xref: /llvm-project/llvm/test/CodeGen/PowerPC/sms-simple.ll (revision 5403c59c608c08c8ecd4303763f08eb046eb5e4d)
1; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
2; RUN: llc < %s -ppc-vsr-nums-as-vr -mtriple=powerpc64-unknown-linux-gnu \
3; RUN:       -verify-machineinstrs -ppc-asm-full-reg-names -mcpu=pwr9 --ppc-enable-pipeliner \
4; RUN:       | FileCheck %s
5
6@x = dso_local local_unnamed_addr global <{ i32, i32, i32, i32, [1020 x i32] }> <{ i32 1, i32 2, i32 3, i32 4, [1020 x i32] zeroinitializer }>, align 4
7@y = dso_local global [1024 x i32] zeroinitializer, align 4
8
9define dso_local ptr @foo() local_unnamed_addr {
10; CHECK-LABEL: foo:
11; CHECK:       # %bb.0: # %entry
12; CHECK-NEXT:    addis r5, r2, y@toc@ha
13; CHECK-NEXT:    li r7, 340
14; CHECK-NEXT:    addi r3, r5, y@toc@l
15; CHECK-NEXT:    lwz r6, y@toc@l(r5)
16; CHECK-NEXT:    addis r5, r2, x@toc@ha
17; CHECK-NEXT:    mtctr r7
18; CHECK-NEXT:    addi r5, r5, x@toc@l
19; CHECK-NEXT:    addi r4, r3, -8
20; CHECK-NEXT:    addi r5, r5, -8
21; CHECK-NEXT:    lwzu r7, 12(r5)
22; CHECK-NEXT:    maddld r6, r7, r7, r6
23; CHECK-NEXT:    lwz r7, 4(r5)
24; CHECK-NEXT:    stwu r6, 12(r4)
25; CHECK-NEXT:    maddld r6, r7, r7, r6
26; CHECK-NEXT:    lwz r7, 8(r5)
27; CHECK-NEXT:    .p2align 4
28; CHECK-NEXT:  .LBB0_1: # %for.body
29; CHECK-NEXT:    #
30; CHECK-NEXT:    maddld r7, r7, r7, r6
31; CHECK-NEXT:    lwzu r8, 12(r5)
32; CHECK-NEXT:    stw r6, 4(r4)
33; CHECK-NEXT:    lwz r6, 4(r5)
34; CHECK-NEXT:    maddld r8, r8, r8, r7
35; CHECK-NEXT:    stw r7, 8(r4)
36; CHECK-NEXT:    lwz r7, 8(r5)
37; CHECK-NEXT:    maddld r6, r6, r6, r8
38; CHECK-NEXT:    stwu r8, 12(r4)
39; CHECK-NEXT:    bdnz .LBB0_1
40; CHECK-NEXT:  # %bb.2:
41; CHECK-NEXT:    maddld r5, r7, r7, r6
42; CHECK-NEXT:    stw r6, 4(r4)
43; CHECK-NEXT:    stw r5, 8(r4)
44; CHECK-NEXT:    blr
45entry:
46  %.pre = load i32, ptr @y, align 4
47  br label %for.body
48
49for.cond.cleanup:                                 ; preds = %for.body
50  ret ptr @y
51
52for.body:                                         ; preds = %for.body, %entry
53  %0 = phi i32 [ %.pre, %entry ], [ %add.2, %for.body ]
54  %indvars.iv = phi i64 [ 1, %entry ], [ %indvars.iv.next.2, %for.body ]
55  %arrayidx2 = getelementptr inbounds [1024 x i32], ptr @x, i64 0, i64 %indvars.iv
56  %1 = load i32, ptr %arrayidx2, align 4
57  %mul = mul nsw i32 %1, %1
58  %add = add nsw i32 %mul, %0
59  %arrayidx6 = getelementptr inbounds [1024 x i32], ptr @y, i64 0, i64 %indvars.iv
60  store i32 %add, ptr %arrayidx6, align 4
61  %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
62  %arrayidx2.1 = getelementptr inbounds [1024 x i32], ptr @x, i64 0, i64 %indvars.iv.next
63  %2 = load i32, ptr %arrayidx2.1, align 4
64  %mul.1 = mul nsw i32 %2, %2
65  %add.1 = add nsw i32 %mul.1, %add
66  %arrayidx6.1 = getelementptr inbounds [1024 x i32], ptr @y, i64 0, i64 %indvars.iv.next
67  store i32 %add.1, ptr %arrayidx6.1, align 4
68  %indvars.iv.next.1 = add nuw nsw i64 %indvars.iv, 2
69  %arrayidx2.2 = getelementptr inbounds [1024 x i32], ptr @x, i64 0, i64 %indvars.iv.next.1
70  %3 = load i32, ptr %arrayidx2.2, align 4
71  %mul.2 = mul nsw i32 %3, %3
72  %add.2 = add nsw i32 %mul.2, %add.1
73  %arrayidx6.2 = getelementptr inbounds [1024 x i32], ptr @y, i64 0, i64 %indvars.iv.next.1
74  store i32 %add.2, ptr %arrayidx6.2, align 4
75  %indvars.iv.next.2 = add nuw nsw i64 %indvars.iv, 3
76  %exitcond.2 = icmp eq i64 %indvars.iv.next.2, 1024
77  br i1 %exitcond.2, label %for.cond.cleanup, label %for.body
78}
79