mirror of
https://github.com/RPCS3/llvm.git
synced 2026-07-19 15:13:49 -04:00
[AMDGPU] Improve fma.f64 test. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@372908 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
Stanislav Mekhanoshin
@@ -4,7 +4,7 @@
|
||||
declare double @llvm.fma.f64(double, double, double) nounwind readnone
|
||||
declare <2 x double> @llvm.fma.v2f64(<2 x double>, <2 x double>, <2 x double>) nounwind readnone
|
||||
declare <4 x double> @llvm.fma.v4f64(<4 x double>, <4 x double>, <4 x double>) nounwind readnone
|
||||
|
||||
declare double @llvm.fabs.f64(double) nounwind readnone
|
||||
|
||||
; FUNC-LABEL: {{^}}fma_f64:
|
||||
; SI: v_fma_f64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\]}}
|
||||
@@ -45,3 +45,156 @@ define amdgpu_kernel void @fma_v4f64(<4 x double> addrspace(1)* %out, <4 x doubl
|
||||
store <4 x double> %r3, <4 x double> addrspace(1)* %out
|
||||
ret void
|
||||
}
|
||||
|
||||
; FUNC-LABEL: {{^}}fma_f64_abs_src0:
|
||||
; SI: v_fma_f64 {{v\[[0-9]+:[0-9]+\], |v\[[0-9]+:[0-9]+\]|, v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\]}}
|
||||
define amdgpu_kernel void @fma_f64_abs_src0(double addrspace(1)* %out, double addrspace(1)* %in1,
|
||||
double addrspace(1)* %in2, double addrspace(1)* %in3) {
|
||||
%r0 = load double, double addrspace(1)* %in1
|
||||
%r1 = load double, double addrspace(1)* %in2
|
||||
%r2 = load double, double addrspace(1)* %in3
|
||||
%fabs = call double @llvm.fabs.f64(double %r0)
|
||||
%r3 = tail call double @llvm.fma.f64(double %fabs, double %r1, double %r2)
|
||||
store double %r3, double addrspace(1)* %out
|
||||
ret void
|
||||
}
|
||||
|
||||
; FUNC-LABEL: {{^}}fma_f64_abs_src1:
|
||||
; SI: v_fma_f64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], \|v\[[0-9]+:[0-9]+\]\|, v\[[0-9]+:[0-9]+\]}}
|
||||
define amdgpu_kernel void @fma_f64_abs_src1(double addrspace(1)* %out, double addrspace(1)* %in1,
|
||||
double addrspace(1)* %in2, double addrspace(1)* %in3) {
|
||||
%r0 = load double, double addrspace(1)* %in1
|
||||
%r1 = load double, double addrspace(1)* %in2
|
||||
%r2 = load double, double addrspace(1)* %in3
|
||||
%fabs = call double @llvm.fabs.f64(double %r1)
|
||||
%r3 = tail call double @llvm.fma.f64(double %r0, double %fabs, double %r2)
|
||||
store double %r3, double addrspace(1)* %out
|
||||
ret void
|
||||
}
|
||||
|
||||
; FUNC-LABEL: {{^}}fma_f64_abs_src2:
|
||||
; SI: v_fma_f64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], \|v\[[0-9]+:[0-9]+\]\|}}
|
||||
define amdgpu_kernel void @fma_f64_abs_src2(double addrspace(1)* %out, double addrspace(1)* %in1,
|
||||
double addrspace(1)* %in2, double addrspace(1)* %in3) {
|
||||
%r0 = load double, double addrspace(1)* %in1
|
||||
%r1 = load double, double addrspace(1)* %in2
|
||||
%r2 = load double, double addrspace(1)* %in3
|
||||
%fabs = call double @llvm.fabs.f64(double %r2)
|
||||
%r3 = tail call double @llvm.fma.f64(double %r0, double %r1, double %fabs)
|
||||
store double %r3, double addrspace(1)* %out
|
||||
ret void
|
||||
}
|
||||
|
||||
; FUNC-LABEL: {{^}}fma_f64_neg_src0:
|
||||
; SI: v_fma_f64 {{v\[[0-9]+:[0-9]+\], -v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\]}}
|
||||
define amdgpu_kernel void @fma_f64_neg_src0(double addrspace(1)* %out, double addrspace(1)* %in1,
|
||||
double addrspace(1)* %in2, double addrspace(1)* %in3) {
|
||||
%r0 = load double, double addrspace(1)* %in1
|
||||
%r1 = load double, double addrspace(1)* %in2
|
||||
%r2 = load double, double addrspace(1)* %in3
|
||||
%fsub = fsub double -0.000000e+00, %r0
|
||||
%r3 = tail call double @llvm.fma.f64(double %fsub, double %r1, double %r2)
|
||||
store double %r3, double addrspace(1)* %out
|
||||
ret void
|
||||
}
|
||||
|
||||
; FUNC-LABEL: {{^}}fma_f64_neg_src1:
|
||||
; SI: v_fma_f64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], -v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\]}}
|
||||
define amdgpu_kernel void @fma_f64_neg_src1(double addrspace(1)* %out, double addrspace(1)* %in1,
|
||||
double addrspace(1)* %in2, double addrspace(1)* %in3) {
|
||||
%r0 = load double, double addrspace(1)* %in1
|
||||
%r1 = load double, double addrspace(1)* %in2
|
||||
%r2 = load double, double addrspace(1)* %in3
|
||||
%fsub = fsub double -0.000000e+00, %r1
|
||||
%r3 = tail call double @llvm.fma.f64(double %r0, double %fsub, double %r2)
|
||||
store double %r3, double addrspace(1)* %out
|
||||
ret void
|
||||
}
|
||||
|
||||
; FUNC-LABEL: {{^}}fma_f64_neg_src2:
|
||||
; SI: v_fma_f64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], -v\[[0-9]+:[0-9]+\]}}
|
||||
define amdgpu_kernel void @fma_f64_neg_src2(double addrspace(1)* %out, double addrspace(1)* %in1,
|
||||
double addrspace(1)* %in2, double addrspace(1)* %in3) {
|
||||
%r0 = load double, double addrspace(1)* %in1
|
||||
%r1 = load double, double addrspace(1)* %in2
|
||||
%r2 = load double, double addrspace(1)* %in3
|
||||
%fsub = fsub double -0.000000e+00, %r2
|
||||
%r3 = tail call double @llvm.fma.f64(double %r0, double %r1, double %fsub)
|
||||
store double %r3, double addrspace(1)* %out
|
||||
ret void
|
||||
}
|
||||
|
||||
; FUNC-LABEL: {{^}}fma_f64_abs_neg_src0:
|
||||
; SI: v_fma_f64 {{v\[[0-9]+:[0-9]+\], -\|v\[[0-9]+:[0-9]+\]\|, v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\]}}
|
||||
define amdgpu_kernel void @fma_f64_abs_neg_src0(double addrspace(1)* %out, double addrspace(1)* %in1,
|
||||
double addrspace(1)* %in2, double addrspace(1)* %in3) {
|
||||
%r0 = load double, double addrspace(1)* %in1
|
||||
%r1 = load double, double addrspace(1)* %in2
|
||||
%r2 = load double, double addrspace(1)* %in3
|
||||
%fabs = call double @llvm.fabs.f64(double %r0)
|
||||
%fsub = fsub double -0.000000e+00, %fabs
|
||||
%r3 = tail call double @llvm.fma.f64(double %fsub, double %r1, double %r2)
|
||||
store double %r3, double addrspace(1)* %out
|
||||
ret void
|
||||
}
|
||||
|
||||
; FUNC-LABEL: {{^}}fma_f64_abs_neg_src1:
|
||||
; SI: v_fma_f64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], -\|v\[[0-9]+:[0-9]+\]\|, v\[[0-9]+:[0-9]+\]}}
|
||||
define amdgpu_kernel void @fma_f64_abs_neg_src1(double addrspace(1)* %out, double addrspace(1)* %in1,
|
||||
double addrspace(1)* %in2, double addrspace(1)* %in3) {
|
||||
%r0 = load double, double addrspace(1)* %in1
|
||||
%r1 = load double, double addrspace(1)* %in2
|
||||
%r2 = load double, double addrspace(1)* %in3
|
||||
%fabs = call double @llvm.fabs.f64(double %r1)
|
||||
%fsub = fsub double -0.000000e+00, %fabs
|
||||
%r3 = tail call double @llvm.fma.f64(double %r0, double %fsub, double %r2)
|
||||
store double %r3, double addrspace(1)* %out
|
||||
ret void
|
||||
}
|
||||
|
||||
; FUNC-LABEL: {{^}}fma_f64_abs_neg_src2:
|
||||
; SI: v_fma_f64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], -\|v\[[0-9]+:[0-9]+\]\|}}
|
||||
define amdgpu_kernel void @fma_f64_abs_neg_src2(double addrspace(1)* %out, double addrspace(1)* %in1,
|
||||
double addrspace(1)* %in2, double addrspace(1)* %in3) {
|
||||
%r0 = load double, double addrspace(1)* %in1
|
||||
%r1 = load double, double addrspace(1)* %in2
|
||||
%r2 = load double, double addrspace(1)* %in3
|
||||
%fabs = call double @llvm.fabs.f64(double %r2)
|
||||
%fsub = fsub double -0.000000e+00, %fabs
|
||||
%r3 = tail call double @llvm.fma.f64(double %r0, double %r1, double %fsub)
|
||||
store double %r3, double addrspace(1)* %out
|
||||
ret void
|
||||
}
|
||||
|
||||
; FUNC-LABEL: {{^}}fma_f64_lit_src0:
|
||||
; SI: v_fma_f64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], 2.0, v\[[0-9]+:[0-9]+\]}}
|
||||
define amdgpu_kernel void @fma_f64_lit_src0(double addrspace(1)* %out,
|
||||
double addrspace(1)* %in2, double addrspace(1)* %in3) {
|
||||
%r1 = load double, double addrspace(1)* %in2
|
||||
%r2 = load double, double addrspace(1)* %in3
|
||||
%r3 = tail call double @llvm.fma.f64(double +2.0, double %r1, double %r2)
|
||||
store double %r3, double addrspace(1)* %out
|
||||
ret void
|
||||
}
|
||||
|
||||
; FUNC-LABEL: {{^}}fma_f64_lit_src1:
|
||||
; SI: v_fma_f64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], 2.0, v\[[0-9]+:[0-9]+\]}}
|
||||
define amdgpu_kernel void @fma_f64_lit_src1(double addrspace(1)* %out, double addrspace(1)* %in1,
|
||||
double addrspace(1)* %in3) {
|
||||
%r0 = load double, double addrspace(1)* %in1
|
||||
%r2 = load double, double addrspace(1)* %in3
|
||||
%r3 = tail call double @llvm.fma.f64(double %r0, double +2.0, double %r2)
|
||||
store double %r3, double addrspace(1)* %out
|
||||
ret void
|
||||
}
|
||||
|
||||
; FUNC-LABEL: {{^}}fma_f64_lit_src2:
|
||||
; SI: v_fma_f64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], 2.0}}
|
||||
define amdgpu_kernel void @fma_f64_lit_src2(double addrspace(1)* %out, double addrspace(1)* %in1,
|
||||
double addrspace(1)* %in2) {
|
||||
%r0 = load double, double addrspace(1)* %in1
|
||||
%r1 = load double, double addrspace(1)* %in2
|
||||
%r3 = tail call double @llvm.fma.f64(double %r0, double %r1, double +2.0)
|
||||
store double %r3, double addrspace(1)* %out
|
||||
ret void
|
||||
}
|
||||
|
||||
Reference in New Issue
Block a user