1 ; RUN: llc -march=amdgcn -mcpu=SI -enable-misched < %s | FileCheck -check-prefix=SI -check-prefix=FUNC %s
3 ; FUNC-LABEL: {{^}}frem_f32:
4 ; SI-DAG: buffer_load_dword [[X:v[0-9]+]], {{.*$}}
5 ; SI-DAG: buffer_load_dword [[Y:v[0-9]+]], {{.*}} offset:16
14 define void @frem_f32(float addrspace(1)* %out, float addrspace(1)* %in1,
15 float addrspace(1)* %in2) #0 {
16 %gep2 = getelementptr float addrspace(1)* %in2, i32 4
17 %r0 = load float addrspace(1)* %in1, align 4
18 %r1 = load float addrspace(1)* %gep2, align 4
19 %r2 = frem float %r0, %r1
20 store float %r2, float addrspace(1)* %out, align 4
24 ; FUNC-LABEL: {{^}}unsafe_frem_f32:
25 ; SI: buffer_load_dword [[Y:v[0-9]+]], {{.*}} offset:16
26 ; SI: buffer_load_dword [[X:v[0-9]+]], {{.*}}
27 ; SI: v_rcp_f32_e32 [[INVY:v[0-9]+]], [[Y]]
28 ; SI: v_mul_f32_e32 [[DIV:v[0-9]+]], [[INVY]], [[X]]
29 ; SI: v_trunc_f32_e32 [[TRUNC:v[0-9]+]], [[DIV]]
30 ; SI: v_mad_f32 [[RESULT:v[0-9]+]], -[[TRUNC]], [[Y]], [[X]]
31 ; SI: buffer_store_dword [[RESULT]]
33 define void @unsafe_frem_f32(float addrspace(1)* %out, float addrspace(1)* %in1,
34 float addrspace(1)* %in2) #1 {
35 %gep2 = getelementptr float addrspace(1)* %in2, i32 4
36 %r0 = load float addrspace(1)* %in1, align 4
37 %r1 = load float addrspace(1)* %gep2, align 4
38 %r2 = frem float %r0, %r1
39 store float %r2, float addrspace(1)* %out, align 4
43 ; TODO: This should check something when f64 fdiv is implemented
46 ; FUNC-LABEL: {{^}}frem_f64:
48 define void @frem_f64(double addrspace(1)* %out, double addrspace(1)* %in1,
49 double addrspace(1)* %in2) #0 {
50 %r0 = load double addrspace(1)* %in1, align 8
51 %r1 = load double addrspace(1)* %in2, align 8
52 %r2 = frem double %r0, %r1
53 store double %r2, double addrspace(1)* %out, align 8
57 ; FUNC-LABEL: {{^}}unsafe_frem_f64:
63 define void @unsafe_frem_f64(double addrspace(1)* %out, double addrspace(1)* %in1,
64 double addrspace(1)* %in2) #1 {
65 %r0 = load double addrspace(1)* %in1, align 8
66 %r1 = load double addrspace(1)* %in2, align 8
67 %r2 = frem double %r0, %r1
68 store double %r2, double addrspace(1)* %out, align 8
72 define void @frem_v2f32(<2 x float> addrspace(1)* %out, <2 x float> addrspace(1)* %in1,
73 <2 x float> addrspace(1)* %in2) #0 {
74 %gep2 = getelementptr <2 x float> addrspace(1)* %in2, i32 4
75 %r0 = load <2 x float> addrspace(1)* %in1, align 8
76 %r1 = load <2 x float> addrspace(1)* %gep2, align 8
77 %r2 = frem <2 x float> %r0, %r1
78 store <2 x float> %r2, <2 x float> addrspace(1)* %out, align 8
82 define void @frem_v4f32(<4 x float> addrspace(1)* %out, <4 x float> addrspace(1)* %in1,
83 <4 x float> addrspace(1)* %in2) #0 {
84 %gep2 = getelementptr <4 x float> addrspace(1)* %in2, i32 4
85 %r0 = load <4 x float> addrspace(1)* %in1, align 16
86 %r1 = load <4 x float> addrspace(1)* %gep2, align 16
87 %r2 = frem <4 x float> %r0, %r1
88 store <4 x float> %r2, <4 x float> addrspace(1)* %out, align 16
92 define void @frem_v2f64(<2 x double> addrspace(1)* %out, <2 x double> addrspace(1)* %in1,
93 <2 x double> addrspace(1)* %in2) #0 {
94 %gep2 = getelementptr <2 x double> addrspace(1)* %in2, i32 4
95 %r0 = load <2 x double> addrspace(1)* %in1, align 16
96 %r1 = load <2 x double> addrspace(1)* %gep2, align 16
97 %r2 = frem <2 x double> %r0, %r1
98 store <2 x double> %r2, <2 x double> addrspace(1)* %out, align 16
102 attributes #0 = { nounwind "unsafe-fp-math"="false" }
103 attributes #1 = { nounwind "unsafe-fp-math"="true" }