new-ir.ll

; ModuleID = 'bench1.cgu-0.rs'
source_filename = "bench1.cgu-0.rs"
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
target triple = "x86_64-unknown-linux-gnu"

%str_slice = type { i8*, i64 }
%"12.test::Bencher" = type { i64, %"1.std::time::Duration", i64 }
%"1.std::time::Duration" = type { i64, i32 }
%"15.ndarray::ArrayBase<ndarray::ViewRepr<&i32>, (usize, usize)>" = type { %"15.ndarray::ViewRepr<&i32>", i32*, { i64, i64 }, { i64, i64 } }
%"15.ndarray::ViewRepr<&i32>" = type { %"2.std::marker::PhantomData<&i32>" }
%"2.std::marker::PhantomData<&i32>" = type {}
%"1.std::time::Instant" = type { %"1.std::sys::time::inner::Instant" }
%"1.std::sys::time::inner::Instant" = type { %"1.std::sys::time::Timespec" }
%"1.std::sys::time::Timespec" = type { %"7.libc::unix::timespec" }
%"7.libc::unix::timespec" = type { i64, i64 }
%"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>" = type { %"3.std::vec::Vec<i32>", i32*, { i64, i64 }, { i64, i64 } }
%"3.std::vec::Vec<i32>" = type { %"5.alloc::raw_vec::RawVec<i32>", i64 }
%"5.alloc::raw_vec::RawVec<i32>" = type { %"2.std::ptr::Unique<i32>", i64 }
%"2.std::ptr::Unique<i32>" = type { %"2.core::nonzero::NonZero<*const i32>", %"2.std::marker::PhantomData<i32>" }
%"2.core::nonzero::NonZero<*const i32>" = type { i32* }
%"2.std::marker::PhantomData<i32>" = type {}
%"12.test::TestDescAndFn" = type { %"12.test::TestDesc", %"12.test::TestFn" }
%"12.test::TestDesc" = type { %"12.test::TestName", i8, %"12.test::ShouldPanic" }
%"12.test::TestName" = type { i64, [0 x i64], [3 x i64] }
%"12.test::ShouldPanic" = type { i64, [0 x i64], [2 x i64] }
%"12.test::TestFn" = type { i64, [0 x i64], [2 x i64] }
%"8.unwind::libunwind::_Unwind_Exception" = type { i64, void (i32, %"8.unwind::libunwind::_Unwind_Exception"*)*, [6 x i64] }
%"8.unwind::libunwind::_Unwind_Context" = type {}

@str10697 = internal constant [29 x i8] c"iter_sum_2d_transpose_regular"
@ref10698 = internal unnamed_addr constant { { { { i64, %str_slice, [8 x i8] }, i1, [7 x i8], { i64, [16 x i8] } }, { i64, void (%"12.test::Bencher"*)*, [8 x i8] } } } { { { { i64, %str_slice, [8 x i8] }, i1, [7 x i8], { i64, [16 x i8] } }, { i64, void (%"12.test::Bencher"*)*, [8 x i8] } } { { { i64, %str_slice, [8 x i8] }, i1, [7 x i8], { i64, [16 x i8] } } { { i64, %str_slice, [8 x i8] } { i64 0, %str_slice { i8* getelementptr inbounds ([29 x i8], [29 x i8]* @str10697, i32 0, i32 0), i64 29 }, [8 x i8] undef }, i1 false, [7 x i8] undef, { i64, [16 x i8] } { i64 0, [16 x i8] undef } }, { i64, void (%"12.test::Bencher"*)*, [8 x i8] } { i64 1, void (%"12.test::Bencher"*)* @_ZN6bench129iter_sum_2d_transpose_regular17h06748f26b74ef457E, [8 x i8] undef } } }, align 8

; Function Attrs: uwtable
define internal void @_ZN6bench129iter_sum_2d_transpose_regular17h06748f26b74ef457E(%"12.test::Bencher"* nocapture dereferenceable(32)) unnamed_addr #0 personality i32 (i32, i32, i64, %"8.unwind::libunwind::_Unwind_Exception"*, %"8.unwind::libunwind::_Unwind_Context"*)* @rust_eh_personality {
bb6.i.i.i.i.i:
  %defaults.i.i.i.i.i.i = alloca { i64, i64 }, align 8
  %self.i.i.i.i = alloca %"15.ndarray::ArrayBase<ndarray::ViewRepr<&i32>, (usize, usize)>", align 8
  %dummy.i.i = alloca i32, align 4
  %start1.i = alloca %"1.std::time::Instant", align 8
  %tmp_ret2.i = alloca %"1.std::time::Duration", align 8
  %dummy.i = alloca %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>", align 8
  %1 = tail call i8* @__rust_allocate(i64 16384, i64 4) #2, !noalias !1
  %2 = icmp eq i8* %1, null
  br i1 %2, label %bb10.i.i.i.i.i, label %bb4

bb10.i.i.i.i.i:                                   ; preds = %bb6.i.i.i.i.i
  invoke void @_ZN5alloc3oom3oom17hb4382286715c86b5E()
          to label %.noexc20 unwind label %cleanup

.noexc20:                                         ; preds = %bb10.i.i.i.i.i
  unreachable

bb1:                                              ; preds = %bb6.i.i.i.i.i29, %bb9, %cleanup
  %personalityslot.sroa.5.0 = phi i32 [ %.fca.1.extract, %cleanup ], [ %.fca.1.extract7, %bb9 ], [ %.fca.1.extract7, %bb6.i.i.i.i.i29 ]
  %personalityslot.sroa.0.0 = phi i8* [ %.fca.0.extract4, %cleanup ], [ %.fca.0.extract5, %bb9 ], [ %.fca.0.extract5, %bb6.i.i.i.i.i29 ]
  %.fca.0.insert = insertvalue { i8*, i32 } undef, i8* %personalityslot.sroa.0.0, 0
  %.fca.1.insert = insertvalue { i8*, i32 } %.fca.0.insert, i32 %personalityslot.sroa.5.0, 1
  resume { i8*, i32 } %.fca.1.insert

bb4:                                              ; preds = %bb6.i.i.i.i.i
  %3 = ptrtoint i8* %1 to i64
  %4 = bitcast %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>"* %dummy.i to i8*
  call void @llvm.memset.p0i8.i64(i8* nonnull %1, i8 0, i64 16384, i32 4, i1 false)
  call void @llvm.lifetime.start(i64 64, i8* %4) #2, !noalias !12
  %_9.sroa.0.0..sroa_cast = bitcast %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>"* %dummy.i to i64*
  store i64 %3, i64* %_9.sroa.0.0..sroa_cast, align 8
  %_9.sroa.4.0..sroa_idx66 = getelementptr inbounds %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>", %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>"* %dummy.i, i64 0, i32 0, i32 0, i32 1
  store i64 4096, i64* %_9.sroa.4.0..sroa_idx66, align 8
  %_9.sroa.5.0..sroa_idx68 = getelementptr inbounds %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>", %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>"* %dummy.i, i64 0, i32 0, i32 1
  store i64 4096, i64* %_9.sroa.5.0..sroa_idx68, align 8
  %_9.sroa.6.0..sroa_idx70 = getelementptr inbounds %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>", %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>"* %dummy.i, i64 0, i32 1
  %_9.sroa.6.0..sroa_cast = bitcast i32** %_9.sroa.6.0..sroa_idx70 to i64*
  store i64 %3, i64* %_9.sroa.6.0..sroa_cast, align 8
  %_9.sroa.7.0..sroa_idx72 = getelementptr inbounds %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>", %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>"* %dummy.i, i64 0, i32 2, i32 0
  %_9.sroa.8.0..sroa_idx74 = getelementptr inbounds %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>", %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>"* %dummy.i, i64 0, i32 2, i32 1
  %_9.sroa.9.0..sroa_idx76 = getelementptr inbounds %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>", %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>"* %dummy.i, i64 0, i32 3, i32 0
  %_9.sroa.10.0..sroa_idx78 = getelementptr inbounds %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>", %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>"* %dummy.i, i64 0, i32 3, i32 1
  %5 = bitcast i64* %_9.sroa.7.0..sroa_idx72 to <4 x i64>*
  store <4 x i64> <i64 64, i64 64, i64 1, i64 64>, <4 x i64>* %5, align 8
  call void asm "", "r,~{dirflag},~{fpsr},~{flags}"(%"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>"* nonnull %dummy.i) #2, !noalias !12, !srcloc !16
  %a1.sroa.0.0..sroa_cast = bitcast %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>"* %dummy.i to i8**
  %a1.sroa.0.0.copyload = load i8*, i8** %a1.sroa.0.0..sroa_cast, align 8
  %a1.sroa.5.0.copyload = load i64, i64* %_9.sroa.4.0..sroa_idx66, align 8
  %a1.sroa.739.0.copyload = load i64, i64* %_9.sroa.6.0..sroa_cast, align 8
  %a1.sroa.8.0.copyload = load i64, i64* %_9.sroa.7.0..sroa_idx72, align 8
  %a1.sroa.9.0.copyload = load i64, i64* %_9.sroa.8.0..sroa_idx74, align 8
  %a1.sroa.10.0.copyload = load i64, i64* %_9.sroa.9.0..sroa_idx76, align 8
  %a1.sroa.11.0.copyload = load i64, i64* %_9.sroa.10.0..sroa_idx78, align 8
  call void @llvm.lifetime.end(i64 64, i8* %4) #2, !noalias !12
  %6 = bitcast %"1.std::time::Duration"* %tmp_ret2.i to i8*
  call void @llvm.lifetime.start(i64 16, i8* %6)
  %7 = bitcast %"1.std::time::Instant"* %start1.i to i8*
  call void @llvm.lifetime.start(i64 16, i8* %7)
  invoke void @_ZN3std4time7Instant3now17h92a4d74662f63e98E(%"1.std::time::Instant"* noalias nocapture nonnull sret dereferenceable(16) %start1.i)
          to label %.noexc22 unwind label %bb9

.noexc22:                                         ; preds = %bb4
  %8 = getelementptr inbounds %"12.test::Bencher", %"12.test::Bencher"* %0, i64 0, i32 0
  %9 = load i64, i64* %8, align 8
  %10 = icmp eq i64 %9, 0
  br i1 %10, label %bb8.i, label %bb7.lr.ph.i

bb7.lr.ph.i:                                      ; preds = %.noexc22
  %11 = bitcast %"15.ndarray::ArrayBase<ndarray::ViewRepr<&i32>, (usize, usize)>"* %self.i.i.i.i to i8*
  %_11.sroa.0.0..sroa_cast.i.i.i = bitcast %"15.ndarray::ArrayBase<ndarray::ViewRepr<&i32>, (usize, usize)>"* %self.i.i.i.i to i64*
  %_11.sroa.4.0..sroa_idx8.i.i.i = getelementptr inbounds %"15.ndarray::ArrayBase<ndarray::ViewRepr<&i32>, (usize, usize)>", %"15.ndarray::ArrayBase<ndarray::ViewRepr<&i32>, (usize, usize)>"* %self.i.i.i.i, i64 0, i32 2, i32 0
  %_11.sroa.5.0..sroa_idx10.i.i.i = getelementptr inbounds %"15.ndarray::ArrayBase<ndarray::ViewRepr<&i32>, (usize, usize)>", %"15.ndarray::ArrayBase<ndarray::ViewRepr<&i32>, (usize, usize)>"* %self.i.i.i.i, i64 0, i32 2, i32 1
  %_11.sroa.6.0..sroa_idx12.i.i.i = getelementptr inbounds %"15.ndarray::ArrayBase<ndarray::ViewRepr<&i32>, (usize, usize)>", %"15.ndarray::ArrayBase<ndarray::ViewRepr<&i32>, (usize, usize)>"* %self.i.i.i.i, i64 0, i32 3, i32 0
  %_11.sroa.7.0..sroa_idx14.i.i.i = getelementptr inbounds %"15.ndarray::ArrayBase<ndarray::ViewRepr<&i32>, (usize, usize)>", %"15.ndarray::ArrayBase<ndarray::ViewRepr<&i32>, (usize, usize)>"* %self.i.i.i.i, i64 0, i32 3, i32 1
  %12 = bitcast { i64, i64 }* %defaults.i.i.i.i.i.i to i8*
  %13 = getelementptr inbounds { i64, i64 }, { i64, i64 }* %defaults.i.i.i.i.i.i, i64 0, i32 0
  %14 = getelementptr inbounds { i64, i64 }, { i64, i64 }* %defaults.i.i.i.i.i.i, i64 0, i32 1
  %15 = bitcast i32* %dummy.i.i to i8*
  %16 = icmp eq i64 %a1.sroa.10.0.copyload, %a1.sroa.9.0.copyload
  %17 = icmp eq i64 %a1.sroa.11.0.copyload, 1
  %or.cond.i.i.i.i.i.i.i = and i1 %16, %17
  %18 = inttoptr i64 %a1.sroa.739.0.copyload to i32*
  %switchtmp.i.i.i.i = icmp eq i64 %a1.sroa.739.0.copyload, 0
  %19 = mul i64 %a1.sroa.9.0.copyload, %a1.sroa.8.0.copyload
  %20 = getelementptr inbounds i32, i32* %18, i64 %19
  %21 = icmp eq i64 %19, 0
  %22 = icmp ne i64 %a1.sroa.8.0.copyload, 0
  %23 = icmp ne i64 %a1.sroa.9.0.copyload, 0
  %or.cond.i.i.i.i.i.i.i.i = and i1 %22, %23
  %24 = mul i64 %a1.sroa.9.0.copyload, %a1.sroa.8.0.copyload
  %25 = shl i64 %24, 2
  %26 = add i64 %25, -4
  %27 = lshr exact i64 %26, 2
  %28 = add nuw nsw i64 %27, 1
  br i1 %or.cond.i.i.i.i.i.i.i, label %bb7.i.us.preheader, label %bb7.i.preheader

bb7.i.preheader:                                  ; preds = %bb7.lr.ph.i
  %29 = and i64 %28, 9223372036854775792
  %30 = add nsw i64 %29, -16
  %31 = lshr exact i64 %30, 4
  %32 = add nuw nsw i64 %31, 1
  %min.iters.check = icmp ult i64 %28, 16
  %n.vec = and i64 %28, 9223372036854775792
  %cmp.zero = icmp eq i64 %n.vec, 0
  %ind.end = getelementptr i32, i32* %18, i64 %n.vec
  %xtraiter212 = and i64 %32, 1
  %33 = icmp eq i64 %31, 0
  %lcmp.mod213 = icmp eq i64 %xtraiter212, 0
  %cmp.n = icmp eq i64 %28, %n.vec
  %unroll_iter214 = sub nsw i64 %32, %xtraiter212
  br label %bb7.i

bb7.i.us.preheader:                               ; preds = %bb7.lr.ph.i
  %34 = and i64 %28, 9223372036854775792
  %35 = add nsw i64 %34, -16
  %36 = lshr exact i64 %35, 4
  %37 = add nuw nsw i64 %36, 1
  %min.iters.check137 = icmp ult i64 %28, 16
  %n.vec140 = and i64 %28, 9223372036854775792
  %cmp.zero141 = icmp eq i64 %n.vec140, 0
  %ind.end146 = getelementptr i32, i32* %18, i64 %n.vec140
  %xtraiter = and i64 %37, 1
  %38 = icmp eq i64 %36, 0
  %lcmp.mod = icmp eq i64 %xtraiter, 0
  %cmp.n147 = icmp eq i64 %28, %n.vec140
  %unroll_iter = sub nsw i64 %37, %xtraiter
  br label %bb7.i.us

bb7.i.us:                                         ; preds = %bb7.i.us.preheader, %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i.us"
  %iter.sroa.0.035.i.us = phi i64 [ %39, %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i.us" ], [ 0, %bb7.i.us.preheader ]
  %39 = add nuw i64 %iter.sroa.0.035.i.us, 1
  br i1 %switchtmp.i.i.i.i, label %bb3.us.preheader.i.i.us, label %bb3.preheader.i.i.us

bb3.preheader.i.i.us:                             ; preds = %bb7.i.us
  br i1 %21, label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i.us", label %bb6.i.i21.us.preheader

bb6.i.i21.us.preheader:                           ; preds = %bb3.preheader.i.i.us
  br i1 %min.iters.check137, label %bb6.i.i21.us.preheader198, label %min.iters.checked138

min.iters.checked138:                             ; preds = %bb6.i.i21.us.preheader
  br i1 %cmp.zero141, label %bb6.i.i21.us.preheader198, label %vector.body134.preheader

vector.body134.preheader:                         ; preds = %min.iters.checked138
  br i1 %38, label %middle.block135.unr-lcssa, label %vector.body134.preheader.new

vector.body134.preheader.new:                     ; preds = %vector.body134.preheader
  br label %vector.body134

vector.body134:                                   ; preds = %vector.body134, %vector.body134.preheader.new
  %index143 = phi i64 [ 0, %vector.body134.preheader.new ], [ %index.next144.1, %vector.body134 ]
  %vec.phi148 = phi <4 x i32> [ zeroinitializer, %vector.body134.preheader.new ], [ %58, %vector.body134 ]
  %vec.phi149 = phi <4 x i32> [ zeroinitializer, %vector.body134.preheader.new ], [ %59, %vector.body134 ]
  %vec.phi150 = phi <4 x i32> [ zeroinitializer, %vector.body134.preheader.new ], [ %60, %vector.body134 ]
  %vec.phi151 = phi <4 x i32> [ zeroinitializer, %vector.body134.preheader.new ], [ %61, %vector.body134 ]
  %niter = phi i64 [ %unroll_iter, %vector.body134.preheader.new ], [ %niter.nsub.1, %vector.body134 ]
  %next.gep152 = getelementptr i32, i32* %18, i64 %index143
  %40 = bitcast i32* %next.gep152 to <4 x i32>*
  %wide.load184 = load <4 x i32>, <4 x i32>* %40, align 4
  %41 = getelementptr i32, i32* %next.gep152, i64 4
  %42 = bitcast i32* %41 to <4 x i32>*
  %wide.load185 = load <4 x i32>, <4 x i32>* %42, align 4
  %43 = getelementptr i32, i32* %next.gep152, i64 8
  %44 = bitcast i32* %43 to <4 x i32>*
  %wide.load186 = load <4 x i32>, <4 x i32>* %44, align 4
  %45 = getelementptr i32, i32* %next.gep152, i64 12
  %46 = bitcast i32* %45 to <4 x i32>*
  %wide.load187 = load <4 x i32>, <4 x i32>* %46, align 4
  %47 = add <4 x i32> %wide.load184, %vec.phi148
  %48 = add <4 x i32> %wide.load185, %vec.phi149
  %49 = add <4 x i32> %wide.load186, %vec.phi150
  %50 = add <4 x i32> %wide.load187, %vec.phi151
  %index.next144 = or i64 %index143, 16
  %next.gep152.1 = getelementptr i32, i32* %18, i64 %index.next144
  %51 = bitcast i32* %next.gep152.1 to <4 x i32>*
  %wide.load184.1 = load <4 x i32>, <4 x i32>* %51, align 4
  %52 = getelementptr i32, i32* %next.gep152.1, i64 4
  %53 = bitcast i32* %52 to <4 x i32>*
  %wide.load185.1 = load <4 x i32>, <4 x i32>* %53, align 4
  %54 = getelementptr i32, i32* %next.gep152.1, i64 8
  %55 = bitcast i32* %54 to <4 x i32>*
  %wide.load186.1 = load <4 x i32>, <4 x i32>* %55, align 4
  %56 = getelementptr i32, i32* %next.gep152.1, i64 12
  %57 = bitcast i32* %56 to <4 x i32>*
  %wide.load187.1 = load <4 x i32>, <4 x i32>* %57, align 4
  %58 = add <4 x i32> %wide.load184.1, %47
  %59 = add <4 x i32> %wide.load185.1, %48
  %60 = add <4 x i32> %wide.load186.1, %49
  %61 = add <4 x i32> %wide.load187.1, %50
  %index.next144.1 = add i64 %index143, 32
  %niter.nsub.1 = add i64 %niter, -2
  %niter.ncmp.1 = icmp eq i64 %niter.nsub.1, 0
  br i1 %niter.ncmp.1, label %middle.block135.unr-lcssa.loopexit, label %vector.body134, !llvm.loop !17

middle.block135.unr-lcssa.loopexit:               ; preds = %vector.body134
  br label %middle.block135.unr-lcssa

middle.block135.unr-lcssa:                        ; preds = %middle.block135.unr-lcssa.loopexit, %vector.body134.preheader
  %.lcssa205.ph = phi <4 x i32> [ undef, %vector.body134.preheader ], [ %58, %middle.block135.unr-lcssa.loopexit ]
  %.lcssa204.ph = phi <4 x i32> [ undef, %vector.body134.preheader ], [ %59, %middle.block135.unr-lcssa.loopexit ]
  %.lcssa203.ph = phi <4 x i32> [ undef, %vector.body134.preheader ], [ %60, %middle.block135.unr-lcssa.loopexit ]
  %.lcssa.ph = phi <4 x i32> [ undef, %vector.body134.preheader ], [ %61, %middle.block135.unr-lcssa.loopexit ]
  %index143.unr = phi i64 [ 0, %vector.body134.preheader ], [ %index.next144.1, %middle.block135.unr-lcssa.loopexit ]
  %vec.phi148.unr = phi <4 x i32> [ zeroinitializer, %vector.body134.preheader ], [ %58, %middle.block135.unr-lcssa.loopexit ]
  %vec.phi149.unr = phi <4 x i32> [ zeroinitializer, %vector.body134.preheader ], [ %59, %middle.block135.unr-lcssa.loopexit ]
  %vec.phi150.unr = phi <4 x i32> [ zeroinitializer, %vector.body134.preheader ], [ %60, %middle.block135.unr-lcssa.loopexit ]
  %vec.phi151.unr = phi <4 x i32> [ zeroinitializer, %vector.body134.preheader ], [ %61, %middle.block135.unr-lcssa.loopexit ]
  br i1 %lcmp.mod, label %middle.block135, label %vector.body134.epil.preheader

vector.body134.epil.preheader:                    ; preds = %middle.block135.unr-lcssa
  br label %vector.body134.epil

vector.body134.epil:                              ; preds = %vector.body134.epil.preheader
  %next.gep152.epil = getelementptr i32, i32* %18, i64 %index143.unr
  br label %middle.block135.epilog-lcssa

middle.block135.epilog-lcssa:                     ; preds = %vector.body134.epil
  %62 = getelementptr i32, i32* %next.gep152.epil, i64 12
  %63 = bitcast i32* %62 to <4 x i32>*
  %wide.load187.epil = load <4 x i32>, <4 x i32>* %63, align 4
  %64 = add <4 x i32> %wide.load187.epil, %vec.phi151.unr
  %65 = getelementptr i32, i32* %next.gep152.epil, i64 8
  %66 = bitcast i32* %65 to <4 x i32>*
  %wide.load186.epil = load <4 x i32>, <4 x i32>* %66, align 4
  %67 = add <4 x i32> %wide.load186.epil, %vec.phi150.unr
  %68 = getelementptr i32, i32* %next.gep152.epil, i64 4
  %69 = bitcast i32* %68 to <4 x i32>*
  %wide.load185.epil = load <4 x i32>, <4 x i32>* %69, align 4
  %70 = add <4 x i32> %wide.load185.epil, %vec.phi149.unr
  %71 = bitcast i32* %next.gep152.epil to <4 x i32>*
  %wide.load184.epil = load <4 x i32>, <4 x i32>* %71, align 4
  %72 = add <4 x i32> %wide.load184.epil, %vec.phi148.unr
  br label %middle.block135

middle.block135:                                  ; preds = %middle.block135.unr-lcssa, %middle.block135.epilog-lcssa
  %.lcssa205 = phi <4 x i32> [ %.lcssa205.ph, %middle.block135.unr-lcssa ], [ %72, %middle.block135.epilog-lcssa ]
  %.lcssa204 = phi <4 x i32> [ %.lcssa204.ph, %middle.block135.unr-lcssa ], [ %70, %middle.block135.epilog-lcssa ]
  %.lcssa203 = phi <4 x i32> [ %.lcssa203.ph, %middle.block135.unr-lcssa ], [ %67, %middle.block135.epilog-lcssa ]
  %.lcssa = phi <4 x i32> [ %.lcssa.ph, %middle.block135.unr-lcssa ], [ %64, %middle.block135.epilog-lcssa ]
  %bin.rdx190 = add <4 x i32> %.lcssa204, %.lcssa205
  %bin.rdx191 = add <4 x i32> %.lcssa203, %bin.rdx190
  %bin.rdx192 = add <4 x i32> %.lcssa, %bin.rdx191
  %rdx.shuf193 = shufflevector <4 x i32> %bin.rdx192, <4 x i32> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
  %bin.rdx194 = add <4 x i32> %bin.rdx192, %rdx.shuf193
  %rdx.shuf195 = shufflevector <4 x i32> %bin.rdx194, <4 x i32> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
  %bin.rdx196 = add <4 x i32> %bin.rdx194, %rdx.shuf195
  %73 = extractelement <4 x i32> %bin.rdx196, i32 0
  br i1 %cmp.n147, label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i.us", label %bb6.i.i21.us.preheader198

bb6.i.i21.us.preheader198:                        ; preds = %middle.block135, %min.iters.checked138, %bb6.i.i21.us.preheader
  %sum.0.i28.i.us.ph = phi i32 [ 0, %min.iters.checked138 ], [ 0, %bb6.i.i21.us.preheader ], [ %73, %middle.block135 ]
  %iter.sroa.4.0.i27.i.us.ph = phi i32* [ %18, %min.iters.checked138 ], [ %18, %bb6.i.i21.us.preheader ], [ %ind.end146, %middle.block135 ]
  br label %bb6.i.i21.us

bb6.i.i21.us:                                     ; preds = %bb6.i.i21.us.preheader198, %bb6.i.i21.us
  %sum.0.i28.i.us = phi i32 [ %76, %bb6.i.i21.us ], [ %sum.0.i28.i.us.ph, %bb6.i.i21.us.preheader198 ]
  %iter.sroa.4.0.i27.i.us = phi i32* [ %74, %bb6.i.i21.us ], [ %iter.sroa.4.0.i27.i.us.ph, %bb6.i.i21.us.preheader198 ]
  %74 = getelementptr inbounds i32, i32* %iter.sroa.4.0.i27.i.us, i64 1
  %75 = load i32, i32* %iter.sroa.4.0.i27.i.us, align 4
  %76 = add i32 %75, %sum.0.i28.i.us
  %77 = icmp eq i32* %74, %20
  br i1 %77, label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i.us.loopexit199", label %bb6.i.i21.us, !llvm.loop !20

bb3.us.preheader.i.i.us:                          ; preds = %bb7.i.us
  br i1 %or.cond.i.i.i.i.i.i.i.i, label %bb4.i.i.i.i.us.i.i.us.preheader, label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i.us"

bb4.i.i.i.i.us.i.i.us.preheader:                  ; preds = %bb3.us.preheader.i.i.us
  br label %bb4.i.i.i.i.us.i.i.us

bb4.i.i.i.i.us.i.i.us:                            ; preds = %bb4.i.i.i.i.us.i.i.us.preheader, %bb6.us.i.i.us
  %sum.0.us.i32.i.us = phi i32 [ %86, %bb6.us.i.i.us ], [ 0, %bb4.i.i.i.i.us.i.i.us.preheader ]
  %iter.sroa.14.0.us.i31.i.us = phi i64 [ %_13.sroa.5.0.i.i.i.i.us.i.i.us, %bb6.us.i.i.us ], [ 0, %bb4.i.i.i.i.us.i.i.us.preheader ]
  %iter.sroa.16.0.us.i30.i.us = phi i64 [ %_13.sroa.6.0.i.i.i.i.us.i.i.us, %bb6.us.i.i.us ], [ 0, %bb4.i.i.i.i.us.i.i.us.preheader ]
  %78 = mul i64 %iter.sroa.14.0.us.i31.i.us, %a1.sroa.9.0.copyload
  %79 = add i64 %iter.sroa.16.0.us.i30.i.us, %78
  %80 = add i64 %iter.sroa.16.0.us.i30.i.us, 1
  %81 = icmp eq i64 %80, %a1.sroa.9.0.copyload
  br i1 %81, label %bb1.i.i.i.i.i.us.i.i.us, label %bb6.i.i.i.i.i.us.i.i.us

bb1.i.i.i.i.i.us.i.i.us:                          ; preds = %bb4.i.i.i.i.us.i.i.us
  %82 = add i64 %iter.sroa.14.0.us.i31.i.us, 1
  %83 = icmp eq i64 %82, %a1.sroa.8.0.copyload
  br i1 %83, label %"_ZN76_$LT$$LP$usize$C$$u20$usize$RP$$u20$as$u20$ndarray..dimension..Dimension$GT$8next_for17hd8af96c1e9300594E.exit.i.i.i.i.us.i.i.us", label %bb6.i.i.i.i.i.us.i.i.us

bb6.i.i.i.i.i.us.i.i.us:                          ; preds = %bb1.i.i.i.i.i.us.i.i.us, %bb4.i.i.i.i.us.i.i.us
  %j.0.i.i.i.i.i.us.i.i.us = phi i64 [ 0, %bb1.i.i.i.i.i.us.i.i.us ], [ %80, %bb4.i.i.i.i.us.i.i.us ]
  %i.0.i.i.i.i.i.us.i.i.us = phi i64 [ %82, %bb1.i.i.i.i.i.us.i.i.us ], [ %iter.sroa.14.0.us.i31.i.us, %bb4.i.i.i.i.us.i.i.us ]
  br label %"_ZN76_$LT$$LP$usize$C$$u20$usize$RP$$u20$as$u20$ndarray..dimension..Dimension$GT$8next_for17hd8af96c1e9300594E.exit.i.i.i.i.us.i.i.us"

"_ZN76_$LT$$LP$usize$C$$u20$usize$RP$$u20$as$u20$ndarray..dimension..Dimension$GT$8next_for17hd8af96c1e9300594E.exit.i.i.i.i.us.i.i.us": ; preds = %bb6.i.i.i.i.i.us.i.i.us, %bb1.i.i.i.i.i.us.i.i.us
  %_13.sroa.0.0.i.i.i.i.us.i.i.us = phi i64 [ 1, %bb6.i.i.i.i.i.us.i.i.us ], [ 0, %bb1.i.i.i.i.i.us.i.i.us ]
  %_13.sroa.5.0.i.i.i.i.us.i.i.us = phi i64 [ %i.0.i.i.i.i.i.us.i.i.us, %bb6.i.i.i.i.i.us.i.i.us ], [ undef, %bb1.i.i.i.i.i.us.i.i.us ]
  %_13.sroa.6.0.i.i.i.i.us.i.i.us = phi i64 [ %j.0.i.i.i.i.i.us.i.i.us, %bb6.i.i.i.i.i.us.i.i.us ], [ undef, %bb1.i.i.i.i.i.us.i.i.us ]
  %84 = getelementptr inbounds i32, i32* %18, i64 %79
  %switchtmp.us.i.i.us = icmp eq i32* %84, null
  br i1 %switchtmp.us.i.i.us, label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i.us.loopexit", label %bb6.us.i.i.us

bb6.us.i.i.us:                                    ; preds = %"_ZN76_$LT$$LP$usize$C$$u20$usize$RP$$u20$as$u20$ndarray..dimension..Dimension$GT$8next_for17hd8af96c1e9300594E.exit.i.i.i.i.us.i.i.us"
  %85 = load i32, i32* %84, align 4
  %86 = add i32 %85, %sum.0.us.i32.i.us
  %switch.i.i.i.i.us.i.i.us = icmp eq i64 %_13.sroa.0.0.i.i.i.i.us.i.i.us, 1
  br i1 %switch.i.i.i.i.us.i.i.us, label %bb4.i.i.i.i.us.i.i.us, label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i.us.loopexit"

"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i.us.loopexit": ; preds = %bb6.us.i.i.us, %"_ZN76_$LT$$LP$usize$C$$u20$usize$RP$$u20$as$u20$ndarray..dimension..Dimension$GT$8next_for17hd8af96c1e9300594E.exit.i.i.i.i.us.i.i.us"
  %sum.0.lcssa.i.i.us.ph = phi i32 [ %86, %bb6.us.i.i.us ], [ %sum.0.us.i32.i.us, %"_ZN76_$LT$$LP$usize$C$$u20$usize$RP$$u20$as$u20$ndarray..dimension..Dimension$GT$8next_for17hd8af96c1e9300594E.exit.i.i.i.i.us.i.i.us" ]
  br label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i.us"

"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i.us.loopexit199": ; preds = %bb6.i.i21.us
  br label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i.us"

"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i.us": ; preds = %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i.us.loopexit199", %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i.us.loopexit", %middle.block135, %bb3.us.preheader.i.i.us, %bb3.preheader.i.i.us
  %sum.0.lcssa.i.i.us = phi i32 [ 0, %bb3.us.preheader.i.i.us ], [ 0, %bb3.preheader.i.i.us ], [ %73, %middle.block135 ], [ %sum.0.lcssa.i.i.us.ph, %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i.us.loopexit" ], [ %76, %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i.us.loopexit199" ]
  call void @llvm.lifetime.start(i64 4, i8* %15) #2
  store i32 %sum.0.lcssa.i.i.us, i32* %dummy.i.i, align 4
  call void asm "", "r,~{dirflag},~{fpsr},~{flags}"(i32* nonnull %dummy.i.i) #2, !srcloc !16
  call void @llvm.lifetime.end(i64 4, i8* %15) #2
  %exitcond.i.us = icmp eq i64 %39, %9
  br i1 %exitcond.i.us, label %bb8.i.loopexit, label %bb7.i.us

bb7.i:                                            ; preds = %bb7.i.preheader, %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i"
  %iter.sroa.0.035.i = phi i64 [ %87, %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i" ], [ 0, %bb7.i.preheader ]
  %87 = add nuw i64 %iter.sroa.0.035.i, 1
  call void @llvm.lifetime.start(i64 40, i8* %11), !noalias !22
  store i64 %a1.sroa.739.0.copyload, i64* %_11.sroa.0.0..sroa_cast.i.i.i, align 8, !noalias !29
  store i64 %a1.sroa.8.0.copyload, i64* %_11.sroa.4.0..sroa_idx8.i.i.i, align 8, !noalias !29
  store i64 %a1.sroa.9.0.copyload, i64* %_11.sroa.5.0..sroa_idx10.i.i.i, align 8, !noalias !29
  store i64 %a1.sroa.10.0.copyload, i64* %_11.sroa.6.0..sroa_idx12.i.i.i, align 8, !noalias !29
  store i64 %a1.sroa.11.0.copyload, i64* %_11.sroa.7.0..sroa_idx14.i.i.i, align 8, !noalias !29
  call void @llvm.lifetime.start(i64 16, i8* %12), !noalias !30
  store i64 %a1.sroa.9.0.copyload, i64* %13, align 8, !alias.scope !36, !noalias !30
  store i64 1, i64* %14, align 8, !alias.scope !36, !noalias !30
  br label %bb18.i.i.i.i.i.i

bb18.i.i.i.i.i.i:                                 ; preds = %bb7.i, %"_ZN145_$LT$itertools..cons_tuples_impl..ConsTuples$LT$Iter$C$$u20$$LP$$LP$G$C$$u20$H$RP$$C$$u20$X$RP$$GT$$u20$as$u20$core..iter..iterator..Iterator$GT$4next17hb28c8d7d69dbfae8E.exit.i.i.i.i.i.i"
  %iter.sroa.738.0.i.i.i.i.i.i = phi i64 [ %92, %"_ZN145_$LT$itertools..cons_tuples_impl..ConsTuples$LT$Iter$C$$u20$$LP$$LP$G$C$$u20$H$RP$$C$$u20$X$RP$$GT$$u20$as$u20$core..iter..iterator..Iterator$GT$4next17hb28c8d7d69dbfae8E.exit.i.i.i.i.i.i" ], [ 0, %bb7.i ]
  %88 = icmp ult i64 %iter.sroa.738.0.i.i.i.i.i.i, 2
  br i1 %88, label %"_ZN145_$LT$itertools..cons_tuples_impl..ConsTuples$LT$Iter$C$$u20$$LP$$LP$G$C$$u20$H$RP$$C$$u20$X$RP$$GT$$u20$as$u20$core..iter..iterator..Iterator$GT$4next17hb28c8d7d69dbfae8E.exit.i.i.i.i.i.i", label %"_ZN7ndarray12impl_methods54_$LT$impl$u20$ndarray..ArrayBase$LT$S$C$$u20$D$GT$$GT$18is_standard_layout17he690d19e37be350fE.exit.thread.i.i.i.i.i"

"_ZN145_$LT$itertools..cons_tuples_impl..ConsTuples$LT$Iter$C$$u20$$LP$$LP$G$C$$u20$H$RP$$C$$u20$X$RP$$GT$$u20$as$u20$core..iter..iterator..Iterator$GT$4next17hb28c8d7d69dbfae8E.exit.i.i.i.i.i.i": ; preds = %bb18.i.i.i.i.i.i
  %89 = getelementptr inbounds i64, i64* %_11.sroa.4.0..sroa_idx8.i.i.i, i64 %iter.sroa.738.0.i.i.i.i.i.i
  %90 = getelementptr inbounds i64, i64* %_11.sroa.6.0..sroa_idx12.i.i.i, i64 %iter.sroa.738.0.i.i.i.i.i.i
  %91 = getelementptr inbounds i64, i64* %13, i64 %iter.sroa.738.0.i.i.i.i.i.i
  %92 = add nuw nsw i64 %iter.sroa.738.0.i.i.i.i.i.i, 1
  %93 = load i64, i64* %89, align 8, !alias.scope !39, !noalias !40
  %94 = load i64, i64* %90, align 8, !alias.scope !39, !noalias !40
  %95 = load i64, i64* %91, align 8, !noalias !30
  %96 = icmp ne i64 %93, 1
  %97 = icmp ne i64 %94, %95
  %or.cond.i.i.i.i.i.i = and i1 %96, %97
  br i1 %or.cond.i.i.i.i.i.i, label %"_ZN85_$LT$ndarray..ArrayBase$LT$ndarray..ViewRepr$LT$$RF$$u27$a$u20$A$GT$$C$$u20$D$GT$$GT$10into_slice17haf36f7fe3600206dE.exit.thread.i.i.i.i", label %bb18.i.i.i.i.i.i

"_ZN7ndarray12impl_methods54_$LT$impl$u20$ndarray..ArrayBase$LT$S$C$$u20$D$GT$$GT$18is_standard_layout17he690d19e37be350fE.exit.thread.i.i.i.i.i": ; preds = %bb18.i.i.i.i.i.i
  call void @llvm.lifetime.end(i64 16, i8* nonnull %12), !noalias !30
  br i1 %switchtmp.i.i.i.i, label %bb3.us.preheader.i.i, label %bb3.preheader.i.i

"_ZN85_$LT$ndarray..ArrayBase$LT$ndarray..ViewRepr$LT$$RF$$u27$a$u20$A$GT$$C$$u20$D$GT$$GT$10into_slice17haf36f7fe3600206dE.exit.thread.i.i.i.i": ; preds = %"_ZN145_$LT$itertools..cons_tuples_impl..ConsTuples$LT$Iter$C$$u20$$LP$$LP$G$C$$u20$H$RP$$C$$u20$X$RP$$GT$$u20$as$u20$core..iter..iterator..Iterator$GT$4next17hb28c8d7d69dbfae8E.exit.i.i.i.i.i.i"
  call void @llvm.lifetime.end(i64 16, i8* nonnull %12), !noalias !30
  br label %bb3.us.preheader.i.i

bb3.preheader.i.i:                                ; preds = %"_ZN7ndarray12impl_methods54_$LT$impl$u20$ndarray..ArrayBase$LT$S$C$$u20$D$GT$$GT$18is_standard_layout17he690d19e37be350fE.exit.thread.i.i.i.i.i"
  call void @llvm.lifetime.end(i64 40, i8* nonnull %11), !noalias !22
  br i1 %21, label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i", label %bb6.i.i21.preheader

bb6.i.i21.preheader:                              ; preds = %bb3.preheader.i.i
  br i1 %min.iters.check, label %bb6.i.i21.preheader200, label %min.iters.checked

bb6.i.i21.preheader200:                           ; preds = %middle.block, %min.iters.checked, %bb6.i.i21.preheader
  %sum.0.i28.i.ph = phi i32 [ 0, %min.iters.checked ], [ 0, %bb6.i.i21.preheader ], [ %131, %middle.block ]
  %iter.sroa.4.0.i27.i.ph = phi i32* [ %18, %min.iters.checked ], [ %18, %bb6.i.i21.preheader ], [ %ind.end, %middle.block ]
  br label %bb6.i.i21

min.iters.checked:                                ; preds = %bb6.i.i21.preheader
  br i1 %cmp.zero, label %bb6.i.i21.preheader200, label %vector.body.preheader

vector.body.preheader:                            ; preds = %min.iters.checked
  br i1 %33, label %middle.block.unr-lcssa, label %vector.body.preheader.new

vector.body.preheader.new:                        ; preds = %vector.body.preheader
  br label %vector.body

vector.body:                                      ; preds = %vector.body, %vector.body.preheader.new
  %index = phi i64 [ 0, %vector.body.preheader.new ], [ %index.next.1, %vector.body ]
  %vec.phi = phi <4 x i32> [ zeroinitializer, %vector.body.preheader.new ], [ %116, %vector.body ]
  %vec.phi93 = phi <4 x i32> [ zeroinitializer, %vector.body.preheader.new ], [ %117, %vector.body ]
  %vec.phi94 = phi <4 x i32> [ zeroinitializer, %vector.body.preheader.new ], [ %118, %vector.body ]
  %vec.phi95 = phi <4 x i32> [ zeroinitializer, %vector.body.preheader.new ], [ %119, %vector.body ]
  %niter215 = phi i64 [ %unroll_iter214, %vector.body.preheader.new ], [ %niter215.nsub.1, %vector.body ]
  %next.gep = getelementptr i32, i32* %18, i64 %index
  %98 = bitcast i32* %next.gep to <4 x i32>*
  %wide.load = load <4 x i32>, <4 x i32>* %98, align 4
  %99 = getelementptr i32, i32* %next.gep, i64 4
  %100 = bitcast i32* %99 to <4 x i32>*
  %wide.load126 = load <4 x i32>, <4 x i32>* %100, align 4
  %101 = getelementptr i32, i32* %next.gep, i64 8
  %102 = bitcast i32* %101 to <4 x i32>*
  %wide.load127 = load <4 x i32>, <4 x i32>* %102, align 4
  %103 = getelementptr i32, i32* %next.gep, i64 12
  %104 = bitcast i32* %103 to <4 x i32>*
  %wide.load128 = load <4 x i32>, <4 x i32>* %104, align 4
  %105 = add <4 x i32> %wide.load, %vec.phi
  %106 = add <4 x i32> %wide.load126, %vec.phi93
  %107 = add <4 x i32> %wide.load127, %vec.phi94
  %108 = add <4 x i32> %wide.load128, %vec.phi95
  %index.next = or i64 %index, 16
  %next.gep.1 = getelementptr i32, i32* %18, i64 %index.next
  %109 = bitcast i32* %next.gep.1 to <4 x i32>*
  %wide.load.1 = load <4 x i32>, <4 x i32>* %109, align 4
  %110 = getelementptr i32, i32* %next.gep.1, i64 4
  %111 = bitcast i32* %110 to <4 x i32>*
  %wide.load126.1 = load <4 x i32>, <4 x i32>* %111, align 4
  %112 = getelementptr i32, i32* %next.gep.1, i64 8
  %113 = bitcast i32* %112 to <4 x i32>*
  %wide.load127.1 = load <4 x i32>, <4 x i32>* %113, align 4
  %114 = getelementptr i32, i32* %next.gep.1, i64 12
  %115 = bitcast i32* %114 to <4 x i32>*
  %wide.load128.1 = load <4 x i32>, <4 x i32>* %115, align 4
  %116 = add <4 x i32> %wide.load.1, %105
  %117 = add <4 x i32> %wide.load126.1, %106
  %118 = add <4 x i32> %wide.load127.1, %107
  %119 = add <4 x i32> %wide.load128.1, %108
  %index.next.1 = add i64 %index, 32
  %niter215.nsub.1 = add i64 %niter215, -2
  %niter215.ncmp.1 = icmp eq i64 %niter215.nsub.1, 0
  br i1 %niter215.ncmp.1, label %middle.block.unr-lcssa.loopexit, label %vector.body, !llvm.loop !41

middle.block.unr-lcssa.loopexit:                  ; preds = %vector.body
  br label %middle.block.unr-lcssa

middle.block.unr-lcssa:                           ; preds = %middle.block.unr-lcssa.loopexit, %vector.body.preheader
  %.lcssa210.ph = phi <4 x i32> [ undef, %vector.body.preheader ], [ %116, %middle.block.unr-lcssa.loopexit ]
  %.lcssa209.ph = phi <4 x i32> [ undef, %vector.body.preheader ], [ %117, %middle.block.unr-lcssa.loopexit ]
  %.lcssa208.ph = phi <4 x i32> [ undef, %vector.body.preheader ], [ %118, %middle.block.unr-lcssa.loopexit ]
  %.lcssa207.ph = phi <4 x i32> [ undef, %vector.body.preheader ], [ %119, %middle.block.unr-lcssa.loopexit ]
  %index.unr = phi i64 [ 0, %vector.body.preheader ], [ %index.next.1, %middle.block.unr-lcssa.loopexit ]
  %vec.phi.unr = phi <4 x i32> [ zeroinitializer, %vector.body.preheader ], [ %116, %middle.block.unr-lcssa.loopexit ]
  %vec.phi93.unr = phi <4 x i32> [ zeroinitializer, %vector.body.preheader ], [ %117, %middle.block.unr-lcssa.loopexit ]
  %vec.phi94.unr = phi <4 x i32> [ zeroinitializer, %vector.body.preheader ], [ %118, %middle.block.unr-lcssa.loopexit ]
  %vec.phi95.unr = phi <4 x i32> [ zeroinitializer, %vector.body.preheader ], [ %119, %middle.block.unr-lcssa.loopexit ]
  br i1 %lcmp.mod213, label %middle.block, label %vector.body.epil.preheader

vector.body.epil.preheader:                       ; preds = %middle.block.unr-lcssa
  br label %vector.body.epil

vector.body.epil:                                 ; preds = %vector.body.epil.preheader
  %next.gep.epil = getelementptr i32, i32* %18, i64 %index.unr
  br label %middle.block.epilog-lcssa

middle.block.epilog-lcssa:                        ; preds = %vector.body.epil
  %120 = getelementptr i32, i32* %next.gep.epil, i64 12
  %121 = bitcast i32* %120 to <4 x i32>*
  %wide.load128.epil = load <4 x i32>, <4 x i32>* %121, align 4
  %122 = add <4 x i32> %wide.load128.epil, %vec.phi95.unr
  %123 = getelementptr i32, i32* %next.gep.epil, i64 8
  %124 = bitcast i32* %123 to <4 x i32>*
  %wide.load127.epil = load <4 x i32>, <4 x i32>* %124, align 4
  %125 = add <4 x i32> %wide.load127.epil, %vec.phi94.unr
  %126 = getelementptr i32, i32* %next.gep.epil, i64 4
  %127 = bitcast i32* %126 to <4 x i32>*
  %wide.load126.epil = load <4 x i32>, <4 x i32>* %127, align 4
  %128 = add <4 x i32> %wide.load126.epil, %vec.phi93.unr
  %129 = bitcast i32* %next.gep.epil to <4 x i32>*
  %wide.load.epil = load <4 x i32>, <4 x i32>* %129, align 4
  %130 = add <4 x i32> %wide.load.epil, %vec.phi.unr
  br label %middle.block

middle.block:                                     ; preds = %middle.block.unr-lcssa, %middle.block.epilog-lcssa
  %.lcssa210 = phi <4 x i32> [ %.lcssa210.ph, %middle.block.unr-lcssa ], [ %130, %middle.block.epilog-lcssa ]
  %.lcssa209 = phi <4 x i32> [ %.lcssa209.ph, %middle.block.unr-lcssa ], [ %128, %middle.block.epilog-lcssa ]
  %.lcssa208 = phi <4 x i32> [ %.lcssa208.ph, %middle.block.unr-lcssa ], [ %125, %middle.block.epilog-lcssa ]
  %.lcssa207 = phi <4 x i32> [ %.lcssa207.ph, %middle.block.unr-lcssa ], [ %122, %middle.block.epilog-lcssa ]
  %bin.rdx = add <4 x i32> %.lcssa209, %.lcssa210
  %bin.rdx129 = add <4 x i32> %.lcssa208, %bin.rdx
  %bin.rdx130 = add <4 x i32> %.lcssa207, %bin.rdx129
  %rdx.shuf = shufflevector <4 x i32> %bin.rdx130, <4 x i32> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
  %bin.rdx131 = add <4 x i32> %bin.rdx130, %rdx.shuf
  %rdx.shuf132 = shufflevector <4 x i32> %bin.rdx131, <4 x i32> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
  %bin.rdx133 = add <4 x i32> %bin.rdx131, %rdx.shuf132
  %131 = extractelement <4 x i32> %bin.rdx133, i32 0
  br i1 %cmp.n, label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i", label %bb6.i.i21.preheader200

bb3.us.preheader.i.i:                             ; preds = %"_ZN7ndarray12impl_methods54_$LT$impl$u20$ndarray..ArrayBase$LT$S$C$$u20$D$GT$$GT$18is_standard_layout17he690d19e37be350fE.exit.thread.i.i.i.i.i", %"_ZN85_$LT$ndarray..ArrayBase$LT$ndarray..ViewRepr$LT$$RF$$u27$a$u20$A$GT$$C$$u20$D$GT$$GT$10into_slice17haf36f7fe3600206dE.exit.thread.i.i.i.i"
  call void @llvm.lifetime.end(i64 40, i8* nonnull %11), !noalias !22
  br i1 %or.cond.i.i.i.i.i.i.i.i, label %bb4.i.i.i.i.us.i.i.preheader, label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i"

bb4.i.i.i.i.us.i.i.preheader:                     ; preds = %bb3.us.preheader.i.i
  br label %bb4.i.i.i.i.us.i.i

bb4.i.i.i.i.us.i.i:                               ; preds = %bb4.i.i.i.i.us.i.i.preheader, %bb6.us.i.i
  %sum.0.us.i32.i = phi i32 [ %141, %bb6.us.i.i ], [ 0, %bb4.i.i.i.i.us.i.i.preheader ]
  %iter.sroa.14.0.us.i31.i = phi i64 [ %_13.sroa.5.0.i.i.i.i.us.i.i, %bb6.us.i.i ], [ 0, %bb4.i.i.i.i.us.i.i.preheader ]
  %iter.sroa.16.0.us.i30.i = phi i64 [ %_13.sroa.6.0.i.i.i.i.us.i.i, %bb6.us.i.i ], [ 0, %bb4.i.i.i.i.us.i.i.preheader ]
  %132 = mul i64 %iter.sroa.14.0.us.i31.i, %a1.sroa.10.0.copyload
  %133 = mul i64 %iter.sroa.16.0.us.i30.i, %a1.sroa.11.0.copyload
  %134 = add i64 %133, %132
  %135 = add i64 %iter.sroa.16.0.us.i30.i, 1
  %136 = icmp eq i64 %135, %a1.sroa.9.0.copyload
  br i1 %136, label %bb1.i.i.i.i.i.us.i.i, label %bb6.i.i.i.i.i.us.i.i

bb1.i.i.i.i.i.us.i.i:                             ; preds = %bb4.i.i.i.i.us.i.i
  %137 = add i64 %iter.sroa.14.0.us.i31.i, 1
  %138 = icmp eq i64 %137, %a1.sroa.8.0.copyload
  br i1 %138, label %"_ZN76_$LT$$LP$usize$C$$u20$usize$RP$$u20$as$u20$ndarray..dimension..Dimension$GT$8next_for17hd8af96c1e9300594E.exit.i.i.i.i.us.i.i", label %bb6.i.i.i.i.i.us.i.i

bb6.i.i.i.i.i.us.i.i:                             ; preds = %bb1.i.i.i.i.i.us.i.i, %bb4.i.i.i.i.us.i.i
  %j.0.i.i.i.i.i.us.i.i = phi i64 [ 0, %bb1.i.i.i.i.i.us.i.i ], [ %135, %bb4.i.i.i.i.us.i.i ]
  %i.0.i.i.i.i.i.us.i.i = phi i64 [ %137, %bb1.i.i.i.i.i.us.i.i ], [ %iter.sroa.14.0.us.i31.i, %bb4.i.i.i.i.us.i.i ]
  br label %"_ZN76_$LT$$LP$usize$C$$u20$usize$RP$$u20$as$u20$ndarray..dimension..Dimension$GT$8next_for17hd8af96c1e9300594E.exit.i.i.i.i.us.i.i"

"_ZN76_$LT$$LP$usize$C$$u20$usize$RP$$u20$as$u20$ndarray..dimension..Dimension$GT$8next_for17hd8af96c1e9300594E.exit.i.i.i.i.us.i.i": ; preds = %bb6.i.i.i.i.i.us.i.i, %bb1.i.i.i.i.i.us.i.i
  %_13.sroa.0.0.i.i.i.i.us.i.i = phi i64 [ 1, %bb6.i.i.i.i.i.us.i.i ], [ 0, %bb1.i.i.i.i.i.us.i.i ]
  %_13.sroa.5.0.i.i.i.i.us.i.i = phi i64 [ %i.0.i.i.i.i.i.us.i.i, %bb6.i.i.i.i.i.us.i.i ], [ undef, %bb1.i.i.i.i.i.us.i.i ]
  %_13.sroa.6.0.i.i.i.i.us.i.i = phi i64 [ %j.0.i.i.i.i.i.us.i.i, %bb6.i.i.i.i.i.us.i.i ], [ undef, %bb1.i.i.i.i.i.us.i.i ]
  %139 = getelementptr inbounds i32, i32* %18, i64 %134
  %switchtmp.us.i.i = icmp eq i32* %139, null
  br i1 %switchtmp.us.i.i, label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i.loopexit", label %bb6.us.i.i

bb6.us.i.i:                                       ; preds = %"_ZN76_$LT$$LP$usize$C$$u20$usize$RP$$u20$as$u20$ndarray..dimension..Dimension$GT$8next_for17hd8af96c1e9300594E.exit.i.i.i.i.us.i.i"
  %140 = load i32, i32* %139, align 4
  %141 = add i32 %140, %sum.0.us.i32.i
  %switch.i.i.i.i.us.i.i = icmp eq i64 %_13.sroa.0.0.i.i.i.i.us.i.i, 1
  br i1 %switch.i.i.i.i.us.i.i, label %bb4.i.i.i.i.us.i.i, label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i.loopexit"

bb6.i.i21:                                        ; preds = %bb6.i.i21.preheader200, %bb6.i.i21
  %sum.0.i28.i = phi i32 [ %144, %bb6.i.i21 ], [ %sum.0.i28.i.ph, %bb6.i.i21.preheader200 ]
  %iter.sroa.4.0.i27.i = phi i32* [ %142, %bb6.i.i21 ], [ %iter.sroa.4.0.i27.i.ph, %bb6.i.i21.preheader200 ]
  %142 = getelementptr inbounds i32, i32* %iter.sroa.4.0.i27.i, i64 1
  %143 = load i32, i32* %iter.sroa.4.0.i27.i, align 4
  %144 = add i32 %143, %sum.0.i28.i
  %145 = icmp eq i32* %142, %20
  br i1 %145, label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i.loopexit201", label %bb6.i.i21, !llvm.loop !42

"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i.loopexit": ; preds = %"_ZN76_$LT$$LP$usize$C$$u20$usize$RP$$u20$as$u20$ndarray..dimension..Dimension$GT$8next_for17hd8af96c1e9300594E.exit.i.i.i.i.us.i.i", %bb6.us.i.i
  %sum.0.lcssa.i.i.ph = phi i32 [ %141, %bb6.us.i.i ], [ %sum.0.us.i32.i, %"_ZN76_$LT$$LP$usize$C$$u20$usize$RP$$u20$as$u20$ndarray..dimension..Dimension$GT$8next_for17hd8af96c1e9300594E.exit.i.i.i.i.us.i.i" ]
  br label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i"

"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i.loopexit201": ; preds = %bb6.i.i21
  br label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i"

"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i": ; preds = %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i.loopexit201", %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i.loopexit", %middle.block, %bb3.us.preheader.i.i, %bb3.preheader.i.i
  %sum.0.lcssa.i.i = phi i32 [ 0, %bb3.us.preheader.i.i ], [ 0, %bb3.preheader.i.i ], [ %131, %middle.block ], [ %sum.0.lcssa.i.i.ph, %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i.loopexit" ], [ %144, %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i.loopexit201" ]
  call void @llvm.lifetime.start(i64 4, i8* %15) #2
  store i32 %sum.0.lcssa.i.i, i32* %dummy.i.i, align 4
  call void asm "", "r,~{dirflag},~{fpsr},~{flags}"(i32* nonnull %dummy.i.i) #2, !srcloc !16
  call void @llvm.lifetime.end(i64 4, i8* %15) #2
  %exitcond.i = icmp eq i64 %87, %9
  br i1 %exitcond.i, label %bb8.i.loopexit202, label %bb7.i

bb8.i.loopexit:                                   ; preds = %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i.us"
  br label %bb8.i

bb8.i.loopexit202:                                ; preds = %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hff6dac657e135198E.exit.i"
  br label %bb8.i

bb8.i:                                            ; preds = %bb8.i.loopexit202, %bb8.i.loopexit, %.noexc22
  invoke void @_ZN3std4time7Instant7elapsed17h49a3c3321963a7a7E(%"1.std::time::Duration"* noalias nocapture nonnull sret dereferenceable(16) %tmp_ret2.i, %"1.std::time::Instant"* noalias nonnull readonly dereferenceable(16) %start1.i)
          to label %bb5 unwind label %bb9

bb5:                                              ; preds = %bb8.i
  %146 = getelementptr inbounds %"1.std::time::Duration", %"1.std::time::Duration"* %tmp_ret2.i, i64 0, i32 0
  %147 = getelementptr inbounds %"1.std::time::Duration", %"1.std::time::Duration"* %tmp_ret2.i, i64 0, i32 1
  %148 = load i64, i64* %146, align 8
  %149 = load i32, i32* %147, align 8
  %150 = getelementptr inbounds %"12.test::Bencher", %"12.test::Bencher"* %0, i64 0, i32 1, i32 0
  store i64 %148, i64* %150, align 8
  %151 = getelementptr inbounds %"12.test::Bencher", %"12.test::Bencher"* %0, i64 0, i32 1, i32 1
  store i32 %149, i32* %151, align 4
  call void @llvm.lifetime.end(i64 16, i8* %7)
  call void @llvm.lifetime.end(i64 16, i8* %6)
  %not..i.i.i.i.i = icmp eq i64 %a1.sroa.5.0.copyload, 0
  br i1 %not..i.i.i.i.i, label %bb6, label %bb6.i.i.i.i.i24

bb6.i.i.i.i.i24:                                  ; preds = %bb5
  %152 = shl i64 %a1.sroa.5.0.copyload, 2
  call void @__rust_deallocate(i8* %a1.sroa.0.0.copyload, i64 %152, i64 4) #2
  br label %bb6

bb6:                                              ; preds = %bb6.i.i.i.i.i24, %bb5
  ret void

bb9:                                              ; preds = %bb4, %bb8.i
  %153 = landingpad { i8*, i32 }
          cleanup
  %.fca.0.extract5 = extractvalue { i8*, i32 } %153, 0
  %.fca.1.extract7 = extractvalue { i8*, i32 } %153, 1
  %not..i.i.i.i.i28 = icmp eq i64 %a1.sroa.5.0.copyload, 0
  br i1 %not..i.i.i.i.i28, label %bb1, label %bb6.i.i.i.i.i29

bb6.i.i.i.i.i29:                                  ; preds = %bb9
  %154 = shl i64 %a1.sroa.5.0.copyload, 2
  call void @__rust_deallocate(i8* %a1.sroa.0.0.copyload, i64 %154, i64 4) #2
  br label %bb1

cleanup:                                          ; preds = %bb10.i.i.i.i.i
  %155 = landingpad { i8*, i32 }
          cleanup
  %.fca.0.extract4 = extractvalue { i8*, i32 } %155, 0
  %.fca.1.extract = extractvalue { i8*, i32 } %155, 1
  br label %bb1
}

; Function Attrs: uwtable
define internal void @_ZN6bench16__test4main17hea89029b865a9639E() unnamed_addr #0 {
entry-block:
  tail call void @_ZN4test16test_main_static17h2986dcfcbcd72e4aE(%"12.test::TestDescAndFn"* noalias nonnull readonly bitcast ({ { { { i64, %str_slice, [8 x i8] }, i1, [7 x i8], { i64, [16 x i8] } }, { i64, void (%"12.test::Bencher"*)*, [8 x i8] } } }* @ref10698 to %"12.test::TestDescAndFn"*), i64 1)
  ret void
}

; Function Attrs: argmemonly nounwind
declare void @llvm.lifetime.start(i64, i8* nocapture) #1

; Function Attrs: nounwind
declare i32 @rust_eh_personality(i32, i32, i64, %"8.unwind::libunwind::_Unwind_Exception"*, %"8.unwind::libunwind::_Unwind_Context"*) unnamed_addr #2

; Function Attrs: argmemonly nounwind
declare void @llvm.lifetime.end(i64, i8* nocapture) #1

; Function Attrs: cold noinline noreturn
declare void @_ZN5alloc3oom3oom17hb4382286715c86b5E() unnamed_addr #3

declare void @_ZN3std4time7Instant3now17h92a4d74662f63e98E(%"1.std::time::Instant"* noalias nocapture sret dereferenceable(16)) unnamed_addr

declare void @_ZN3std4time7Instant7elapsed17h49a3c3321963a7a7E(%"1.std::time::Duration"* noalias nocapture sret dereferenceable(16), %"1.std::time::Instant"* noalias readonly dereferenceable(16)) unnamed_addr

; Function Attrs: nounwind
declare void @__rust_deallocate(i8*, i64, i64) unnamed_addr #2

; Function Attrs: nounwind
declare noalias i8* @__rust_allocate(i64, i64) unnamed_addr #2

declare void @_ZN4test16test_main_static17h2986dcfcbcd72e4aE(%"12.test::TestDescAndFn"* noalias nonnull readonly, i64) unnamed_addr

define i64 @main(i64, i8**) unnamed_addr {
top:
  %2 = tail call i64 @_ZN3std2rt10lang_start17h19515238ab9c0a39E(i8* bitcast (void ()* @_ZN6bench16__test4main17hea89029b865a9639E to i8*), i64 %0, i8** %1)
  ret i64 %2
}

declare i64 @_ZN3std2rt10lang_start17h19515238ab9c0a39E(i8*, i64, i8**) unnamed_addr

; Function Attrs: argmemonly nounwind
declare void @llvm.memset.p0i8.i64(i8* nocapture writeonly, i8, i64, i32, i1) #1

attributes #0 = { uwtable }
attributes #1 = { argmemonly nounwind }
attributes #2 = { nounwind }
attributes #3 = { cold noinline noreturn }

!llvm.module.flags = !{!0}

!0 = !{i32 1, !"PIE Level", i32 2}
!1 = !{!2, !4, !6, !8, !10}
!2 = distinct !{!2, !3, !"_ZN40_$LT$alloc..raw_vec..RawVec$LT$T$GT$$GT$13with_capacity17h442131ee31b75b17E: argument 0"}
!3 = distinct !{!3, !"_ZN40_$LT$alloc..raw_vec..RawVec$LT$T$GT$$GT$13with_capacity17h442131ee31b75b17E"}
!4 = distinct !{!4, !5, !"_ZN39_$LT$collections..vec..Vec$LT$T$GT$$GT$13with_capacity17hafff9947d4a7db1aE: argument 0"}
!5 = distinct !{!5, !"_ZN39_$LT$collections..vec..Vec$LT$T$GT$$GT$13with_capacity17hafff9947d4a7db1aE"}
!6 = distinct !{!6, !7, !"_ZN11collections3vec9from_elem17h2478192a4323e46eE: argument 0"}
!7 = distinct !{!7, !"_ZN11collections3vec9from_elem17h2478192a4323e46eE"}
!8 = distinct !{!8, !9, !"_ZN7ndarray17impl_constructors54_$LT$impl$u20$ndarray..ArrayBase$LT$S$C$$u20$D$GT$$GT$9from_elem17hede15a45bcf452dbE: argument 0"}
!9 = distinct !{!9, !"_ZN7ndarray17impl_constructors54_$LT$impl$u20$ndarray..ArrayBase$LT$S$C$$u20$D$GT$$GT$9from_elem17hede15a45bcf452dbE"}
!10 = distinct !{!10, !11, !"_ZN7ndarray17impl_constructors54_$LT$impl$u20$ndarray..ArrayBase$LT$S$C$$u20$D$GT$$GT$5zeros17h6cd29673e53e99b7E: argument 0"}
!11 = distinct !{!11, !"_ZN7ndarray17impl_constructors54_$LT$impl$u20$ndarray..ArrayBase$LT$S$C$$u20$D$GT$$GT$5zeros17h6cd29673e53e99b7E"}
!12 = !{!13, !15}
!13 = distinct !{!13, !14, !"_ZN4test9black_box17h9c12647707c851e2E: argument 0"}
!14 = distinct !{!14, !"_ZN4test9black_box17h9c12647707c851e2E"}
!15 = distinct !{!15, !14, !"_ZN4test9black_box17h9c12647707c851e2E: argument 1"}
!16 = !{i32 191}
!17 = distinct !{!17, !18, !19}
!18 = !{!"llvm.loop.vectorize.width", i32 1}
!19 = !{!"llvm.loop.interleave.count", i32 1}
!20 = distinct !{!20, !21, !18, !19}
!21 = !{!"llvm.loop.unroll.runtime.disable"}
!22 = !{!23, !25, !26, !28}
!23 = distinct !{!23, !24, !"_ZN85_$LT$ndarray..ArrayBase$LT$ndarray..ViewRepr$LT$$RF$$u27$a$u20$A$GT$$C$$u20$D$GT$$GT$10into_iter_17h07e424d86abf96f4E: argument 0"}
!24 = distinct !{!24, !"_ZN85_$LT$ndarray..ArrayBase$LT$ndarray..ViewRepr$LT$$RF$$u27$a$u20$A$GT$$C$$u20$D$GT$$GT$10into_iter_17h07e424d86abf96f4E"}
!25 = distinct !{!25, !24, !"_ZN85_$LT$ndarray..ArrayBase$LT$ndarray..ViewRepr$LT$$RF$$u27$a$u20$A$GT$$C$$u20$D$GT$$GT$10into_iter_17h07e424d86abf96f4E: argument 1"}
!26 = distinct !{!26, !27, !"_ZN7ndarray12impl_methods54_$LT$impl$u20$ndarray..ArrayBase$LT$S$C$$u20$D$GT$$GT$4iter17h92e629a76f9602eaE: argument 0"}
!27 = distinct !{!27, !"_ZN7ndarray12impl_methods54_$LT$impl$u20$ndarray..ArrayBase$LT$S$C$$u20$D$GT$$GT$4iter17h92e629a76f9602eaE"}
!28 = distinct !{!28, !27, !"_ZN7ndarray12impl_methods54_$LT$impl$u20$ndarray..ArrayBase$LT$S$C$$u20$D$GT$$GT$4iter17h92e629a76f9602eaE: argument 1"}
!29 = !{!26, !28}
!30 = !{!31, !33, !35, !23, !25, !26, !28}
!31 = distinct !{!31, !32, !"_ZN7ndarray12impl_methods54_$LT$impl$u20$ndarray..ArrayBase$LT$S$C$$u20$D$GT$$GT$18is_standard_layout17he690d19e37be350fE: argument 0"}
!32 = distinct !{!32, !"_ZN7ndarray12impl_methods54_$LT$impl$u20$ndarray..ArrayBase$LT$S$C$$u20$D$GT$$GT$18is_standard_layout17he690d19e37be350fE"}
!33 = distinct !{!33, !34, !"_ZN85_$LT$ndarray..ArrayBase$LT$ndarray..ViewRepr$LT$$RF$$u27$a$u20$A$GT$$C$$u20$D$GT$$GT$10into_slice17haf36f7fe3600206dE: argument 0"}
!34 = distinct !{!34, !"_ZN85_$LT$ndarray..ArrayBase$LT$ndarray..ViewRepr$LT$$RF$$u27$a$u20$A$GT$$C$$u20$D$GT$$GT$10into_slice17haf36f7fe3600206dE"}
!35 = distinct !{!35, !34, !"_ZN85_$LT$ndarray..ArrayBase$LT$ndarray..ViewRepr$LT$$RF$$u27$a$u20$A$GT$$C$$u20$D$GT$$GT$10into_slice17haf36f7fe3600206dE: argument 1"}
!36 = !{!37}
!37 = distinct !{!37, !38, !"_ZN76_$LT$$LP$usize$C$$u20$usize$RP$$u20$as$u20$ndarray..dimension..Dimension$GT$15default_strides17h9594c0827058adc3E: argument 0"}
!38 = distinct !{!38, !"_ZN76_$LT$$LP$usize$C$$u20$usize$RP$$u20$as$u20$ndarray..dimension..Dimension$GT$15default_strides17h9594c0827058adc3E"}
!39 = !{!31, !35}
!40 = !{!33, !23, !25, !26, !28}
!41 = distinct !{!41, !18, !19}
!42 = distinct !{!42, !21, !18, !19}

old-ir.ll

; ModuleID = 'bench1.cgu-0.rs'
source_filename = "bench1.cgu-0.rs"
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
target triple = "x86_64-unknown-linux-gnu"

%str_slice = type { i8*, i64 }
%"12.test::Bencher" = type { i64, %"1.std::time::Duration", i64 }
%"1.std::time::Duration" = type { i64, i32 }
%"15.ndarray::ArrayBase<ndarray::ViewRepr<&i32>, (usize, usize)>" = type { %"15.ndarray::ViewRepr<&i32>", i32*, { i64, i64 }, { i64, i64 } }
%"15.ndarray::ViewRepr<&i32>" = type { %"2.std::marker::PhantomData<&i32>" }
%"2.std::marker::PhantomData<&i32>" = type {}
%"1.std::time::Instant" = type { %"1.std::sys::time::inner::Instant" }
%"1.std::sys::time::inner::Instant" = type { %"1.std::sys::time::Timespec" }
%"1.std::sys::time::Timespec" = type { %"7.libc::unix::timespec" }
%"7.libc::unix::timespec" = type { i64, i64 }
%"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>" = type { %"3.std::vec::Vec<i32>", i32*, { i64, i64 }, { i64, i64 } }
%"3.std::vec::Vec<i32>" = type { %"5.alloc::raw_vec::RawVec<i32>", i64 }
%"5.alloc::raw_vec::RawVec<i32>" = type { %"2.std::ptr::Unique<i32>", i64 }
%"2.std::ptr::Unique<i32>" = type { %"2.core::nonzero::NonZero<*const i32>", %"2.std::marker::PhantomData<i32>" }
%"2.core::nonzero::NonZero<*const i32>" = type { i32* }
%"2.std::marker::PhantomData<i32>" = type {}
%"12.test::TestDescAndFn" = type { %"12.test::TestDesc", %"12.test::TestFn" }
%"12.test::TestDesc" = type { %"12.test::TestName", i8, %"12.test::ShouldPanic" }
%"12.test::TestName" = type { i64, [0 x i64], [3 x i64] }
%"12.test::ShouldPanic" = type { i64, [0 x i64], [2 x i64] }
%"12.test::TestFn" = type { i64, [0 x i64], [2 x i64] }
%"8.unwind::libunwind::_Unwind_Exception" = type { i64, void (i32, %"8.unwind::libunwind::_Unwind_Exception"*)*, [6 x i64] }
%"8.unwind::libunwind::_Unwind_Context" = type {}

@str10694 = internal constant [29 x i8] c"iter_sum_2d_transpose_regular"
@ref10695 = internal unnamed_addr constant { { { { i64, %str_slice, [8 x i8] }, i1, [7 x i8], { i64, [16 x i8] } }, { i64, void (%"12.test::Bencher"*)*, [8 x i8] } } } { { { { i64, %str_slice, [8 x i8] }, i1, [7 x i8], { i64, [16 x i8] } }, { i64, void (%"12.test::Bencher"*)*, [8 x i8] } } { { { i64, %str_slice, [8 x i8] }, i1, [7 x i8], { i64, [16 x i8] } } { { i64, %str_slice, [8 x i8] } { i64 0, %str_slice { i8* getelementptr inbounds ([29 x i8], [29 x i8]* @str10694, i32 0, i32 0), i64 29 }, [8 x i8] undef }, i1 false, [7 x i8] undef, { i64, [16 x i8] } { i64 0, [16 x i8] undef } }, { i64, void (%"12.test::Bencher"*)*, [8 x i8] } { i64 1, void (%"12.test::Bencher"*)* @_ZN6bench129iter_sum_2d_transpose_regular17hb420e27248fa1c24E, [8 x i8] undef } } }, align 8

; Function Attrs: uwtable
define internal void @_ZN6bench129iter_sum_2d_transpose_regular17hb420e27248fa1c24E(%"12.test::Bencher"* nocapture dereferenceable(32)) unnamed_addr #0 personality i32 (i32, i32, i64, %"8.unwind::libunwind::_Unwind_Exception"*, %"8.unwind::libunwind::_Unwind_Context"*)* @rust_eh_personality {
bb6.i.i.i.i.i:
  %defaults.i.i.i.i.i.i = alloca { i64, i64 }, align 8
  %self.i.i.i.i = alloca %"15.ndarray::ArrayBase<ndarray::ViewRepr<&i32>, (usize, usize)>", align 8
  %dummy.i.i = alloca i32, align 4
  %start1.i = alloca %"1.std::time::Instant", align 8
  %tmp_ret2.i = alloca %"1.std::time::Duration", align 8
  %dummy.i = alloca %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>", align 8
  %1 = tail call i8* @__rust_allocate(i64 16384, i64 4) #2, !noalias !1
  %2 = icmp eq i8* %1, null
  br i1 %2, label %bb10.i.i.i.i.i, label %bb4

bb10.i.i.i.i.i:                                   ; preds = %bb6.i.i.i.i.i
  invoke void @_ZN5alloc3oom3oom17h97560b93ca17f851E()
          to label %.noexc20 unwind label %cleanup

.noexc20:                                         ; preds = %bb10.i.i.i.i.i
  unreachable

bb1:                                              ; preds = %bb6.i.i.i.i.i29, %bb9, %cleanup
  %personalityslot.sroa.5.0 = phi i32 [ %.fca.1.extract, %cleanup ], [ %.fca.1.extract7, %bb9 ], [ %.fca.1.extract7, %bb6.i.i.i.i.i29 ]
  %personalityslot.sroa.0.0 = phi i8* [ %.fca.0.extract4, %cleanup ], [ %.fca.0.extract5, %bb9 ], [ %.fca.0.extract5, %bb6.i.i.i.i.i29 ]
  %.fca.0.insert = insertvalue { i8*, i32 } undef, i8* %personalityslot.sroa.0.0, 0
  %.fca.1.insert = insertvalue { i8*, i32 } %.fca.0.insert, i32 %personalityslot.sroa.5.0, 1
  resume { i8*, i32 } %.fca.1.insert

bb4:                                              ; preds = %bb6.i.i.i.i.i
  %3 = ptrtoint i8* %1 to i64
  %4 = bitcast %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>"* %dummy.i to i8*
  call void @llvm.memset.p0i8.i64(i8* nonnull %1, i8 0, i64 16384, i32 4, i1 false)
  call void @llvm.lifetime.start(i64 64, i8* %4) #2, !noalias !12
  %_9.sroa.0.0..sroa_cast = bitcast %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>"* %dummy.i to i64*
  store i64 %3, i64* %_9.sroa.0.0..sroa_cast, align 8
  %_9.sroa.4.0..sroa_idx70 = getelementptr inbounds %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>", %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>"* %dummy.i, i64 0, i32 0, i32 0, i32 1
  store i64 4096, i64* %_9.sroa.4.0..sroa_idx70, align 8
  %_9.sroa.5.0..sroa_idx72 = getelementptr inbounds %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>", %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>"* %dummy.i, i64 0, i32 0, i32 1
  store i64 4096, i64* %_9.sroa.5.0..sroa_idx72, align 8
  %_9.sroa.6.0..sroa_idx74 = getelementptr inbounds %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>", %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>"* %dummy.i, i64 0, i32 1
  %_9.sroa.6.0..sroa_cast = bitcast i32** %_9.sroa.6.0..sroa_idx74 to i64*
  store i64 %3, i64* %_9.sroa.6.0..sroa_cast, align 8
  %_9.sroa.7.0..sroa_idx76 = getelementptr inbounds %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>", %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>"* %dummy.i, i64 0, i32 2, i32 0
  %_9.sroa.8.0..sroa_idx78 = getelementptr inbounds %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>", %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>"* %dummy.i, i64 0, i32 2, i32 1
  %_9.sroa.9.0..sroa_idx80 = getelementptr inbounds %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>", %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>"* %dummy.i, i64 0, i32 3, i32 0
  %_9.sroa.10.0..sroa_idx82 = getelementptr inbounds %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>", %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>"* %dummy.i, i64 0, i32 3, i32 1
  %5 = bitcast i64* %_9.sroa.7.0..sroa_idx76 to <4 x i64>*
  store <4 x i64> <i64 64, i64 64, i64 1, i64 64>, <4 x i64>* %5, align 8
  call void asm "", "r,~{dirflag},~{fpsr},~{flags}"(%"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>"* nonnull %dummy.i) #2, !noalias !12, !srcloc !16
  %a1.sroa.0.0..sroa_cast34 = bitcast %"15.ndarray::ArrayBase<std::vec::Vec<i32>, (usize, usize)>"* %dummy.i to i8**
  %a1.sroa.0.0.copyload = load i8*, i8** %a1.sroa.0.0..sroa_cast34, align 8
  %a1.sroa.5.0.copyload = load i64, i64* %_9.sroa.4.0..sroa_idx70, align 8
  %a1.sroa.740.0.copyload = load i64, i64* %_9.sroa.6.0..sroa_cast, align 8
  %a1.sroa.8.0.copyload = load i64, i64* %_9.sroa.7.0..sroa_idx76, align 8
  %a1.sroa.9.0.copyload = load i64, i64* %_9.sroa.8.0..sroa_idx78, align 8
  %a1.sroa.10.0.copyload = load i64, i64* %_9.sroa.9.0..sroa_idx80, align 8
  %a1.sroa.11.0.copyload = load i64, i64* %_9.sroa.10.0..sroa_idx82, align 8
  call void @llvm.lifetime.end(i64 64, i8* %4) #2, !noalias !12
  %6 = bitcast %"1.std::time::Duration"* %tmp_ret2.i to i8*
  call void @llvm.lifetime.start(i64 16, i8* %6)
  %7 = bitcast %"1.std::time::Instant"* %start1.i to i8*
  call void @llvm.lifetime.start(i64 16, i8* %7)
  invoke void @_ZN3std4time7Instant3now17h5851ec34f9ff4cdfE(%"1.std::time::Instant"* noalias nocapture nonnull sret dereferenceable(16) %start1.i)
          to label %.noexc22 unwind label %bb9

.noexc22:                                         ; preds = %bb4
  %8 = getelementptr inbounds %"12.test::Bencher", %"12.test::Bencher"* %0, i64 0, i32 0
  %9 = load i64, i64* %8, align 8
  %10 = icmp eq i64 %9, 0
  br i1 %10, label %bb8.i, label %bb7.lr.ph.i

bb7.lr.ph.i:                                      ; preds = %.noexc22
  %11 = bitcast %"15.ndarray::ArrayBase<ndarray::ViewRepr<&i32>, (usize, usize)>"* %self.i.i.i.i to i8*
  %_11.sroa.0.0..sroa_cast7.i.i.i = bitcast %"15.ndarray::ArrayBase<ndarray::ViewRepr<&i32>, (usize, usize)>"* %self.i.i.i.i to i64*
  %_11.sroa.4.0..sroa_idx9.i.i.i = getelementptr inbounds %"15.ndarray::ArrayBase<ndarray::ViewRepr<&i32>, (usize, usize)>", %"15.ndarray::ArrayBase<ndarray::ViewRepr<&i32>, (usize, usize)>"* %self.i.i.i.i, i64 0, i32 2, i32 0
  %_11.sroa.5.0..sroa_idx11.i.i.i = getelementptr inbounds %"15.ndarray::ArrayBase<ndarray::ViewRepr<&i32>, (usize, usize)>", %"15.ndarray::ArrayBase<ndarray::ViewRepr<&i32>, (usize, usize)>"* %self.i.i.i.i, i64 0, i32 2, i32 1
  %_11.sroa.6.0..sroa_idx13.i.i.i = getelementptr inbounds %"15.ndarray::ArrayBase<ndarray::ViewRepr<&i32>, (usize, usize)>", %"15.ndarray::ArrayBase<ndarray::ViewRepr<&i32>, (usize, usize)>"* %self.i.i.i.i, i64 0, i32 3, i32 0
  %_11.sroa.7.0..sroa_idx15.i.i.i = getelementptr inbounds %"15.ndarray::ArrayBase<ndarray::ViewRepr<&i32>, (usize, usize)>", %"15.ndarray::ArrayBase<ndarray::ViewRepr<&i32>, (usize, usize)>"* %self.i.i.i.i, i64 0, i32 3, i32 1
  %12 = bitcast { i64, i64 }* %defaults.i.i.i.i.i.i to i8*
  %13 = getelementptr inbounds { i64, i64 }, { i64, i64 }* %defaults.i.i.i.i.i.i, i64 0, i32 0
  %14 = getelementptr inbounds { i64, i64 }, { i64, i64 }* %defaults.i.i.i.i.i.i, i64 0, i32 1
  %15 = bitcast i32* %dummy.i.i to i8*
  %16 = icmp eq i64 %a1.sroa.10.0.copyload, %a1.sroa.9.0.copyload
  %17 = icmp eq i64 %a1.sroa.11.0.copyload, 1
  %or.cond.i.i.i.i.i.i.i = and i1 %16, %17
  %18 = inttoptr i64 %a1.sroa.740.0.copyload to i32*
  %switchtmp.i.i.i.i = icmp eq i64 %a1.sroa.740.0.copyload, 0
  %19 = mul i64 %a1.sroa.9.0.copyload, %a1.sroa.8.0.copyload
  %20 = getelementptr inbounds i32, i32* %18, i64 %19
  %21 = icmp eq i64 %19, 0
  %22 = icmp ne i64 %a1.sroa.8.0.copyload, 0
  %23 = icmp ne i64 %a1.sroa.9.0.copyload, 0
  %or.cond.i.i.i.i.i.i.i.i = and i1 %22, %23
  %24 = mul i64 %a1.sroa.9.0.copyload, %a1.sroa.8.0.copyload
  %25 = shl i64 %24, 2
  %26 = add i64 %25, -4
  %27 = lshr exact i64 %26, 2
  %28 = add nuw nsw i64 %27, 1
  br i1 %or.cond.i.i.i.i.i.i.i, label %bb7.i.us.preheader, label %bb7.i.preheader

bb7.i.preheader:                                  ; preds = %bb7.lr.ph.i
  %29 = and i64 %28, 9223372036854775792
  %30 = add nsw i64 %29, -16
  %31 = lshr exact i64 %30, 4
  %32 = add nuw nsw i64 %31, 1
  %min.iters.check = icmp ult i64 %28, 16
  %n.vec = and i64 %28, 9223372036854775792
  %cmp.zero = icmp eq i64 %n.vec, 0
  %ind.end = getelementptr i32, i32* %18, i64 %n.vec
  %xtraiter216 = and i64 %32, 1
  %33 = icmp eq i64 %31, 0
  %lcmp.mod217 = icmp eq i64 %xtraiter216, 0
  %cmp.n = icmp eq i64 %28, %n.vec
  %unroll_iter218 = sub nsw i64 %32, %xtraiter216
  br label %bb7.i

bb7.i.us.preheader:                               ; preds = %bb7.lr.ph.i
  %34 = and i64 %28, 9223372036854775792
  %35 = add nsw i64 %34, -16
  %36 = lshr exact i64 %35, 4
  %37 = add nuw nsw i64 %36, 1
  %min.iters.check141 = icmp ult i64 %28, 16
  %n.vec144 = and i64 %28, 9223372036854775792
  %cmp.zero145 = icmp eq i64 %n.vec144, 0
  %ind.end150 = getelementptr i32, i32* %18, i64 %n.vec144
  %xtraiter = and i64 %37, 1
  %38 = icmp eq i64 %36, 0
  %lcmp.mod = icmp eq i64 %xtraiter, 0
  %cmp.n151 = icmp eq i64 %28, %n.vec144
  %unroll_iter = sub nsw i64 %37, %xtraiter
  br label %bb7.i.us

bb7.i.us:                                         ; preds = %bb7.i.us.preheader, %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i.us"
  %iter.sroa.0.036.i.us = phi i64 [ %39, %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i.us" ], [ 0, %bb7.i.us.preheader ]
  %39 = add nuw i64 %iter.sroa.0.036.i.us, 1
  br i1 %switchtmp.i.i.i.i, label %bb3.us.preheader.i.i.us, label %bb3.preheader.i.i.us

bb3.preheader.i.i.us:                             ; preds = %bb7.i.us
  br i1 %21, label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i.us", label %bb6.i.i21.us.preheader

bb6.i.i21.us.preheader:                           ; preds = %bb3.preheader.i.i.us
  br i1 %min.iters.check141, label %bb6.i.i21.us.preheader202, label %min.iters.checked142

min.iters.checked142:                             ; preds = %bb6.i.i21.us.preheader
  br i1 %cmp.zero145, label %bb6.i.i21.us.preheader202, label %vector.body138.preheader

vector.body138.preheader:                         ; preds = %min.iters.checked142
  br i1 %38, label %middle.block139.unr-lcssa, label %vector.body138.preheader.new

vector.body138.preheader.new:                     ; preds = %vector.body138.preheader
  br label %vector.body138

vector.body138:                                   ; preds = %vector.body138, %vector.body138.preheader.new
  %index147 = phi i64 [ 0, %vector.body138.preheader.new ], [ %index.next148.1, %vector.body138 ]
  %vec.phi152 = phi <4 x i32> [ zeroinitializer, %vector.body138.preheader.new ], [ %58, %vector.body138 ]
  %vec.phi153 = phi <4 x i32> [ zeroinitializer, %vector.body138.preheader.new ], [ %59, %vector.body138 ]
  %vec.phi154 = phi <4 x i32> [ zeroinitializer, %vector.body138.preheader.new ], [ %60, %vector.body138 ]
  %vec.phi155 = phi <4 x i32> [ zeroinitializer, %vector.body138.preheader.new ], [ %61, %vector.body138 ]
  %niter = phi i64 [ %unroll_iter, %vector.body138.preheader.new ], [ %niter.nsub.1, %vector.body138 ]
  %next.gep156 = getelementptr i32, i32* %18, i64 %index147
  %40 = bitcast i32* %next.gep156 to <4 x i32>*
  %wide.load188 = load <4 x i32>, <4 x i32>* %40, align 4
  %41 = getelementptr i32, i32* %next.gep156, i64 4
  %42 = bitcast i32* %41 to <4 x i32>*
  %wide.load189 = load <4 x i32>, <4 x i32>* %42, align 4
  %43 = getelementptr i32, i32* %next.gep156, i64 8
  %44 = bitcast i32* %43 to <4 x i32>*
  %wide.load190 = load <4 x i32>, <4 x i32>* %44, align 4
  %45 = getelementptr i32, i32* %next.gep156, i64 12
  %46 = bitcast i32* %45 to <4 x i32>*
  %wide.load191 = load <4 x i32>, <4 x i32>* %46, align 4
  %47 = add <4 x i32> %wide.load188, %vec.phi152
  %48 = add <4 x i32> %wide.load189, %vec.phi153
  %49 = add <4 x i32> %wide.load190, %vec.phi154
  %50 = add <4 x i32> %wide.load191, %vec.phi155
  %index.next148 = or i64 %index147, 16
  %next.gep156.1 = getelementptr i32, i32* %18, i64 %index.next148
  %51 = bitcast i32* %next.gep156.1 to <4 x i32>*
  %wide.load188.1 = load <4 x i32>, <4 x i32>* %51, align 4
  %52 = getelementptr i32, i32* %next.gep156.1, i64 4
  %53 = bitcast i32* %52 to <4 x i32>*
  %wide.load189.1 = load <4 x i32>, <4 x i32>* %53, align 4
  %54 = getelementptr i32, i32* %next.gep156.1, i64 8
  %55 = bitcast i32* %54 to <4 x i32>*
  %wide.load190.1 = load <4 x i32>, <4 x i32>* %55, align 4
  %56 = getelementptr i32, i32* %next.gep156.1, i64 12
  %57 = bitcast i32* %56 to <4 x i32>*
  %wide.load191.1 = load <4 x i32>, <4 x i32>* %57, align 4
  %58 = add <4 x i32> %wide.load188.1, %47
  %59 = add <4 x i32> %wide.load189.1, %48
  %60 = add <4 x i32> %wide.load190.1, %49
  %61 = add <4 x i32> %wide.load191.1, %50
  %index.next148.1 = add i64 %index147, 32
  %niter.nsub.1 = add i64 %niter, -2
  %niter.ncmp.1 = icmp eq i64 %niter.nsub.1, 0
  br i1 %niter.ncmp.1, label %middle.block139.unr-lcssa.loopexit, label %vector.body138, !llvm.loop !17

middle.block139.unr-lcssa.loopexit:               ; preds = %vector.body138
  br label %middle.block139.unr-lcssa

middle.block139.unr-lcssa:                        ; preds = %middle.block139.unr-lcssa.loopexit, %vector.body138.preheader
  %.lcssa209.ph = phi <4 x i32> [ undef, %vector.body138.preheader ], [ %58, %middle.block139.unr-lcssa.loopexit ]
  %.lcssa208.ph = phi <4 x i32> [ undef, %vector.body138.preheader ], [ %59, %middle.block139.unr-lcssa.loopexit ]
  %.lcssa207.ph = phi <4 x i32> [ undef, %vector.body138.preheader ], [ %60, %middle.block139.unr-lcssa.loopexit ]
  %.lcssa.ph = phi <4 x i32> [ undef, %vector.body138.preheader ], [ %61, %middle.block139.unr-lcssa.loopexit ]
  %index147.unr = phi i64 [ 0, %vector.body138.preheader ], [ %index.next148.1, %middle.block139.unr-lcssa.loopexit ]
  %vec.phi152.unr = phi <4 x i32> [ zeroinitializer, %vector.body138.preheader ], [ %58, %middle.block139.unr-lcssa.loopexit ]
  %vec.phi153.unr = phi <4 x i32> [ zeroinitializer, %vector.body138.preheader ], [ %59, %middle.block139.unr-lcssa.loopexit ]
  %vec.phi154.unr = phi <4 x i32> [ zeroinitializer, %vector.body138.preheader ], [ %60, %middle.block139.unr-lcssa.loopexit ]
  %vec.phi155.unr = phi <4 x i32> [ zeroinitializer, %vector.body138.preheader ], [ %61, %middle.block139.unr-lcssa.loopexit ]
  br i1 %lcmp.mod, label %middle.block139, label %vector.body138.epil.preheader

vector.body138.epil.preheader:                    ; preds = %middle.block139.unr-lcssa
  br label %vector.body138.epil

vector.body138.epil:                              ; preds = %vector.body138.epil.preheader
  %next.gep156.epil = getelementptr i32, i32* %18, i64 %index147.unr
  br label %middle.block139.epilog-lcssa

middle.block139.epilog-lcssa:                     ; preds = %vector.body138.epil
  %62 = getelementptr i32, i32* %next.gep156.epil, i64 12
  %63 = bitcast i32* %62 to <4 x i32>*
  %wide.load191.epil = load <4 x i32>, <4 x i32>* %63, align 4
  %64 = add <4 x i32> %wide.load191.epil, %vec.phi155.unr
  %65 = getelementptr i32, i32* %next.gep156.epil, i64 8
  %66 = bitcast i32* %65 to <4 x i32>*
  %wide.load190.epil = load <4 x i32>, <4 x i32>* %66, align 4
  %67 = add <4 x i32> %wide.load190.epil, %vec.phi154.unr
  %68 = getelementptr i32, i32* %next.gep156.epil, i64 4
  %69 = bitcast i32* %68 to <4 x i32>*
  %wide.load189.epil = load <4 x i32>, <4 x i32>* %69, align 4
  %70 = add <4 x i32> %wide.load189.epil, %vec.phi153.unr
  %71 = bitcast i32* %next.gep156.epil to <4 x i32>*
  %wide.load188.epil = load <4 x i32>, <4 x i32>* %71, align 4
  %72 = add <4 x i32> %wide.load188.epil, %vec.phi152.unr
  br label %middle.block139

middle.block139:                                  ; preds = %middle.block139.unr-lcssa, %middle.block139.epilog-lcssa
  %.lcssa209 = phi <4 x i32> [ %.lcssa209.ph, %middle.block139.unr-lcssa ], [ %72, %middle.block139.epilog-lcssa ]
  %.lcssa208 = phi <4 x i32> [ %.lcssa208.ph, %middle.block139.unr-lcssa ], [ %70, %middle.block139.epilog-lcssa ]
  %.lcssa207 = phi <4 x i32> [ %.lcssa207.ph, %middle.block139.unr-lcssa ], [ %67, %middle.block139.epilog-lcssa ]
  %.lcssa = phi <4 x i32> [ %.lcssa.ph, %middle.block139.unr-lcssa ], [ %64, %middle.block139.epilog-lcssa ]
  %bin.rdx194 = add <4 x i32> %.lcssa208, %.lcssa209
  %bin.rdx195 = add <4 x i32> %.lcssa207, %bin.rdx194
  %bin.rdx196 = add <4 x i32> %.lcssa, %bin.rdx195
  %rdx.shuf197 = shufflevector <4 x i32> %bin.rdx196, <4 x i32> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
  %bin.rdx198 = add <4 x i32> %bin.rdx196, %rdx.shuf197
  %rdx.shuf199 = shufflevector <4 x i32> %bin.rdx198, <4 x i32> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
  %bin.rdx200 = add <4 x i32> %bin.rdx198, %rdx.shuf199
  %73 = extractelement <4 x i32> %bin.rdx200, i32 0
  br i1 %cmp.n151, label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i.us", label %bb6.i.i21.us.preheader202

bb6.i.i21.us.preheader202:                        ; preds = %middle.block139, %min.iters.checked142, %bb6.i.i21.us.preheader
  %sum.0.i29.i.us.ph = phi i32 [ 0, %min.iters.checked142 ], [ 0, %bb6.i.i21.us.preheader ], [ %73, %middle.block139 ]
  %iter.sroa.4.0.i28.i.us.ph = phi i32* [ %18, %min.iters.checked142 ], [ %18, %bb6.i.i21.us.preheader ], [ %ind.end150, %middle.block139 ]
  br label %bb6.i.i21.us

bb6.i.i21.us:                                     ; preds = %bb6.i.i21.us.preheader202, %bb6.i.i21.us
  %sum.0.i29.i.us = phi i32 [ %76, %bb6.i.i21.us ], [ %sum.0.i29.i.us.ph, %bb6.i.i21.us.preheader202 ]
  %iter.sroa.4.0.i28.i.us = phi i32* [ %74, %bb6.i.i21.us ], [ %iter.sroa.4.0.i28.i.us.ph, %bb6.i.i21.us.preheader202 ]
  %74 = getelementptr inbounds i32, i32* %iter.sroa.4.0.i28.i.us, i64 1
  %75 = load i32, i32* %iter.sroa.4.0.i28.i.us, align 4
  %76 = add i32 %75, %sum.0.i29.i.us
  %77 = icmp eq i32* %74, %20
  br i1 %77, label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i.us.loopexit203", label %bb6.i.i21.us, !llvm.loop !20

bb3.us.preheader.i.i.us:                          ; preds = %bb7.i.us
  br i1 %or.cond.i.i.i.i.i.i.i.i, label %bb4.i.i.i.i.us.i.i.us.preheader, label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i.us"

bb4.i.i.i.i.us.i.i.us.preheader:                  ; preds = %bb3.us.preheader.i.i.us
  br label %bb4.i.i.i.i.us.i.i.us

bb4.i.i.i.i.us.i.i.us:                            ; preds = %bb4.i.i.i.i.us.i.i.us.preheader, %bb6.us.i.i.us
  %sum.0.us.i33.i.us = phi i32 [ %86, %bb6.us.i.i.us ], [ 0, %bb4.i.i.i.i.us.i.i.us.preheader ]
  %iter.sroa.14.0.us.i32.i.us = phi i64 [ %_13.sroa.5.0.i.i.i.i.us.i.i.us, %bb6.us.i.i.us ], [ 0, %bb4.i.i.i.i.us.i.i.us.preheader ]
  %iter.sroa.16.0.us.i31.i.us = phi i64 [ %_13.sroa.6.0.i.i.i.i.us.i.i.us, %bb6.us.i.i.us ], [ 0, %bb4.i.i.i.i.us.i.i.us.preheader ]
  %78 = mul i64 %iter.sroa.14.0.us.i32.i.us, %a1.sroa.9.0.copyload
  %79 = add i64 %iter.sroa.16.0.us.i31.i.us, %78
  %80 = add i64 %iter.sroa.16.0.us.i31.i.us, 1
  %81 = icmp eq i64 %80, %a1.sroa.9.0.copyload
  br i1 %81, label %bb1.i.i.i.i.i.us.i.i.us, label %bb6.i.i.i.i.i.us.i.i.us

bb1.i.i.i.i.i.us.i.i.us:                          ; preds = %bb4.i.i.i.i.us.i.i.us
  %82 = add i64 %iter.sroa.14.0.us.i32.i.us, 1
  %83 = icmp eq i64 %82, %a1.sroa.8.0.copyload
  br i1 %83, label %"_ZN76_$LT$$LP$usize$C$$u20$usize$RP$$u20$as$u20$ndarray..dimension..Dimension$GT$8next_for17h04b0fe300a278d57E.exit.i.i.i.i.us.i.i.us", label %bb6.i.i.i.i.i.us.i.i.us

bb6.i.i.i.i.i.us.i.i.us:                          ; preds = %bb1.i.i.i.i.i.us.i.i.us, %bb4.i.i.i.i.us.i.i.us
  %j.0.i.i.i.i.i.us.i.i.us = phi i64 [ 0, %bb1.i.i.i.i.i.us.i.i.us ], [ %80, %bb4.i.i.i.i.us.i.i.us ]
  %i.0.i.i.i.i.i.us.i.i.us = phi i64 [ %82, %bb1.i.i.i.i.i.us.i.i.us ], [ %iter.sroa.14.0.us.i32.i.us, %bb4.i.i.i.i.us.i.i.us ]
  br label %"_ZN76_$LT$$LP$usize$C$$u20$usize$RP$$u20$as$u20$ndarray..dimension..Dimension$GT$8next_for17h04b0fe300a278d57E.exit.i.i.i.i.us.i.i.us"

"_ZN76_$LT$$LP$usize$C$$u20$usize$RP$$u20$as$u20$ndarray..dimension..Dimension$GT$8next_for17h04b0fe300a278d57E.exit.i.i.i.i.us.i.i.us": ; preds = %bb6.i.i.i.i.i.us.i.i.us, %bb1.i.i.i.i.i.us.i.i.us
  %_13.sroa.0.0.i.i.i.i.us.i.i.us = phi i64 [ 1, %bb6.i.i.i.i.i.us.i.i.us ], [ 0, %bb1.i.i.i.i.i.us.i.i.us ]
  %_13.sroa.5.0.i.i.i.i.us.i.i.us = phi i64 [ %i.0.i.i.i.i.i.us.i.i.us, %bb6.i.i.i.i.i.us.i.i.us ], [ undef, %bb1.i.i.i.i.i.us.i.i.us ]
  %_13.sroa.6.0.i.i.i.i.us.i.i.us = phi i64 [ %j.0.i.i.i.i.i.us.i.i.us, %bb6.i.i.i.i.i.us.i.i.us ], [ undef, %bb1.i.i.i.i.i.us.i.i.us ]
  %84 = getelementptr inbounds i32, i32* %18, i64 %79
  %switchtmp.us.i.i.us = icmp eq i32* %84, null
  br i1 %switchtmp.us.i.i.us, label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i.us.loopexit", label %bb6.us.i.i.us

bb6.us.i.i.us:                                    ; preds = %"_ZN76_$LT$$LP$usize$C$$u20$usize$RP$$u20$as$u20$ndarray..dimension..Dimension$GT$8next_for17h04b0fe300a278d57E.exit.i.i.i.i.us.i.i.us"
  %85 = load i32, i32* %84, align 4
  %86 = add i32 %85, %sum.0.us.i33.i.us
  %switch.i.i.i.i.us.i.i.us = icmp eq i64 %_13.sroa.0.0.i.i.i.i.us.i.i.us, 1
  br i1 %switch.i.i.i.i.us.i.i.us, label %bb4.i.i.i.i.us.i.i.us, label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i.us.loopexit"

"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i.us.loopexit": ; preds = %bb6.us.i.i.us, %"_ZN76_$LT$$LP$usize$C$$u20$usize$RP$$u20$as$u20$ndarray..dimension..Dimension$GT$8next_for17h04b0fe300a278d57E.exit.i.i.i.i.us.i.i.us"
  %sum.0.lcssa.i.i.us.ph = phi i32 [ %86, %bb6.us.i.i.us ], [ %sum.0.us.i33.i.us, %"_ZN76_$LT$$LP$usize$C$$u20$usize$RP$$u20$as$u20$ndarray..dimension..Dimension$GT$8next_for17h04b0fe300a278d57E.exit.i.i.i.i.us.i.i.us" ]
  br label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i.us"

"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i.us.loopexit203": ; preds = %bb6.i.i21.us
  br label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i.us"

"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i.us": ; preds = %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i.us.loopexit203", %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i.us.loopexit", %middle.block139, %bb3.us.preheader.i.i.us, %bb3.preheader.i.i.us
  %sum.0.lcssa.i.i.us = phi i32 [ 0, %bb3.us.preheader.i.i.us ], [ 0, %bb3.preheader.i.i.us ], [ %73, %middle.block139 ], [ %sum.0.lcssa.i.i.us.ph, %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i.us.loopexit" ], [ %76, %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i.us.loopexit203" ]
  call void @llvm.lifetime.start(i64 4, i8* %15) #2
  store i32 %sum.0.lcssa.i.i.us, i32* %dummy.i.i, align 4
  call void asm "", "r,~{dirflag},~{fpsr},~{flags}"(i32* nonnull %dummy.i.i) #2, !srcloc !16
  call void @llvm.lifetime.end(i64 4, i8* %15) #2
  %exitcond.i.us = icmp eq i64 %39, %9
  br i1 %exitcond.i.us, label %bb8.i.loopexit, label %bb7.i.us

bb7.i:                                            ; preds = %bb7.i.preheader, %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i"
  %iter.sroa.0.036.i = phi i64 [ %87, %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i" ], [ 0, %bb7.i.preheader ]
  %87 = add nuw i64 %iter.sroa.0.036.i, 1
  call void @llvm.lifetime.start(i64 40, i8* %11), !noalias !22
  store i64 %a1.sroa.740.0.copyload, i64* %_11.sroa.0.0..sroa_cast7.i.i.i, align 8, !noalias !29
  store i64 %a1.sroa.8.0.copyload, i64* %_11.sroa.4.0..sroa_idx9.i.i.i, align 8, !noalias !29
  store i64 %a1.sroa.9.0.copyload, i64* %_11.sroa.5.0..sroa_idx11.i.i.i, align 8, !noalias !29
  store i64 %a1.sroa.10.0.copyload, i64* %_11.sroa.6.0..sroa_idx13.i.i.i, align 8, !noalias !29
  store i64 %a1.sroa.11.0.copyload, i64* %_11.sroa.7.0..sroa_idx15.i.i.i, align 8, !noalias !29
  call void @llvm.lifetime.start(i64 16, i8* %12), !noalias !30
  store i64 %a1.sroa.9.0.copyload, i64* %13, align 8, !alias.scope !36, !noalias !30
  store i64 1, i64* %14, align 8, !alias.scope !36, !noalias !30
  br label %bb18.i.i.i.i.i.i

bb18.i.i.i.i.i.i:                                 ; preds = %bb7.i, %"_ZN145_$LT$itertools..cons_tuples_impl..ConsTuples$LT$Iter$C$$u20$$LP$$LP$G$C$$u20$H$RP$$C$$u20$X$RP$$GT$$u20$as$u20$core..iter..iterator..Iterator$GT$4next17hef7423d175aa3a15E.exit.i.i.i.i.i.i"
  %iter.sroa.738.0.i.i.i.i.i.i = phi i64 [ %92, %"_ZN145_$LT$itertools..cons_tuples_impl..ConsTuples$LT$Iter$C$$u20$$LP$$LP$G$C$$u20$H$RP$$C$$u20$X$RP$$GT$$u20$as$u20$core..iter..iterator..Iterator$GT$4next17hef7423d175aa3a15E.exit.i.i.i.i.i.i" ], [ 0, %bb7.i ]
  %88 = icmp ult i64 %iter.sroa.738.0.i.i.i.i.i.i, 2
  br i1 %88, label %"_ZN145_$LT$itertools..cons_tuples_impl..ConsTuples$LT$Iter$C$$u20$$LP$$LP$G$C$$u20$H$RP$$C$$u20$X$RP$$GT$$u20$as$u20$core..iter..iterator..Iterator$GT$4next17hef7423d175aa3a15E.exit.i.i.i.i.i.i", label %"_ZN7ndarray12impl_methods54_$LT$impl$u20$ndarray..ArrayBase$LT$S$C$$u20$D$GT$$GT$18is_standard_layout17he690d19e37be350fE.exit.thread.i.i.i.i.i"

"_ZN145_$LT$itertools..cons_tuples_impl..ConsTuples$LT$Iter$C$$u20$$LP$$LP$G$C$$u20$H$RP$$C$$u20$X$RP$$GT$$u20$as$u20$core..iter..iterator..Iterator$GT$4next17hef7423d175aa3a15E.exit.i.i.i.i.i.i": ; preds = %bb18.i.i.i.i.i.i
  %89 = getelementptr inbounds i64, i64* %_11.sroa.4.0..sroa_idx9.i.i.i, i64 %iter.sroa.738.0.i.i.i.i.i.i
  %90 = getelementptr inbounds i64, i64* %_11.sroa.6.0..sroa_idx13.i.i.i, i64 %iter.sroa.738.0.i.i.i.i.i.i
  %91 = getelementptr inbounds i64, i64* %13, i64 %iter.sroa.738.0.i.i.i.i.i.i
  %92 = add nuw nsw i64 %iter.sroa.738.0.i.i.i.i.i.i, 1
  %93 = load i64, i64* %89, align 8, !alias.scope !39, !noalias !40
  %94 = load i64, i64* %90, align 8, !alias.scope !39, !noalias !40
  %95 = load i64, i64* %91, align 8, !noalias !30
  %96 = icmp ne i64 %93, 1
  %97 = icmp ne i64 %94, %95
  %or.cond.i.i.i.i.i.i = and i1 %96, %97
  br i1 %or.cond.i.i.i.i.i.i, label %"_ZN85_$LT$ndarray..ArrayBase$LT$ndarray..ViewRepr$LT$$RF$$u27$a$u20$A$GT$$C$$u20$D$GT$$GT$10into_slice17hbd80c1dc13b299edE.exit.thread.i.i.i.i", label %bb18.i.i.i.i.i.i

"_ZN7ndarray12impl_methods54_$LT$impl$u20$ndarray..ArrayBase$LT$S$C$$u20$D$GT$$GT$18is_standard_layout17he690d19e37be350fE.exit.thread.i.i.i.i.i": ; preds = %bb18.i.i.i.i.i.i
  call void @llvm.lifetime.end(i64 16, i8* nonnull %12), !noalias !30
  br i1 %switchtmp.i.i.i.i, label %bb3.us.preheader.i.i, label %bb3.preheader.i.i

"_ZN85_$LT$ndarray..ArrayBase$LT$ndarray..ViewRepr$LT$$RF$$u27$a$u20$A$GT$$C$$u20$D$GT$$GT$10into_slice17hbd80c1dc13b299edE.exit.thread.i.i.i.i": ; preds = %"_ZN145_$LT$itertools..cons_tuples_impl..ConsTuples$LT$Iter$C$$u20$$LP$$LP$G$C$$u20$H$RP$$C$$u20$X$RP$$GT$$u20$as$u20$core..iter..iterator..Iterator$GT$4next17hef7423d175aa3a15E.exit.i.i.i.i.i.i"
  call void @llvm.lifetime.end(i64 16, i8* nonnull %12), !noalias !30
  br label %bb3.us.preheader.i.i

bb3.preheader.i.i:                                ; preds = %"_ZN7ndarray12impl_methods54_$LT$impl$u20$ndarray..ArrayBase$LT$S$C$$u20$D$GT$$GT$18is_standard_layout17he690d19e37be350fE.exit.thread.i.i.i.i.i"
  call void @llvm.lifetime.end(i64 40, i8* nonnull %11), !noalias !22
  br i1 %21, label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i", label %bb6.i.i21.preheader

bb6.i.i21.preheader:                              ; preds = %bb3.preheader.i.i
  br i1 %min.iters.check, label %bb6.i.i21.preheader204, label %min.iters.checked

bb6.i.i21.preheader204:                           ; preds = %middle.block, %min.iters.checked, %bb6.i.i21.preheader
  %sum.0.i29.i.ph = phi i32 [ 0, %min.iters.checked ], [ 0, %bb6.i.i21.preheader ], [ %131, %middle.block ]
  %iter.sroa.4.0.i28.i.ph = phi i32* [ %18, %min.iters.checked ], [ %18, %bb6.i.i21.preheader ], [ %ind.end, %middle.block ]
  br label %bb6.i.i21

min.iters.checked:                                ; preds = %bb6.i.i21.preheader
  br i1 %cmp.zero, label %bb6.i.i21.preheader204, label %vector.body.preheader

vector.body.preheader:                            ; preds = %min.iters.checked
  br i1 %33, label %middle.block.unr-lcssa, label %vector.body.preheader.new

vector.body.preheader.new:                        ; preds = %vector.body.preheader
  br label %vector.body

vector.body:                                      ; preds = %vector.body, %vector.body.preheader.new
  %index = phi i64 [ 0, %vector.body.preheader.new ], [ %index.next.1, %vector.body ]
  %vec.phi = phi <4 x i32> [ zeroinitializer, %vector.body.preheader.new ], [ %116, %vector.body ]
  %vec.phi97 = phi <4 x i32> [ zeroinitializer, %vector.body.preheader.new ], [ %117, %vector.body ]
  %vec.phi98 = phi <4 x i32> [ zeroinitializer, %vector.body.preheader.new ], [ %118, %vector.body ]
  %vec.phi99 = phi <4 x i32> [ zeroinitializer, %vector.body.preheader.new ], [ %119, %vector.body ]
  %niter219 = phi i64 [ %unroll_iter218, %vector.body.preheader.new ], [ %niter219.nsub.1, %vector.body ]
  %next.gep = getelementptr i32, i32* %18, i64 %index
  %98 = bitcast i32* %next.gep to <4 x i32>*
  %wide.load = load <4 x i32>, <4 x i32>* %98, align 4
  %99 = getelementptr i32, i32* %next.gep, i64 4
  %100 = bitcast i32* %99 to <4 x i32>*
  %wide.load130 = load <4 x i32>, <4 x i32>* %100, align 4
  %101 = getelementptr i32, i32* %next.gep, i64 8
  %102 = bitcast i32* %101 to <4 x i32>*
  %wide.load131 = load <4 x i32>, <4 x i32>* %102, align 4
  %103 = getelementptr i32, i32* %next.gep, i64 12
  %104 = bitcast i32* %103 to <4 x i32>*
  %wide.load132 = load <4 x i32>, <4 x i32>* %104, align 4
  %105 = add <4 x i32> %wide.load, %vec.phi
  %106 = add <4 x i32> %wide.load130, %vec.phi97
  %107 = add <4 x i32> %wide.load131, %vec.phi98
  %108 = add <4 x i32> %wide.load132, %vec.phi99
  %index.next = or i64 %index, 16
  %next.gep.1 = getelementptr i32, i32* %18, i64 %index.next
  %109 = bitcast i32* %next.gep.1 to <4 x i32>*
  %wide.load.1 = load <4 x i32>, <4 x i32>* %109, align 4
  %110 = getelementptr i32, i32* %next.gep.1, i64 4
  %111 = bitcast i32* %110 to <4 x i32>*
  %wide.load130.1 = load <4 x i32>, <4 x i32>* %111, align 4
  %112 = getelementptr i32, i32* %next.gep.1, i64 8
  %113 = bitcast i32* %112 to <4 x i32>*
  %wide.load131.1 = load <4 x i32>, <4 x i32>* %113, align 4
  %114 = getelementptr i32, i32* %next.gep.1, i64 12
  %115 = bitcast i32* %114 to <4 x i32>*
  %wide.load132.1 = load <4 x i32>, <4 x i32>* %115, align 4
  %116 = add <4 x i32> %wide.load.1, %105
  %117 = add <4 x i32> %wide.load130.1, %106
  %118 = add <4 x i32> %wide.load131.1, %107
  %119 = add <4 x i32> %wide.load132.1, %108
  %index.next.1 = add i64 %index, 32
  %niter219.nsub.1 = add i64 %niter219, -2
  %niter219.ncmp.1 = icmp eq i64 %niter219.nsub.1, 0
  br i1 %niter219.ncmp.1, label %middle.block.unr-lcssa.loopexit, label %vector.body, !llvm.loop !41

middle.block.unr-lcssa.loopexit:                  ; preds = %vector.body
  br label %middle.block.unr-lcssa

middle.block.unr-lcssa:                           ; preds = %middle.block.unr-lcssa.loopexit, %vector.body.preheader
  %.lcssa214.ph = phi <4 x i32> [ undef, %vector.body.preheader ], [ %116, %middle.block.unr-lcssa.loopexit ]
  %.lcssa213.ph = phi <4 x i32> [ undef, %vector.body.preheader ], [ %117, %middle.block.unr-lcssa.loopexit ]
  %.lcssa212.ph = phi <4 x i32> [ undef, %vector.body.preheader ], [ %118, %middle.block.unr-lcssa.loopexit ]
  %.lcssa211.ph = phi <4 x i32> [ undef, %vector.body.preheader ], [ %119, %middle.block.unr-lcssa.loopexit ]
  %index.unr = phi i64 [ 0, %vector.body.preheader ], [ %index.next.1, %middle.block.unr-lcssa.loopexit ]
  %vec.phi.unr = phi <4 x i32> [ zeroinitializer, %vector.body.preheader ], [ %116, %middle.block.unr-lcssa.loopexit ]
  %vec.phi97.unr = phi <4 x i32> [ zeroinitializer, %vector.body.preheader ], [ %117, %middle.block.unr-lcssa.loopexit ]
  %vec.phi98.unr = phi <4 x i32> [ zeroinitializer, %vector.body.preheader ], [ %118, %middle.block.unr-lcssa.loopexit ]
  %vec.phi99.unr = phi <4 x i32> [ zeroinitializer, %vector.body.preheader ], [ %119, %middle.block.unr-lcssa.loopexit ]
  br i1 %lcmp.mod217, label %middle.block, label %vector.body.epil.preheader

vector.body.epil.preheader:                       ; preds = %middle.block.unr-lcssa
  br label %vector.body.epil

vector.body.epil:                                 ; preds = %vector.body.epil.preheader
  %next.gep.epil = getelementptr i32, i32* %18, i64 %index.unr
  br label %middle.block.epilog-lcssa

middle.block.epilog-lcssa:                        ; preds = %vector.body.epil
  %120 = getelementptr i32, i32* %next.gep.epil, i64 12
  %121 = bitcast i32* %120 to <4 x i32>*
  %wide.load132.epil = load <4 x i32>, <4 x i32>* %121, align 4
  %122 = add <4 x i32> %wide.load132.epil, %vec.phi99.unr
  %123 = getelementptr i32, i32* %next.gep.epil, i64 8
  %124 = bitcast i32* %123 to <4 x i32>*
  %wide.load131.epil = load <4 x i32>, <4 x i32>* %124, align 4
  %125 = add <4 x i32> %wide.load131.epil, %vec.phi98.unr
  %126 = getelementptr i32, i32* %next.gep.epil, i64 4
  %127 = bitcast i32* %126 to <4 x i32>*
  %wide.load130.epil = load <4 x i32>, <4 x i32>* %127, align 4
  %128 = add <4 x i32> %wide.load130.epil, %vec.phi97.unr
  %129 = bitcast i32* %next.gep.epil to <4 x i32>*
  %wide.load.epil = load <4 x i32>, <4 x i32>* %129, align 4
  %130 = add <4 x i32> %wide.load.epil, %vec.phi.unr
  br label %middle.block

middle.block:                                     ; preds = %middle.block.unr-lcssa, %middle.block.epilog-lcssa
  %.lcssa214 = phi <4 x i32> [ %.lcssa214.ph, %middle.block.unr-lcssa ], [ %130, %middle.block.epilog-lcssa ]
  %.lcssa213 = phi <4 x i32> [ %.lcssa213.ph, %middle.block.unr-lcssa ], [ %128, %middle.block.epilog-lcssa ]
  %.lcssa212 = phi <4 x i32> [ %.lcssa212.ph, %middle.block.unr-lcssa ], [ %125, %middle.block.epilog-lcssa ]
  %.lcssa211 = phi <4 x i32> [ %.lcssa211.ph, %middle.block.unr-lcssa ], [ %122, %middle.block.epilog-lcssa ]
  %bin.rdx = add <4 x i32> %.lcssa213, %.lcssa214
  %bin.rdx133 = add <4 x i32> %.lcssa212, %bin.rdx
  %bin.rdx134 = add <4 x i32> %.lcssa211, %bin.rdx133
  %rdx.shuf = shufflevector <4 x i32> %bin.rdx134, <4 x i32> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
  %bin.rdx135 = add <4 x i32> %bin.rdx134, %rdx.shuf
  %rdx.shuf136 = shufflevector <4 x i32> %bin.rdx135, <4 x i32> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
  %bin.rdx137 = add <4 x i32> %bin.rdx135, %rdx.shuf136
  %131 = extractelement <4 x i32> %bin.rdx137, i32 0
  br i1 %cmp.n, label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i", label %bb6.i.i21.preheader204

bb3.us.preheader.i.i:                             ; preds = %"_ZN7ndarray12impl_methods54_$LT$impl$u20$ndarray..ArrayBase$LT$S$C$$u20$D$GT$$GT$18is_standard_layout17he690d19e37be350fE.exit.thread.i.i.i.i.i", %"_ZN85_$LT$ndarray..ArrayBase$LT$ndarray..ViewRepr$LT$$RF$$u27$a$u20$A$GT$$C$$u20$D$GT$$GT$10into_slice17hbd80c1dc13b299edE.exit.thread.i.i.i.i"
  call void @llvm.lifetime.end(i64 40, i8* nonnull %11), !noalias !22
  br i1 %or.cond.i.i.i.i.i.i.i.i, label %bb4.i.i.i.i.us.i.i.preheader, label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i"

bb4.i.i.i.i.us.i.i.preheader:                     ; preds = %bb3.us.preheader.i.i
  br label %bb4.i.i.i.i.us.i.i

bb4.i.i.i.i.us.i.i:                               ; preds = %bb4.i.i.i.i.us.i.i.preheader, %bb6.us.i.i
  %sum.0.us.i33.i = phi i32 [ %141, %bb6.us.i.i ], [ 0, %bb4.i.i.i.i.us.i.i.preheader ]
  %iter.sroa.14.0.us.i32.i = phi i64 [ %_13.sroa.5.0.i.i.i.i.us.i.i, %bb6.us.i.i ], [ 0, %bb4.i.i.i.i.us.i.i.preheader ]
  %iter.sroa.16.0.us.i31.i = phi i64 [ %_13.sroa.6.0.i.i.i.i.us.i.i, %bb6.us.i.i ], [ 0, %bb4.i.i.i.i.us.i.i.preheader ]
  %132 = mul i64 %iter.sroa.14.0.us.i32.i, %a1.sroa.10.0.copyload
  %133 = mul i64 %iter.sroa.16.0.us.i31.i, %a1.sroa.11.0.copyload
  %134 = add i64 %133, %132
  %135 = add i64 %iter.sroa.16.0.us.i31.i, 1
  %136 = icmp eq i64 %135, %a1.sroa.9.0.copyload
  br i1 %136, label %bb1.i.i.i.i.i.us.i.i, label %bb6.i.i.i.i.i.us.i.i

bb1.i.i.i.i.i.us.i.i:                             ; preds = %bb4.i.i.i.i.us.i.i
  %137 = add i64 %iter.sroa.14.0.us.i32.i, 1
  %138 = icmp eq i64 %137, %a1.sroa.8.0.copyload
  br i1 %138, label %"_ZN76_$LT$$LP$usize$C$$u20$usize$RP$$u20$as$u20$ndarray..dimension..Dimension$GT$8next_for17h04b0fe300a278d57E.exit.i.i.i.i.us.i.i", label %bb6.i.i.i.i.i.us.i.i

bb6.i.i.i.i.i.us.i.i:                             ; preds = %bb1.i.i.i.i.i.us.i.i, %bb4.i.i.i.i.us.i.i
  %j.0.i.i.i.i.i.us.i.i = phi i64 [ 0, %bb1.i.i.i.i.i.us.i.i ], [ %135, %bb4.i.i.i.i.us.i.i ]
  %i.0.i.i.i.i.i.us.i.i = phi i64 [ %137, %bb1.i.i.i.i.i.us.i.i ], [ %iter.sroa.14.0.us.i32.i, %bb4.i.i.i.i.us.i.i ]
  br label %"_ZN76_$LT$$LP$usize$C$$u20$usize$RP$$u20$as$u20$ndarray..dimension..Dimension$GT$8next_for17h04b0fe300a278d57E.exit.i.i.i.i.us.i.i"

"_ZN76_$LT$$LP$usize$C$$u20$usize$RP$$u20$as$u20$ndarray..dimension..Dimension$GT$8next_for17h04b0fe300a278d57E.exit.i.i.i.i.us.i.i": ; preds = %bb6.i.i.i.i.i.us.i.i, %bb1.i.i.i.i.i.us.i.i
  %_13.sroa.0.0.i.i.i.i.us.i.i = phi i64 [ 1, %bb6.i.i.i.i.i.us.i.i ], [ 0, %bb1.i.i.i.i.i.us.i.i ]
  %_13.sroa.5.0.i.i.i.i.us.i.i = phi i64 [ %i.0.i.i.i.i.i.us.i.i, %bb6.i.i.i.i.i.us.i.i ], [ undef, %bb1.i.i.i.i.i.us.i.i ]
  %_13.sroa.6.0.i.i.i.i.us.i.i = phi i64 [ %j.0.i.i.i.i.i.us.i.i, %bb6.i.i.i.i.i.us.i.i ], [ undef, %bb1.i.i.i.i.i.us.i.i ]
  %139 = getelementptr inbounds i32, i32* %18, i64 %134
  %switchtmp.us.i.i = icmp eq i32* %139, null
  br i1 %switchtmp.us.i.i, label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i.loopexit", label %bb6.us.i.i

bb6.us.i.i:                                       ; preds = %"_ZN76_$LT$$LP$usize$C$$u20$usize$RP$$u20$as$u20$ndarray..dimension..Dimension$GT$8next_for17h04b0fe300a278d57E.exit.i.i.i.i.us.i.i"
  %140 = load i32, i32* %139, align 4
  %141 = add i32 %140, %sum.0.us.i33.i
  %switch.i.i.i.i.us.i.i = icmp eq i64 %_13.sroa.0.0.i.i.i.i.us.i.i, 1
  br i1 %switch.i.i.i.i.us.i.i, label %bb4.i.i.i.i.us.i.i, label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i.loopexit"

bb6.i.i21:                                        ; preds = %bb6.i.i21.preheader204, %bb6.i.i21
  %sum.0.i29.i = phi i32 [ %144, %bb6.i.i21 ], [ %sum.0.i29.i.ph, %bb6.i.i21.preheader204 ]
  %iter.sroa.4.0.i28.i = phi i32* [ %142, %bb6.i.i21 ], [ %iter.sroa.4.0.i28.i.ph, %bb6.i.i21.preheader204 ]
  %142 = getelementptr inbounds i32, i32* %iter.sroa.4.0.i28.i, i64 1
  %143 = load i32, i32* %iter.sroa.4.0.i28.i, align 4
  %144 = add i32 %143, %sum.0.i29.i
  %145 = icmp eq i32* %142, %20
  br i1 %145, label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i.loopexit205", label %bb6.i.i21, !llvm.loop !42

"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i.loopexit": ; preds = %"_ZN76_$LT$$LP$usize$C$$u20$usize$RP$$u20$as$u20$ndarray..dimension..Dimension$GT$8next_for17h04b0fe300a278d57E.exit.i.i.i.i.us.i.i", %bb6.us.i.i
  %sum.0.lcssa.i.i.ph = phi i32 [ %141, %bb6.us.i.i ], [ %sum.0.us.i33.i, %"_ZN76_$LT$$LP$usize$C$$u20$usize$RP$$u20$as$u20$ndarray..dimension..Dimension$GT$8next_for17h04b0fe300a278d57E.exit.i.i.i.i.us.i.i" ]
  br label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i"

"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i.loopexit205": ; preds = %bb6.i.i21
  br label %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i"

"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i": ; preds = %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i.loopexit205", %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i.loopexit", %middle.block, %bb3.us.preheader.i.i, %bb3.preheader.i.i
  %sum.0.lcssa.i.i = phi i32 [ 0, %bb3.us.preheader.i.i ], [ 0, %bb3.preheader.i.i ], [ %131, %middle.block ], [ %sum.0.lcssa.i.i.ph, %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i.loopexit" ], [ %144, %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i.loopexit205" ]
  call void @llvm.lifetime.start(i64 4, i8* %15) #2
  store i32 %sum.0.lcssa.i.i, i32* %dummy.i.i, align 4
  call void asm "", "r,~{dirflag},~{fpsr},~{flags}"(i32* nonnull %dummy.i.i) #2, !srcloc !16
  call void @llvm.lifetime.end(i64 4, i8* %15) #2
  %exitcond.i = icmp eq i64 %87, %9
  br i1 %exitcond.i, label %bb8.i.loopexit206, label %bb7.i

bb8.i.loopexit:                                   ; preds = %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i.us"
  br label %bb8.i

bb8.i.loopexit206:                                ; preds = %"_ZN6bench129iter_sum_2d_transpose_regular28_$u7b$$u7b$closure$u7d$$u7d$17hde09c7434608be03E.exit.i"
  br label %bb8.i

bb8.i:                                            ; preds = %bb8.i.loopexit206, %bb8.i.loopexit, %.noexc22
  invoke void @_ZN3std4time7Instant7elapsed17h5bac07511bd73b7cE(%"1.std::time::Duration"* noalias nocapture nonnull sret dereferenceable(16) %tmp_ret2.i, %"1.std::time::Instant"* noalias nonnull readonly dereferenceable(16) %start1.i)
          to label %bb5 unwind label %bb9

bb5:                                              ; preds = %bb8.i
  %146 = getelementptr inbounds %"1.std::time::Duration", %"1.std::time::Duration"* %tmp_ret2.i, i64 0, i32 0
  %147 = getelementptr inbounds %"1.std::time::Duration", %"1.std::time::Duration"* %tmp_ret2.i, i64 0, i32 1
  %148 = load i64, i64* %146, align 8
  %149 = load i32, i32* %147, align 8
  %150 = getelementptr inbounds %"12.test::Bencher", %"12.test::Bencher"* %0, i64 0, i32 1, i32 0
  store i64 %148, i64* %150, align 8
  %151 = getelementptr inbounds %"12.test::Bencher", %"12.test::Bencher"* %0, i64 0, i32 1, i32 1
  store i32 %149, i32* %151, align 4
  call void @llvm.lifetime.end(i64 16, i8* %7)
  call void @llvm.lifetime.end(i64 16, i8* %6)
  %not..i.i.i.i.i = icmp eq i64 %a1.sroa.5.0.copyload, 0
  br i1 %not..i.i.i.i.i, label %bb6, label %bb6.i.i.i.i.i24

bb6.i.i.i.i.i24:                                  ; preds = %bb5
  %152 = shl i64 %a1.sroa.5.0.copyload, 2
  call void @__rust_deallocate(i8* %a1.sroa.0.0.copyload, i64 %152, i64 4) #2
  br label %bb6

bb6:                                              ; preds = %bb6.i.i.i.i.i24, %bb5
  ret void

bb9:                                              ; preds = %bb4, %bb8.i
  %153 = landingpad { i8*, i32 }
          cleanup
  %.fca.0.extract5 = extractvalue { i8*, i32 } %153, 0
  %.fca.1.extract7 = extractvalue { i8*, i32 } %153, 1
  %not..i.i.i.i.i28 = icmp eq i64 %a1.sroa.5.0.copyload, 0
  br i1 %not..i.i.i.i.i28, label %bb1, label %bb6.i.i.i.i.i29

bb6.i.i.i.i.i29:                                  ; preds = %bb9
  %154 = shl i64 %a1.sroa.5.0.copyload, 2
  call void @__rust_deallocate(i8* %a1.sroa.0.0.copyload, i64 %154, i64 4) #2
  br label %bb1

cleanup:                                          ; preds = %bb10.i.i.i.i.i
  %155 = landingpad { i8*, i32 }
          cleanup
  %.fca.0.extract4 = extractvalue { i8*, i32 } %155, 0
  %.fca.1.extract = extractvalue { i8*, i32 } %155, 1
  br label %bb1
}

; Function Attrs: uwtable
define internal void @_ZN6bench16__test4main17hea89029b865a9639E() unnamed_addr #0 {
entry-block:
  tail call void @_ZN4test16test_main_static17h9d2a63238910810dE(%"12.test::TestDescAndFn"* noalias nonnull readonly bitcast ({ { { { i64, %str_slice, [8 x i8] }, i1, [7 x i8], { i64, [16 x i8] } }, { i64, void (%"12.test::Bencher"*)*, [8 x i8] } } }* @ref10695 to %"12.test::TestDescAndFn"*), i64 1)
  ret void
}

; Function Attrs: argmemonly nounwind
declare void @llvm.lifetime.start(i64, i8* nocapture) #1

; Function Attrs: nounwind
declare i32 @rust_eh_personality(i32, i32, i64, %"8.unwind::libunwind::_Unwind_Exception"*, %"8.unwind::libunwind::_Unwind_Context"*) unnamed_addr #2

; Function Attrs: argmemonly nounwind
declare void @llvm.lifetime.end(i64, i8* nocapture) #1

; Function Attrs: cold noinline noreturn
declare void @_ZN5alloc3oom3oom17h97560b93ca17f851E() unnamed_addr #3

declare void @_ZN3std4time7Instant3now17h5851ec34f9ff4cdfE(%"1.std::time::Instant"* noalias nocapture sret dereferenceable(16)) unnamed_addr

declare void @_ZN3std4time7Instant7elapsed17h5bac07511bd73b7cE(%"1.std::time::Duration"* noalias nocapture sret dereferenceable(16), %"1.std::time::Instant"* noalias readonly dereferenceable(16)) unnamed_addr

; Function Attrs: nounwind
declare void @__rust_deallocate(i8*, i64, i64) unnamed_addr #2

; Function Attrs: nounwind
declare noalias i8* @__rust_allocate(i64, i64) unnamed_addr #2

declare void @_ZN4test16test_main_static17h9d2a63238910810dE(%"12.test::TestDescAndFn"* noalias nonnull readonly, i64) unnamed_addr

define i64 @main(i64, i8**) unnamed_addr {
top:
  %2 = tail call i64 @_ZN3std2rt10lang_start17h14cbded5fe3cd915E(i8* bitcast (void ()* @_ZN6bench16__test4main17hea89029b865a9639E to i8*), i64 %0, i8** %1)
  ret i64 %2
}

declare i64 @_ZN3std2rt10lang_start17h14cbded5fe3cd915E(i8*, i64, i8**) unnamed_addr

; Function Attrs: argmemonly nounwind
declare void @llvm.memset.p0i8.i64(i8* nocapture writeonly, i8, i64, i32, i1) #1

attributes #0 = { uwtable }
attributes #1 = { argmemonly nounwind }
attributes #2 = { nounwind }
attributes #3 = { cold noinline noreturn }

!llvm.module.flags = !{!0}

!0 = !{i32 1, !"PIE Level", i32 2}
!1 = !{!2, !4, !6, !8, !10}
!2 = distinct !{!2, !3, !"_ZN40_$LT$alloc..raw_vec..RawVec$LT$T$GT$$GT$13with_capacity17h26e462af932cd471E: argument 0"}
!3 = distinct !{!3, !"_ZN40_$LT$alloc..raw_vec..RawVec$LT$T$GT$$GT$13with_capacity17h26e462af932cd471E"}
!4 = distinct !{!4, !5, !"_ZN39_$LT$collections..vec..Vec$LT$T$GT$$GT$13with_capacity17hcda81f751e6a416cE: argument 0"}
!5 = distinct !{!5, !"_ZN39_$LT$collections..vec..Vec$LT$T$GT$$GT$13with_capacity17hcda81f751e6a416cE"}
!6 = distinct !{!6, !7, !"_ZN11collections3vec9from_elem17h7a3e257b3f6bd558E: argument 0"}
!7 = distinct !{!7, !"_ZN11collections3vec9from_elem17h7a3e257b3f6bd558E"}
!8 = distinct !{!8, !9, !"_ZN7ndarray17impl_constructors54_$LT$impl$u20$ndarray..ArrayBase$LT$S$C$$u20$D$GT$$GT$9from_elem17h9578b637100c96adE: argument 0"}
!9 = distinct !{!9, !"_ZN7ndarray17impl_constructors54_$LT$impl$u20$ndarray..ArrayBase$LT$S$C$$u20$D$GT$$GT$9from_elem17h9578b637100c96adE"}
!10 = distinct !{!10, !11, !"_ZN7ndarray17impl_constructors54_$LT$impl$u20$ndarray..ArrayBase$LT$S$C$$u20$D$GT$$GT$5zeros17h454d1cf2928f6f14E: argument 0"}
!11 = distinct !{!11, !"_ZN7ndarray17impl_constructors54_$LT$impl$u20$ndarray..ArrayBase$LT$S$C$$u20$D$GT$$GT$5zeros17h454d1cf2928f6f14E"}
!12 = !{!13, !15}
!13 = distinct !{!13, !14, !"_ZN4test9black_box17he41b3a9d6419a6a4E: argument 0"}
!14 = distinct !{!14, !"_ZN4test9black_box17he41b3a9d6419a6a4E"}
!15 = distinct !{!15, !14, !"_ZN4test9black_box17he41b3a9d6419a6a4E: argument 1"}
!16 = !{i32 191}
!17 = distinct !{!17, !18, !19}
!18 = !{!"llvm.loop.vectorize.width", i32 1}
!19 = !{!"llvm.loop.interleave.count", i32 1}
!20 = distinct !{!20, !21, !18, !19}
!21 = !{!"llvm.loop.unroll.runtime.disable"}
!22 = !{!23, !25, !26, !28}
!23 = distinct !{!23, !24, !"_ZN85_$LT$ndarray..ArrayBase$LT$ndarray..ViewRepr$LT$$RF$$u27$a$u20$A$GT$$C$$u20$D$GT$$GT$10into_iter_17h07e424d86abf96f4E: argument 0"}
!24 = distinct !{!24, !"_ZN85_$LT$ndarray..ArrayBase$LT$ndarray..ViewRepr$LT$$RF$$u27$a$u20$A$GT$$C$$u20$D$GT$$GT$10into_iter_17h07e424d86abf96f4E"}
!25 = distinct !{!25, !24, !"_ZN85_$LT$ndarray..ArrayBase$LT$ndarray..ViewRepr$LT$$RF$$u27$a$u20$A$GT$$C$$u20$D$GT$$GT$10into_iter_17h07e424d86abf96f4E: argument 1"}
!26 = distinct !{!26, !27, !"_ZN7ndarray12impl_methods54_$LT$impl$u20$ndarray..ArrayBase$LT$S$C$$u20$D$GT$$GT$4iter17h2e29b5598ac77580E: argument 0"}
!27 = distinct !{!27, !"_ZN7ndarray12impl_methods54_$LT$impl$u20$ndarray..ArrayBase$LT$S$C$$u20$D$GT$$GT$4iter17h2e29b5598ac77580E"}
!28 = distinct !{!28, !27, !"_ZN7ndarray12impl_methods54_$LT$impl$u20$ndarray..ArrayBase$LT$S$C$$u20$D$GT$$GT$4iter17h2e29b5598ac77580E: argument 1"}
!29 = !{!26, !28}
!30 = !{!31, !33, !35, !23, !25, !26, !28}
!31 = distinct !{!31, !32, !"_ZN7ndarray12impl_methods54_$LT$impl$u20$ndarray..ArrayBase$LT$S$C$$u20$D$GT$$GT$18is_standard_layout17he690d19e37be350fE: argument 0"}
!32 = distinct !{!32, !"_ZN7ndarray12impl_methods54_$LT$impl$u20$ndarray..ArrayBase$LT$S$C$$u20$D$GT$$GT$18is_standard_layout17he690d19e37be350fE"}
!33 = distinct !{!33, !34, !"_ZN85_$LT$ndarray..ArrayBase$LT$ndarray..ViewRepr$LT$$RF$$u27$a$u20$A$GT$$C$$u20$D$GT$$GT$10into_slice17hbd80c1dc13b299edE: argument 0"}
!34 = distinct !{!34, !"_ZN85_$LT$ndarray..ArrayBase$LT$ndarray..ViewRepr$LT$$RF$$u27$a$u20$A$GT$$C$$u20$D$GT$$GT$10into_slice17hbd80c1dc13b299edE"}
!35 = distinct !{!35, !34, !"_ZN85_$LT$ndarray..ArrayBase$LT$ndarray..ViewRepr$LT$$RF$$u27$a$u20$A$GT$$C$$u20$D$GT$$GT$10into_slice17hbd80c1dc13b299edE: argument 1"}
!36 = !{!37}
!37 = distinct !{!37, !38, !"_ZN76_$LT$$LP$usize$C$$u20$usize$RP$$u20$as$u20$ndarray..dimension..Dimension$GT$15default_strides17h9594c0827058adc3E: argument 0"}
!38 = distinct !{!38, !"_ZN76_$LT$$LP$usize$C$$u20$usize$RP$$u20$as$u20$ndarray..dimension..Dimension$GT$15default_strides17h9594c0827058adc3E"}
!39 = !{!31, !35}
!40 = !{!33, !23, !25, !26, !28}
!41 = distinct !{!41, !18, !19}
!42 = distinct !{!42, !21, !18, !19}