mirror of
https://github.com/RPCSX/llvm.git
synced 2024-11-23 19:59:57 +00:00
d706d030af
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). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@298444 91177308-0d34-0410-b5e6-96231b3b80d8
47 lines
2.0 KiB
LLVM
47 lines
2.0 KiB
LLVM
;RUN: llc < %s -march=r600 -mcpu=cayman -stress-sched -verify-misched -verify-machineinstrs
|
|
;REQUIRES: asserts
|
|
|
|
define amdgpu_kernel void @main() {
|
|
main_body:
|
|
%0 = load <4 x float>, <4 x float> addrspace(8)* getelementptr ([1024 x <4 x float>], [1024 x <4 x float>] addrspace(8)* null, i64 0, i32 1)
|
|
%1 = extractelement <4 x float> %0, i32 0
|
|
%2 = bitcast float %1 to i32
|
|
%3 = icmp eq i32 %2, 0
|
|
%4 = sext i1 %3 to i32
|
|
%5 = bitcast i32 %4 to float
|
|
%6 = bitcast float %5 to i32
|
|
%7 = icmp ne i32 %6, 0
|
|
br i1 %7, label %ENDIF, label %ELSE
|
|
|
|
ELSE: ; preds = %main_body
|
|
%8 = load <4 x float>, <4 x float> addrspace(8)* getelementptr ([1024 x <4 x float>], [1024 x <4 x float>] addrspace(8)* null, i64 0, i32 1)
|
|
%9 = extractelement <4 x float> %8, i32 0
|
|
%10 = bitcast float %9 to i32
|
|
%11 = icmp eq i32 %10, 1
|
|
%12 = sext i1 %11 to i32
|
|
%13 = bitcast i32 %12 to float
|
|
%14 = bitcast float %13 to i32
|
|
%15 = icmp ne i32 %14, 0
|
|
br i1 %15, label %IF13, label %ENDIF
|
|
|
|
ENDIF: ; preds = %IF13, %ELSE, %main_body
|
|
%temp.0 = phi float [ 1.000000e+03, %main_body ], [ 1.000000e+00, %IF13 ], [ 0.000000e+00, %ELSE ]
|
|
%temp1.0 = phi float [ 0.000000e+00, %main_body ], [ %23, %IF13 ], [ 0.000000e+00, %ELSE ]
|
|
%temp3.0 = phi float [ 1.000000e+00, %main_body ], [ 0.000000e+00, %ELSE ], [ 0.000000e+00, %IF13 ]
|
|
%16 = insertelement <4 x float> undef, float %temp.0, i32 0
|
|
%17 = insertelement <4 x float> %16, float %temp1.0, i32 1
|
|
%18 = insertelement <4 x float> %17, float 0.000000e+00, i32 2
|
|
%19 = insertelement <4 x float> %18, float %temp3.0, i32 3
|
|
call void @llvm.r600.store.swizzle(<4 x float> %19, i32 0, i32 0)
|
|
ret void
|
|
|
|
IF13: ; preds = %ELSE
|
|
%20 = load <4 x float>, <4 x float> addrspace(8)* null
|
|
%21 = extractelement <4 x float> %20, i32 0
|
|
%22 = fsub float -0.000000e+00, %21
|
|
%23 = fadd float 1.000000e+03, %22
|
|
br label %ENDIF
|
|
}
|
|
|
|
declare void @llvm.r600.store.swizzle(<4 x float>, i32, i32)
|