maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/llvm/test/CodeGen/AMDGPU/llvm.r600.recipsqrt.ieee.ll

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

29 lines
1.2 KiB
LLVM
Raw Normal View History

AMDGPU: Remove legacy rsq.clamped intrinsic Mesa still has a use of llvm.AMDGPU.rsq.f64 remaining. Also fix mismatch with non-IEEE rsq selecting to IEEE rsq. llvm-svn: 275617
2016-07-16 05:26:52 +08:00
; RUN: llc -march=r600 -mcpu=cypress -verify-machineinstrs < %s | FileCheck -check-prefix=EG %s
declare float @llvm.r600.recipsqrt.ieee.f32(float) nounwind readnone
; EG-LABEL: {{^}}recipsqrt.ieee_f32:
; EG: RECIPSQRT_IEEE
AMDGPU: Mark all unspecified CC functions in tests as amdgpu_kernel 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
2017-03-22 05:39:51 +08:00
define amdgpu_kernel void @recipsqrt.ieee_f32(float addrspace(1)* %out, float %src) nounwind {
AMDGPU: Remove legacy rsq.clamped intrinsic Mesa still has a use of llvm.AMDGPU.rsq.f64 remaining. Also fix mismatch with non-IEEE rsq selecting to IEEE rsq. llvm-svn: 275617
2016-07-16 05:26:52 +08:00
%recipsqrt.ieee = call float @llvm.r600.recipsqrt.ieee.f32(float %src) nounwind readnone
store float %recipsqrt.ieee, float addrspace(1)* %out, align 4
ret void
}
; TODO: Really these should be constant folded
; EG-LABEL: {{^}}recipsqrt.ieee_f32_constant_4.0
; EG: RECIPSQRT_IEEE
AMDGPU: Mark all unspecified CC functions in tests as amdgpu_kernel 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
2017-03-22 05:39:51 +08:00
define amdgpu_kernel void @recipsqrt.ieee_f32_constant_4.0(float addrspace(1)* %out) nounwind {
AMDGPU: Remove legacy rsq.clamped intrinsic Mesa still has a use of llvm.AMDGPU.rsq.f64 remaining. Also fix mismatch with non-IEEE rsq selecting to IEEE rsq. llvm-svn: 275617
2016-07-16 05:26:52 +08:00
%recipsqrt.ieee = call float @llvm.r600.recipsqrt.ieee.f32(float 4.0) nounwind readnone
store float %recipsqrt.ieee, float addrspace(1)* %out, align 4
ret void
}
; EG-LABEL: {{^}}recipsqrt.ieee_f32_constant_100.0
; EG: RECIPSQRT_IEEE
AMDGPU: Mark all unspecified CC functions in tests as amdgpu_kernel 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
2017-03-22 05:39:51 +08:00
define amdgpu_kernel void @recipsqrt.ieee_f32_constant_100.0(float addrspace(1)* %out) nounwind {
AMDGPU: Remove legacy rsq.clamped intrinsic Mesa still has a use of llvm.AMDGPU.rsq.f64 remaining. Also fix mismatch with non-IEEE rsq selecting to IEEE rsq. llvm-svn: 275617
2016-07-16 05:26:52 +08:00
%recipsqrt.ieee = call float @llvm.r600.recipsqrt.ieee.f32(float 100.0) nounwind readnone
store float %recipsqrt.ieee, float addrspace(1)* %out, align 4
ret void
}