 x i64&gt; %99, &lt;i64 506381209866536711, i64 506381209866536711, i64 506381209866536711, i64 506381209866536711, i64 506381209866536711, i64 506381209866536711, i64 506381209866536711, i64 506381209866536711&gt;
  %189 = bitcast &lt;8 x i64&gt; %188 to &lt;64 x i8&gt;
  %190 = icmp eq &lt;64 x i8&gt; %189, &lt;i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4&gt;
  %191 = select &lt;64 x i1&gt; %190, &lt;64 x i8&gt; &lt;i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32&gt;, &lt;64 x i8&gt; zeroinitializer
  %192 = bitcast &lt;64 x i8&gt; %191 to &lt;8 x i64&gt;
  %193 = or disjoint &lt;8 x i64&gt; %188, %192
  %194 = select &lt;64 x i1&gt; %190, &lt;64 x i8&gt; %187, &lt;64 x i8&gt; &lt;i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15&gt;
  %195 = add nuw nsw &lt;64 x i8&gt; %187, &lt;i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1&gt;
  %196 = select &lt;64 x i1&gt; %190, &lt;64 x i8&gt; %195, &lt;64 x i8&gt; zeroinitializer
  call void @<!-- -->llvm.lifetime.start.p0(i64 64, ptr nonnull %15) #<!-- -->3
  call void @<!-- -->llvm.lifetime.start.p0(i64 64, ptr nonnull %16) #<!-- -->3
  store &lt;8 x i64&gt; %193, ptr %15, align 16
  store &lt;64 x i8&gt; %194, ptr %16, align 16
  call void @<!-- -->llvm.lifetime.start.p0(i64 1024, ptr nonnull %17) #<!-- -->3
  call void @<!-- -->llvm.lifetime.start.p0(i64 64, ptr nonnull %18) #<!-- -->3
  store &lt;64 x i8&gt; %196, ptr %18, align 16
  %197 = extractelement &lt;64 x i8&gt; %196, i64 0
  br label %201

198:                                              ; preds = %201
  %199 = add i8 %197, -1
  %200 = icmp ult i8 %199, 15
  br i1 %200, label %290, label %307

201:                                              ; preds = %201, %2
  %202 = phi i64 [ 0, %2 ], [ %288, %201 ]
  %203 = getelementptr [64 x i8], ptr %3, i64 0, i64 %202
  %204 = load i8, ptr %203, align 2
  %205 = getelementptr [64 x i8], ptr %4, i64 0, i64 %202
  %206 = load i8, ptr %205, align 2
  %207 = zext i8 %206 to i64
  %208 = getelementptr [64 x [16 x i8]], ptr %17, i64 0, i64 %202, i64 %207
  store i8 %204, ptr %208, align 1
  %209 = getelementptr [64 x i8], ptr %5, i64 0, i64 %202
  %210 = load i8, ptr %209, align 2
  %211 = getelementptr [64 x i8], ptr %6, i64 0, i64 %202
  %212 = load i8, ptr %211, align 2
  %213 = zext i8 %212 to i64
  %214 = getelementptr [64 x [16 x i8]], ptr %17, i64 0, i64 %202, i64 %213
  store i8 %210, ptr %214, align 1
  %215 = getelementptr [64 x i8], ptr %7, i64 0, i64 %202
  %216 = load i8, ptr %215, align 2
  %217 = getelementptr [64 x i8], ptr %8, i64 0, i64 %202
  %218 = load i8, ptr %217, align 2
  %219 = zext i8 %218 to i64
  %220 = getelementptr [64 x [16 x i8]], ptr %17, i64 0, i64 %202, i64 %219
  store i8 %216, ptr %220, align 1
  %221 = getelementptr [64 x i8], ptr %9, i64 0, i64 %202
  %222 = load i8, ptr %221, align 2
  %223 = getelementptr [64 x i8], ptr %10, i64 0, i64 %202
  %224 = load i8, ptr %223, align 2
  %225 = zext i8 %224 to i64
  %226 = getelementptr [64 x [16 x i8]], ptr %17, i64 0, i64 %202, i64 %225
  store i8 %222, ptr %226, align 1
  %227 = getelementptr [64 x i8], ptr %11, i64 0, i64 %202
  %228 = load i8, ptr %227, align 2
  %229 = getelementptr [64 x i8], ptr %12, i64 0, i64 %202
  %230 = load i8, ptr %229, align 2
  %231 = zext i8 %230 to i64
  %232 = getelementptr [64 x [16 x i8]], ptr %17, i64 0, i64 %202, i64 %231
  store i8 %228, ptr %232, align 1
  %233 = getelementptr [64 x i8], ptr %13, i64 0, i64 %202
  %234 = load i8, ptr %233, align 2
  %235 = getelementptr [64 x i8], ptr %14, i64 0, i64 %202
  %236 = load i8, ptr %235, align 2
  %237 = zext i8 %236 to i64
  %238 = getelementptr [64 x [16 x i8]], ptr %17, i64 0, i64 %202, i64 %237
  store i8 %234, ptr %238, align 1
  %239 = getelementptr [64 x i8], ptr %15, i64 0, i64 %202
  %240 = load i8, ptr %239, align 2
  %241 = getelementptr [64 x i8], ptr %16, i64 0, i64 %202
  %242 = load i8, ptr %241, align 2
  %243 = zext i8 %242 to i64
  %244 = getelementptr [64 x [16 x i8]], ptr %17, i64 0, i64 %202, i64 %243
  store i8 %240, ptr %244, align 1
  %245 = or disjoint i64 %202, 1
  %246 = getelementptr [64 x i8], ptr %3, i64 0, i64 %245
  %247 = load i8, ptr %246, align 1
  %248 = getelementptr [64 x i8], ptr %4, i64 0, i64 %245
  %249 = load i8, ptr %248, align 1
  %250 = zext i8 %249 to i64
  %251 = getelementptr [64 x [16 x i8]], ptr %17, i64 0, i64 %245, i64 %250
  store i8 %247, ptr %251, align 1
  %252 = getelementptr [64 x i8], ptr %5, i64 0, i64 %245
  %253 = load i8, ptr %252, align 1
  %254 = getelementptr [64 x i8], ptr %6, i64 0, i64 %245
  %255 = load i8, ptr %254, align 1
  %256 = zext i8 %255 to i64
  %257 = getelementptr [64 x [16 x i8]], ptr %17, i64 0, i64 %245, i64 %256
  store i8 %253, ptr %257, align 1
  %258 = getelementptr [64 x i8], ptr %7, i64 0, i64 %245
  %259 = load i8, ptr %258, align 1
  %260 = getelementptr [64 x i8], ptr %8, i64 0, i64 %245
  %261 = load i8, ptr %260, align 1
  %262 = zext i8 %261 to i64
  %263 = getelementptr [64 x [16 x i8]], ptr %17, i64 0, i64 %245, i64 %262
  store i8 %259, ptr %263, align 1
  %264 = getelementptr [64 x i8], ptr %9, i64 0, i64 %245
  %265 = load i8, ptr %264, align 1
  %266 = getelementptr [64 x i8], ptr %10, i64 0, i64 %245
  %267 = load i8, ptr %266, align 1
  %268 = zext i8 %267 to i64
  %269 = getelementptr [64 x [16 x i8]], ptr %17, i64 0, i64 %245, i64 %268
  store i8 %265, ptr %269, align 1
  %270 = getelementptr [64 x i8], ptr %11, i64 0, i64 %245
  %271 = load i8, ptr %270, align 1
  %272 = getelementptr [64 x i8], ptr %12, i64 0, i64 %245
  %273 = load i8, ptr %272, align 1
  %274 = zext i8 %273 to i64
  %275 = getelementptr [64 x [16 x i8]], ptr %17, i64 0, i64 %245, i64 %274
  store i8 %271, ptr %275, align 1
  %276 = getelementptr [64 x i8], ptr %13, i64 0, i64 %245
  %277 = load i8, ptr %276, align 1
  %278 = getelementptr [64 x i8], ptr %14, i64 0, i64 %245
  %279 = load i8, ptr %278, align 1
  %280 = zext i8 %279 to i64
  %281 = getelementptr [64 x [16 x i8]], ptr %17, i64 0, i64 %245, i64 %280
  store i8 %277, ptr %281, align 1
  %282 = getelementptr [64 x i8], ptr %15, i64 0, i64 %245
  %283 = load i8, ptr %282, align 1
  %284 = getelementptr [64 x i8], ptr %16, i64 0, i64 %245
  %285 = load i8, ptr %284, align 1
  %286 = zext i8 %285 to i64
  %287 = getelementptr [64 x [16 x i8]], ptr %17, i64 0, i64 %245, i64 %286
  store i8 %283, ptr %287, align 1
  %288 = add nuw nsw i64 %202, 2
  %289 = icmp eq i64 %288, 64
  br i1 %289, label %198, label %201, !llvm.loop !5

290:                                              ; preds = %290, %198
  %291 = phi i8 [ %304, %290 ], [ %197, %198 ]
  %292 = phi i64 [ %302, %290 ], [ 0, %198 ]
  %293 = phi i32 [ %301, %290 ], [ 0, %198 ]
  %294 = phi i32 [ %300, %290 ], [ 0, %198 ]
  %295 = zext nneg i8 %291 to i32
  %296 = zext i32 %294 to i64
  %297 = getelementptr i8, ptr %0, i64 %296
  %298 = getelementptr [64 x [16 x i8]], ptr %17, i64 0, i64 %292
  %299 = zext nneg i8 %291 to i64
  call void @<!-- -->llvm.memcpy.p0.p0.i64(ptr align 1 %297, ptr align 16 %298, i64 %299, i1 false)
  %300 = add i32 %294, %295
  %301 = add i32 %293, 1
  %302 = zext i32 %301 to i64
  %303 = getelementptr [64 x i8], ptr %18, i64 0, i64 %302
  %304 = load i8, ptr %303, align 1
  %305 = add i8 %304, -1
  %306 = icmp ult i8 %305, 15
  br i1 %306, label %290, label %307

307:                                              ; preds = %290, %198
  unreachable
}

; Function Attrs: nocallback nofree nosync nounwind willreturn memory(argmem: readwrite)
declare void @<!-- -->llvm.lifetime.start.p0(i64 immarg, ptr nocapture) #<!-- -->1

; Function Attrs: nocallback nofree nounwind willreturn memory(argmem: readwrite)
declare void @<!-- -->llvm.memcpy.p0.p0.i64(ptr noalias nocapture writeonly, ptr noalias nocapture readonly, i64, i1 immarg) #<!-- -->2

attributes #<!-- -->0 = { noreturn nounwind sspstrong uwtable "min-legal-vector-width"="512" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-cpu"="znver4" "target-features"="+64bit,+adx,+aes,+avx,+avx2,+avx512bf16,+avx512bitalg,+avx512bw,+avx512cd,+avx512dq,+avx512f,+avx512ifma,+avx512vbmi,+avx512vbmi2,+avx512vl,+avx512vnni,+avx512vpopcntdq,+bmi,+bmi2,+clflushopt,+clwb,+clzero,+cmov,+crc32,+cx16,+cx8,+evex512,+f16c,+fma,+fsgsbase,+fxsr,+gfni,+invpcid,+lzcnt,+mmx,+movbe,+mwaitx,+pclmul,+pku,+popcnt,+prfchw,+rdpid,+rdpru,+rdrnd,+rdseed,+sahf,+sha,+shstk,+sse,+sse2,+sse3,+sse4.1,+sse4.2,+sse4a,+ssse3,+vaes,+vpclmulqdq,+wbnoinvd,+x87,+xsave,+xsavec,+xsaveopt,+xsaves,-amx-bf16,-amx-complex,-amx-fp16,-amx-int8,-amx-tile,-avx10.1-256,-avx10.1-512,-avx512fp16,-avx512vp2intersect,-avxifma,-avxneconvert,-avxvnni,-avxvnniint16,-avxvnniint8,-ccmp,-cf,-cldemote,-cmpccxadd,-egpr,-enqcmd,-fma4,-hreset,-kl,-lwp,-movdir64b,-movdiri,-ndd,-pconfig,-ppx,-prefetchi,-ptwrite,-push2pop2,-raoint,-rtm,-serialize,-sgx,-sha512,-sm3,-sm4,-tbm,-tsxldtrk,-uintr,-usermsr,-waitpkg,-widekl,-xop" }
attributes #<!-- -->1 = { nocallback nofree nosync nounwind willreturn memory(argmem: readwrite) }
attributes #<!-- -->2 = { nocallback nofree nounwind willreturn memory(argmem: readwrite) }
attributes #<!-- -->3 = { nounwind }

!llvm.module.flags = !{!0, !1, !2, !3}
!llvm.ident = !{!4}

!0 = !{i32 1, !"wchar_size", i32 4}
!1 = !{i32 8, !"PIC Level", i32 2}
!2 = !{i32 7, !"PIE Level", i32 2}
!3 = !{i32 7, !"uwtable", i32 2}
!4 = !{!"clang version 19.1.6"}
!5 = distinct !{!5, !6}
!6 = !{!"llvm.loop.mustprogress"}

Original C function: https://github.com/birth-software/bloat-buster/blob/515617c05b21ed3b526e15dc9cebd5815543f463/bootstrap/bloat-buster/bb.c#L314

I attach the stripped log of llc -debug faulty.ll (you can see it repeats at the end)

log_faulty.log

[abhishek-kaushik22]

Guilty commit seems to be 89172e9.

[RKSimon]

Guilty commit seems to be 89172e9.

OK - will take a look

[abhishek-kaushik22]

@RKSimon Thanks.

Here's what I found, it's this combine in DAGCombiner::visitOR

// Simplify: (or (op x...), (op y...))  -> (op (or x, y))
  if (N0.getOpcode() == N1.getOpcode())
    if (SDValue V = hoistLogicOpWithSameOpcodeHands(N))
      return V;

and this part that is causing the infinite loop

// Does the inner bitcast already exist?
  if (Depth > 0)
    if (SDNode *Alt = DAG.getNodeIfExists(ISD::BITCAST, DAG.getVTList(VT), {V}))
      return SDValue(Alt, 0);

[RKSimon]

Nice catch - hoistLogicOpWithSameOpcodeHands has a history of bad oneuse checking iirc