mirror of
https://github.com/RPCS3/llvm-mirror.git
synced 2024-10-24 05:23:45 +02:00
dd9ab77318
Currently the default C calling convention functions are treated the same as compute kernels. Make this explicit so the default calling convention can be changed to a non-kernel. Converted with perl -pi -e 's/define void/define amdgpu_kernel void/' on the relevant test directories (and undoing in one place that actually wanted a non-kernel). llvm-svn: 298444
75 lines
2.9 KiB
LLVM
75 lines
2.9 KiB
LLVM
; RUN: llc < %s -march=amdgcn -mcpu=tahiti -verify-machineinstrs | FileCheck %s
|
|
; RUN: llc < %s -march=amdgcn -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs | FileCheck %s
|
|
|
|
; CHECK-LABEL: {{^}}flt_f64:
|
|
; CHECK: v_cmp_nge_f64_e32 vcc, {{v[[0-9]+:[0-9]+], v[[0-9]+:[0-9]+]}}
|
|
define amdgpu_kernel void @flt_f64(i32 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
|
|
%r2 = fcmp ult double %r0, %r1
|
|
%r3 = zext i1 %r2 to i32
|
|
store i32 %r3, i32 addrspace(1)* %out
|
|
ret void
|
|
}
|
|
|
|
; CHECK-LABEL: {{^}}fle_f64:
|
|
; CHECK: v_cmp_ngt_f64_e32 vcc, {{v[[0-9]+:[0-9]+], v[[0-9]+:[0-9]+]}}
|
|
define amdgpu_kernel void @fle_f64(i32 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
|
|
%r2 = fcmp ule double %r0, %r1
|
|
%r3 = zext i1 %r2 to i32
|
|
store i32 %r3, i32 addrspace(1)* %out
|
|
ret void
|
|
}
|
|
|
|
; CHECK-LABEL: {{^}}fgt_f64:
|
|
; CHECK: v_cmp_nle_f64_e32 vcc, {{v[[0-9]+:[0-9]+], v[[0-9]+:[0-9]+]}}
|
|
define amdgpu_kernel void @fgt_f64(i32 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
|
|
%r2 = fcmp ugt double %r0, %r1
|
|
%r3 = zext i1 %r2 to i32
|
|
store i32 %r3, i32 addrspace(1)* %out
|
|
ret void
|
|
}
|
|
|
|
; CHECK-LABEL: {{^}}fge_f64:
|
|
; CHECK: v_cmp_nlt_f64_e32 vcc, {{v[[0-9]+:[0-9]+], v[[0-9]+:[0-9]+]}}
|
|
define amdgpu_kernel void @fge_f64(i32 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
|
|
%r2 = fcmp uge double %r0, %r1
|
|
%r3 = zext i1 %r2 to i32
|
|
store i32 %r3, i32 addrspace(1)* %out
|
|
ret void
|
|
}
|
|
|
|
; CHECK-LABEL: {{^}}fne_f64:
|
|
; CHECK: v_cmp_neq_f64_e32 vcc, {{v[[0-9]+:[0-9]+], v[[0-9]+:[0-9]+]}}
|
|
define amdgpu_kernel void @fne_f64(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
|
|
%r2 = fcmp une double %r0, %r1
|
|
%r3 = select i1 %r2, double %r0, double %r1
|
|
store double %r3, double addrspace(1)* %out
|
|
ret void
|
|
}
|
|
|
|
; CHECK-LABEL: {{^}}feq_f64:
|
|
; CHECK: v_cmp_nlg_f64_e32 vcc, {{v[[0-9]+:[0-9]+], v[[0-9]+:[0-9]+]}}
|
|
define amdgpu_kernel void @feq_f64(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
|
|
%r2 = fcmp ueq double %r0, %r1
|
|
%r3 = select i1 %r2, double %r0, double %r1
|
|
store double %r3, double addrspace(1)* %out
|
|
ret void
|
|
}
|