R600/SI: Match cttz_zero_undef

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211116 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
Matt Arsenault 2014-06-17 17:36:27 +00:00
parent 62e378b057
commit 3f1f259c22
3 changed files with 63 additions and 1 deletions

View File

@ -679,6 +679,7 @@ unsigned SIInstrInfo::getVALUOp(const MachineInstr &MI) {
case AMDGPU::S_LOAD_DWORDX4_IMM:
case AMDGPU::S_LOAD_DWORDX4_SGPR: return AMDGPU::BUFFER_LOAD_DWORDX4_ADDR64;
case AMDGPU::S_BCNT1_I32_B32: return AMDGPU::V_BCNT_U32_B32_e32;
case AMDGPU::S_FF1_I32_B32: return AMDGPU::V_FFBL_B32_e32;
case AMDGPU::S_FLBIT_I32_B32: return AMDGPU::V_FFBH_U32_e32;
}
}

View File

@ -114,11 +114,15 @@ def S_BCNT1_I32_B64 : SOP1_32_64 <0x00000010, "S_BCNT1_I32_B64", []>;
////def S_FF0_I32_B32 : SOP1_32 <0x00000011, "S_FF0_I32_B32", []>;
////def S_FF0_I32_B64 : SOP1_FF0 <0x00000012, "S_FF0_I32_B64", []>;
////def S_FF1_I32_B32 : SOP1_32 <0x00000013, "S_FF1_I32_B32", []>;
def S_FF1_I32_B32 : SOP1_32 <0x00000013, "S_FF1_I32_B32",
[(set i32:$dst, (cttz_zero_undef i32:$src0))]
>;
////def S_FF1_I32_B64 : SOP1_FF1 <0x00000014, "S_FF1_I32_B64", []>;
def S_FLBIT_I32_B32 : SOP1_32 <0x00000015, "S_FLBIT_I32_B32",
[(set i32:$dst, (ctlz_zero_undef i32:$src0))]
>;
//def S_FLBIT_I32_B64 : SOP1_32 <0x00000016, "S_FLBIT_I32_B64", []>;
def S_FLBIT_I32 : SOP1_32 <0x00000017, "S_FLBIT_I32", []>;
//def S_FLBIT_I32_I64 : SOP1_32 <0x00000018, "S_FLBIT_I32_I64", []>;

View File

@ -0,0 +1,57 @@
; RUN: llc -march=r600 -mcpu=SI -verify-machineinstrs < %s | FileCheck -check-prefix=SI -check-prefix=FUNC %s
declare i32 @llvm.cttz.i32(i32, i1) nounwind readnone
declare <2 x i32> @llvm.cttz.v2i32(<2 x i32>, i1) nounwind readnone
declare <4 x i32> @llvm.cttz.v4i32(<4 x i32>, i1) nounwind readnone
; FUNC-LABEL: @s_cttz_zero_undef_i32:
; SI: S_LOAD_DWORD [[VAL:s[0-9]+]],
; SI: S_FF1_I32_B32 [[SRESULT:s[0-9]+]], [[VAL]]
; SI: V_MOV_B32_e32 [[VRESULT:v[0-9]+]], [[SRESULT]]
; SI: BUFFER_STORE_DWORD [[VRESULT]],
; SI: S_ENDPGM
define void @s_cttz_zero_undef_i32(i32 addrspace(1)* noalias %out, i32 %val) nounwind {
%cttz = call i32 @llvm.cttz.i32(i32 %val, i1 true) nounwind readnone
store i32 %cttz, i32 addrspace(1)* %out, align 4
ret void
}
; FUNC-LABEL: @v_cttz_zero_undef_i32:
; SI: BUFFER_LOAD_DWORD [[VAL:v[0-9]+]],
; SI: V_FFBL_B32_e32 [[RESULT:v[0-9]+]], [[VAL]]
; SI: BUFFER_STORE_DWORD [[RESULT]],
; SI: S_ENDPGM
define void @v_cttz_zero_undef_i32(i32 addrspace(1)* noalias %out, i32 addrspace(1)* noalias %valptr) nounwind {
%val = load i32 addrspace(1)* %valptr, align 4
%cttz = call i32 @llvm.cttz.i32(i32 %val, i1 true) nounwind readnone
store i32 %cttz, i32 addrspace(1)* %out, align 4
ret void
}
; FUNC-LABEL: @v_cttz_zero_undef_v2i32:
; SI: BUFFER_LOAD_DWORDX2
; SI: V_FFBL_B32_e32
; SI: V_FFBL_B32_e32
; SI: BUFFER_STORE_DWORDX2
; SI: S_ENDPGM
define void @v_cttz_zero_undef_v2i32(<2 x i32> addrspace(1)* noalias %out, <2 x i32> addrspace(1)* noalias %valptr) nounwind {
%val = load <2 x i32> addrspace(1)* %valptr, align 8
%cttz = call <2 x i32> @llvm.cttz.v2i32(<2 x i32> %val, i1 true) nounwind readnone
store <2 x i32> %cttz, <2 x i32> addrspace(1)* %out, align 8
ret void
}
; FUNC-LABEL: @v_cttz_zero_undef_v4i32:
; SI: BUFFER_LOAD_DWORDX4
; SI: V_FFBL_B32_e32
; SI: V_FFBL_B32_e32
; SI: V_FFBL_B32_e32
; SI: V_FFBL_B32_e32
; SI: BUFFER_STORE_DWORDX4
; SI: S_ENDPGM
define void @v_cttz_zero_undef_v4i32(<4 x i32> addrspace(1)* noalias %out, <4 x i32> addrspace(1)* noalias %valptr) nounwind {
%val = load <4 x i32> addrspace(1)* %valptr, align 16
%cttz = call <4 x i32> @llvm.cttz.v4i32(<4 x i32> %val, i1 true) nounwind readnone
store <4 x i32> %cttz, <4 x i32> addrspace(1)* %out, align 16
ret void
}