RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-24 11:42:57 +01:00
Code Issues Projects Releases Wiki Activity

2 more vdup.32 cases

Browse Source 
llvm-svn: 78419
This commit is contained in:
Anton Korobeynikov 2009-08-07 22:36:50 +00:00
parent 897663328b
commit 9b52601704
3 changed files with 34 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
lib/Target/ARM/ARMInstrNEON.td
View File

@ -1769,6 +1769,20 @@ def VDUPLN16q : VDUPLNQ<0b00, 0b10, "vdup.16", v8i16, v4i16>;
def VDUPLN32q : VDUPLNQ<0b01, 0b00, "vdup.32", v4i32, v2i32>;
def VDUPLNfq : VDUPLNQ<0b01, 0b00, "vdup.32", v4f32, v2f32>;
def VDUPfdf : N2V<0b11, 0b11, 0b01, 0b00, 0b11000, 0, 0,
(outs DPR:$dst), (ins SPR:$src),
"vdup.32\t$dst, ${src:lane}", "",
[(set DPR:$dst, (v2f32 (splat_lo
(scalar_to_vector SPR:$src),
undef)))]>;
def VDUPfqf : N2V<0b11, 0b11, 0b01, 0b00, 0b11000, 1, 0,
(outs QPR:$dst), (ins SPR:$src),
"vdup.32\t$dst, ${src:lane}", "",
[(set QPR:$dst, (v4f32 (splat_lo
(scalar_to_vector SPR:$src),
undef)))]>;
// VMOVN : Vector Narrowing Move
defm VMOVN : N2VNInt_HSD<0b11,0b11,0b10,0b00100,0,0, "vmovn.i",
int_arm_neon_vmovn>;

5
lib/Target/ARM/AsmPrinter/ARMAsmPrinter.cpp
View File

@ -345,6 +345,11 @@ void ARMAsmPrinter::printOperand(const MachineInstr *MI, int OpNum,
O << '{'
<< TRI->getAsmName(DRegLo) << ',' << TRI->getAsmName(DRegHi)
<< '}';
} else if (Modifier && strcmp(Modifier, "lane") == 0) {
unsigned RegNum = ARMRegisterInfo::getRegisterNumbering(Reg);
unsigned DReg = TRI->getMatchingSuperReg(Reg, RegNum & 1 ? 0 : 1,
&ARM::DPRRegClass);
O << TRI->getAsmName(DReg) << '[' << (RegNum & 1) << ']';
} else {
O << TRI->getAsmName(Reg);
}

16
test/CodeGen/ARM/vdup.ll
View File

@ -1,7 +1,7 @@
; RUN: llvm-as < %s | llc -march=arm -mattr=+neon > %t
; RUN: grep vdup.8 %t | count 4
; RUN: grep vdup.16 %t | count 4
; RUN: grep vdup.32 %t | count 8
; RUN: grep vdup.32 %t | count 10
define <8 x i8> @v_dup8(i8 %A) nounwind {
%tmp1 = insertelement <8 x i8> zeroinitializer, i8 %A, i32 0
@ -132,3 +132,17 @@ define <4 x float> @v_shuffledupQfloat(float %A) nounwind {
%tmp2 = shufflevector <4 x float> %tmp1, <4 x float> undef, <4 x i32> zeroinitializer
ret <4 x float> %tmp2
}
define <2 x float> @v_shuffledupfloat2(float* %A) nounwind {
%tmp0 = load float* %A
%tmp1 = insertelement <2 x float> undef, float %tmp0, i32 0
%tmp2 = shufflevector <2 x float> %tmp1, <2 x float> undef, <2 x i32> zeroinitializer
ret <2 x float> %tmp2
}
define <4 x float> @v_shuffledupQfloat2(float* %A) nounwind {
%tmp0 = load float* %A
%tmp1 = insertelement <4 x float> undef, float %tmp0, i32 0
%tmp2 = shufflevector <4 x float> %tmp1, <4 x float> undef, <4 x i32> zeroinitializer
ret <4 x float> %tmp2
}