mirror of
https://github.com/RPCS3/llvm.git
synced 2024-12-26 14:15:53 +00:00
95eb45c5d9
R600 was using a clamped version of rsq, but SI was not. Add a new rsq_clamped intrinsic and use them consistently. It's unclear to me from the documentation what behavior the R600 instructions have, so I assume they have the legacy behavior described by the SI documents. For R600, use RECIPSQRT_IEEE for both llvm.AMDGPU.rsq.legacy and llvm.AMDGPU.rsq. R600 also has RECIPSQRT_FF, which I'm not sure how it fits in here. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211637 91177308-0d34-0410-b5e6-96231b3b80d8
15 lines
617 B
LLVM
15 lines
617 B
LLVM
; RUN: llc -march=r600 -mcpu=SI -verify-machineinstrs < %s | FileCheck -check-prefix=SI -check-prefix=FUNC %s
|
|
; RUN: llc -march=r600 -mcpu=cypress -verify-machineinstrs < %s | FileCheck -check-prefix=EG -check-prefix=FUNC %s
|
|
|
|
|
|
declare float @llvm.AMDGPU.rsq.clamped.f32(float) nounwind readnone
|
|
|
|
; FUNC-LABEL: @rsq_clamped_f32
|
|
; SI: V_RSQ_CLAMP_F32_e32
|
|
; EG: RECIPSQRT_CLAMPED
|
|
define void @rsq_clamped_f32(float addrspace(1)* %out, float %src) nounwind {
|
|
%rsq_clamped = call float @llvm.AMDGPU.rsq.clamped.f32(float %src) nounwind readnone
|
|
store float %rsq_clamped, float addrspace(1)* %out, align 4
|
|
ret void
|
|
}
|