1 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
2 ; RUN: llc -mtriple=thumbv8.1m.main -mattr=+mve.fp -o - %s | FileCheck %s
3 ; RUN: llc -mtriple=thumbv8.1m.main -mattr=+mve -o - %s | FileCheck %s
5 define arm_aapcs_vfpcc <4 x i32> @vector_add_by_value(<4 x i32> %lhs, <4 x i32>%rhs) {
6 ; CHECK-LABEL: vector_add_by_value:
9 ; CHECK-NEXT: vadd.i32 q0, q0, q1
12 %result = tail call <4 x i32> asm "vadd.i32 $0,$1,$2", "=t,t,t"(<4 x i32> %lhs, <4 x i32> %rhs)
16 define void @vector_add_by_reference(<4 x i32>* %resultp, <4 x i32>* %lhsp, <4 x i32>* %rhsp) {
17 ; CHECK-LABEL: vector_add_by_reference:
19 ; CHECK-NEXT: vldrw.u32 q0, [r1]
20 ; CHECK-NEXT: vldrw.u32 q1, [r2]
22 ; CHECK-NEXT: vadd.i32 q0, q0, q1
24 ; CHECK-NEXT: vstrw.32 q0, [r0]
26 %lhs = load <4 x i32>, <4 x i32>* %lhsp, align 16
27 %rhs = load <4 x i32>, <4 x i32>* %rhsp, align 16
28 %result = tail call <4 x i32> asm "vadd.i32 $0,$1,$2", "=t,t,t"(<4 x i32> %lhs, <4 x i32> %rhs)
29 store <4 x i32> %result, <4 x i32>* %resultp, align 16
33 define void @vector_f64_copy(<2 x double>* %from, <2 x double>* %to) {
34 ; CHECK-LABEL: vector_f64_copy:
36 ; CHECK-NEXT: vldrw.u32 q0, [r0]
37 ; CHECK-NEXT: vstrw.32 q0, [r1]
39 %v = load <2 x double>, <2 x double>* %from, align 16
40 store <2 x double> %v, <2 x double>* %to, align 16
44 define arm_aapcs_vfpcc <16 x i8> @stack_slot_handling(<16 x i8> %a) #0 {
45 ; CHECK-LABEL: stack_slot_handling:
46 ; CHECK: @ %bb.0: @ %entry
47 ; CHECK-NEXT: push {r4, r6, r7, lr}
48 ; CHECK-NEXT: add r7, sp, #8
49 ; CHECK-NEXT: sub sp, #16
50 ; CHECK-NEXT: mov r4, sp
51 ; CHECK-NEXT: bfc r4, #0, #4
52 ; CHECK-NEXT: mov sp, r4
53 ; CHECK-NEXT: mov r0, sp
54 ; CHECK-NEXT: vstrw.32 q0, [r0]
55 ; CHECK-NEXT: vldrw.u32 q0, [r0]
56 ; CHECK-NEXT: sub.w r4, r7, #8
57 ; CHECK-NEXT: mov sp, r4
58 ; CHECK-NEXT: pop {r4, r6, r7, pc}
60 %a.addr = alloca <16 x i8>, align 8
61 store <16 x i8> %a, <16 x i8>* %a.addr, align 8
62 %0 = load <16 x i8>, <16 x i8>* %a.addr, align 8
66 attributes #0 = { noinline optnone }