/external/llvm/lib/Target/ARM/ |
ARMTargetTransformInfo.cpp | 189 { ISD::FP_EXTEND, MVT::v4f32, 4 } 235 { ISD::SINT_TO_FP, MVT::v4f32, MVT::v4i32, 1 }, 236 { ISD::UINT_TO_FP, MVT::v4f32, MVT::v4i32, 1 }, 244 { ISD::SINT_TO_FP, MVT::v4f32, MVT::v4i1, 3 }, 245 { ISD::UINT_TO_FP, MVT::v4f32, MVT::v4i1, 3 }, 246 { ISD::SINT_TO_FP, MVT::v4f32, MVT::v4i8, 3 }, 247 { ISD::UINT_TO_FP, MVT::v4f32, MVT::v4i8, 3 }, 248 { ISD::SINT_TO_FP, MVT::v4f32, MVT::v4i16, 2 }, 249 { ISD::UINT_TO_FP, MVT::v4f32, MVT::v4i16, 2 }, 259 { ISD::FP_TO_SINT, MVT::v4i32, MVT::v4f32, 1 } [all...] |
ARMCallingConv.td | 28 CCIfType<[v2i64, v4i32, v8i16, v16i8, v4f32], CCBitConvertToType<v2f64>>, 47 CCIfType<[v2i64, v4i32, v8i16, v16i8, v4f32], CCBitConvertToType<v2f64>>, 61 CCIfType<[v2i64, v4i32, v8i16, v16i8, v4f32], CCBitConvertToType<v2f64>>, 73 CCIfType<[v2i64, v4i32, v8i16, v16i8, v4f32], CCBitConvertToType<v2f64>>, 89 CCIfType<[v2i64, v4i32, v8i16, v16i8, v4f32], CCBitConvertToType<v2f64>>, 139 CCIfType<[v2i64, v4i32, v8i16, v16i8, v4f32], CCBitConvertToType<v2f64>>, 149 CCIfType<[v2i64, v4i32, v8i16, v16i8, v4f32], CCBitConvertToType<v2f64>>, 167 CCIfType<[v2i64, v4i32, v8i16, v16i8, v4f32], CCBitConvertToType<v2f64>>, 179 CCIfType<[v2i64, v4i32, v8i16, v16i8, v4f32], CCBitConvertToType<v2f64>>,
|
ARMInstrNEON.td | [all...] |
/external/llvm/test/CodeGen/R600/ |
store.r600.ll | 15 ; v4f32 store
|
/frameworks/rs/driver/runtime/ |
math.ll | 9 declare <4 x float> @llvm.fabs.v4f32(<4 x float>)
|
/external/llvm/test/CodeGen/ARM/ |
coalesce-subregs.ll | 17 %vld2 = tail call { <4 x float>, <4 x float> } @llvm.arm.neon.vld2.v4f32(i8* %0, i32 4) 21 tail call void @llvm.arm.neon.vst2.v4f32(i8* %1, <4 x float> %vld221, <4 x float> undef, i32 4) 30 %vld2 = tail call { <4 x float>, <4 x float> } @llvm.arm.neon.vld2.v4f32(i8* %0, i32 4) 34 %vld22 = tail call { <4 x float>, <4 x float> } @llvm.arm.neon.vld2.v4f32(i8* %1, i32 4) 36 tail call void @llvm.arm.neon.vst2.v4f32(i8* %1, <4 x float> %vld221, <4 x float> %vld2215, i32 4) 45 %vld2 = tail call { <4 x float>, <4 x float> } @llvm.arm.neon.vld2.v4f32(i8* %0, i32 4) 55 %vld22 = tail call { <4 x float>, <4 x float> } @llvm.arm.neon.vld2.v4f32(i8* %1, i32 4) 58 tail call void @llvm.arm.neon.vst2.v4f32(i8* %1, <4 x float> %qq0.0.1.0, <4 x float> %vld2215, i32 4) 67 declare { <4 x float>, <4 x float> } @llvm.arm.neon.vld2.v4f32(i8*, i32) nounwind readonly 68 declare void @llvm.arm.neon.vst2.v4f32(i8*, <4 x float>, <4 x float>, i32) nounwin [all...] |
vld1.ll | 108 %tmp1 = call <4 x float> @llvm.arm.neon.vld1.v4f32(i8* %tmp0, i32 1) 129 declare <4 x float> @llvm.arm.neon.vld1.v4f32(i8*, i32) nounwind readonly 136 %0 = call <4 x float> @llvm.arm.neon.vld1.v4f32(i8* undef, i32 1)
|
vst2.ll | 109 call void @llvm.arm.neon.vst2.v4f32(i8* %tmp0, <4 x float> %tmp1, <4 x float> %tmp1, i32 1) 126 call void @llvm.arm.neon.vst2.v4f32(i8* %out, <4 x float> %tmp1, <4 x float> %tmp1, i32 4) nounwind 140 declare void @llvm.arm.neon.vst2.v4f32(i8*, <4 x float>, <4 x float>, i32) nounwind
|
/external/llvm/lib/Target/X86/ |
X86InstrAVX512.td | 22 def : Pat<(v2i64 (bitconvert (v4f32 VR128X:$src))), (v2i64 VR128X:$src)>; 27 def : Pat<(v4i32 (bitconvert (v4f32 VR128X:$src))), (v4i32 VR128X:$src)>; 32 def : Pat<(v8i16 (bitconvert (v4f32 VR128X:$src))), (v8i16 VR128X:$src)>; 37 def : Pat<(v16i8 (bitconvert (v4f32 VR128X:$src))), (v16i8 VR128X:$src)>; 38 def : Pat<(v4f32 (bitconvert (v2i64 VR128X:$src))), (v4f32 VR128X:$src)>; 39 def : Pat<(v4f32 (bitconvert (v4i32 VR128X:$src))), (v4f32 VR128X:$src)>; 40 def : Pat<(v4f32 (bitconvert (v8i16 VR128X:$src))), (v4f32 VR128X:$src)> [all...] |
X86InstrFMA.td | 78 memopv8f32, X86Fmadd, v4f32, v8f32>; 80 memopv8f32, X86Fmsub, v4f32, v8f32>; 83 v4f32, v8f32>; 86 v4f32, v8f32>; 105 memopv8f32, X86Fnmadd, v4f32, v8f32>; 107 memopv8f32, X86Fnmsub, v4f32, v8f32>; 340 defm VFMADDPS4 : fma4p<0x68, "vfmaddps", X86Fmadd, v4f32, v8f32, 342 defm VFMSUBPS4 : fma4p<0x6C, "vfmsubps", X86Fmsub, v4f32, v8f32, 344 defm VFNMADDPS4 : fma4p<0x78, "vfnmaddps", X86Fnmadd, v4f32, v8f32, 346 defm VFNMSUBPS4 : fma4p<0x7C, "vfnmsubps", X86Fnmsub, v4f32, v8f32 [all...] |
X86InstrSSE.td | 250 def : Pat<(f32 (vector_extract (v4f32 VR128:$src), (iPTR 0))), 251 (COPY_TO_REGCLASS (v4f32 VR128:$src), FR32)>; 259 def : Pat<(v4f32 (extract_subvector (v8f32 VR256:$src), (iPTR 0))), 260 (v4f32 (EXTRACT_SUBREG (v8f32 VR256:$src), sub_xmm))>; 281 def : Pat<(insert_subvector undef, (v4f32 VR128:$src), (iPTR 0)), 290 def : Pat<(v4f32 (scalar_to_vector FR32:$src)), 307 def : Pat<(v2i64 (bitconvert (v4f32 VR128:$src))), (v2i64 VR128:$src)>; 312 def : Pat<(v4i32 (bitconvert (v4f32 VR128:$src))), (v4i32 VR128:$src)>; 317 def : Pat<(v8i16 (bitconvert (v4f32 VR128:$src))), (v8i16 VR128:$src)>; 322 def : Pat<(v16i8 (bitconvert (v4f32 VR128:$src))), (v16i8 VR128:$src)> [all...] |
X86CallingConv.td | 43 CCIfType<[v16i8, v8i16, v4i32, v2i64, v4f32, v2f64], 100 CCIfType<[f32, f64, v4i32, v2i64, v4f32, v2f64], 216 CCIfType<[f32, f64, v16i8, v8i16, v4i32, v2i64, v4f32, v2f64], 244 CCIfType<[v16i8, v8i16, v4i32, v2i64, v4f32, v2f64], CCAssignToStack<16, 16>>, 267 CCIfType<[v16i8, v8i16, v4i32, v2i64, v4f32, v2f64], CCPassIndirect<i64>>, 293 CCIfType<[f32, f64, v16i8, v8i16, v4i32, v2i64, v4f32, v2f64], 315 CCIfType<[f32, f64, v16i8, v8i16, v4i32, v2i64, v4f32, v2f64], 365 CCIfNotVarArg<CCIfType<[v16i8, v8i16, v4i32, v2i64, v4f32, v2f64], 374 CCIfType<[v16i8, v8i16, v4i32, v2i64, v4f32, v2f64], CCAssignToStack<16, 16>>, 487 CCIfType<[f32, f64, v4i32, v2i64, v4f32, v2f64] [all...] |
X86InstrFragmentsSIMD.td | 84 SDTypeProfile<1, 3, [SDTCisVT<0, v4f32>, SDTCisSameAs<0,1>, 85 SDTCisVT<2, v4f32>, SDTCisPtrTy<3>]>>; 229 def sse_load_f32 : ComplexPattern<v4f32, 5, "SelectScalarSSELoad", [], 236 def ssmem : Operand<v4f32> { 255 def loadv4f32 : PatFrag<(ops node:$ptr), (v4f32 (load node:$ptr))>; 304 (v4f32 (alignedload node:$ptr))>; 335 def memopv4f32 : PatFrag<(ops node:$ptr), (v4f32 (memop node:$ptr))>; 381 def bc_v4f32 : PatFrag<(ops node:$in), (v4f32 (bitconvert node:$in))>;
|
/external/llvm/test/CodeGen/Thumb2/ |
crash.ll | 56 %vld1 = tail call <4 x float> @llvm.arm.neon.vld1.v4f32(i8* %p, i32 1) 58 tail call void @llvm.arm.neon.vst1.v4f32(i8* %p, <4 x float> %vld1, i32 1) 62 declare <4 x float> @llvm.arm.neon.vld1.v4f32(i8*, i32) nounwind readonly 64 declare void @llvm.arm.neon.vst1.v4f32(i8*, <4 x float>, i32) nounwind
|
/external/llvm/lib/Target/AArch64/ |
AArch64InstrNEON.td | 163 (ResTy4S (opnode4S (v4f32 VPR128:$Rn), (v4f32 VPR128:$Rm))))], 185 v2f32, v4f32, v2f64, 1>; 191 v2f32, v4f32, v2f64, 0>; 197 v2f32, v4f32, v2f64, 1>; 264 def FMLAvvv_4S: NeonI_3VSame_Constraint_impl<"fmla", ".4s", VPR128, v4f32, 271 def FMLSvvv_4S: NeonI_3VSame_Constraint_impl<"fmls", ".4s", VPR128, v4f32, 281 def : Pat<(v4f32 (fma VPR128:$Rn, VPR128:$Rm, VPR128:$Ra)), 288 def : Pat<(v4f32 (fma (fneg VPR128:$Rn), VPR128:$Rm, VPR128:$Ra)), 296 v2f32, v4f32, v2f64, 0> [all...] |
/external/llvm/test/CodeGen/AArch64/ |
neon-fma.ll | 50 declare <4 x float> @llvm.fma.v4f32(<4 x float>, <4 x float>, <4 x float>) 61 %val = call <4 x float> @llvm.fma.v4f32(<4 x float> %A, <4 x float> %B, <4 x float> %C) 81 %val = call <4 x float> @llvm.fma.v4f32(<4 x float> %negA, <4 x float> %B, <4 x float> %C) 93 declare <4 x float> @llvm.fmuladd.v4f32(<4 x float>, <4 x float>, <4 x float>) 104 %val = call <4 x float> @llvm.fmuladd.v4f32(<4 x float> %A, <4 x float> %B, <4 x float> %C)
|
neon-add-pairwise.ll | 69 declare <4 x float> @llvm.arm.neon.vpadd.v4f32(<4 x float>, <4 x float>) 81 %val = call <4 x float> @llvm.arm.neon.vpadd.v4f32(<4 x float> %lhs, <4 x float> %rhs)
|
neon-max-min-pairwise.ll | 212 declare <4 x float> @llvm.arm.neon.vpmaxs.v4f32(<4 x float>, <4 x float>) 224 %val = call <4 x float> @llvm.arm.neon.vpmaxs.v4f32(<4 x float> %lhs, <4 x float> %rhs) 237 declare <4 x float> @llvm.arm.neon.vpmins.v4f32(<4 x float>, <4 x float>) 249 %val = call <4 x float> @llvm.arm.neon.vpmins.v4f32(<4 x float> %lhs, <4 x float> %rhs) 262 declare <4 x float> @llvm.aarch64.neon.vpmaxnm.v4f32(<4 x float>, <4 x float>) 274 %val = call <4 x float> @llvm.aarch64.neon.vpmaxnm.v4f32(<4 x float> %lhs, <4 x float> %rhs) 287 declare <4 x float> @llvm.aarch64.neon.vpminnm.v4f32(<4 x float>, <4 x float>) 299 %val = call <4 x float> @llvm.aarch64.neon.vpminnm.v4f32(<4 x float> %lhs, <4 x float> %rhs)
|
neon-max-min.ll | 212 declare <4 x float> @llvm.arm.neon.vmaxs.v4f32(<4 x float>, <4 x float>) 224 %val = call <4 x float> @llvm.arm.neon.vmaxs.v4f32(<4 x float> %lhs, <4 x float> %rhs) 237 declare <4 x float> @llvm.arm.neon.vmins.v4f32(<4 x float>, <4 x float>) 249 %val = call <4 x float> @llvm.arm.neon.vmins.v4f32(<4 x float> %lhs, <4 x float> %rhs) 263 declare <4 x float> @llvm.aarch64.neon.vmaxnm.v4f32(<4 x float>, <4 x float>) 275 %val = call <4 x float> @llvm.aarch64.neon.vmaxnm.v4f32(<4 x float> %lhs, <4 x float> %rhs) 288 declare <4 x float> @llvm.aarch64.neon.vminnm.v4f32(<4 x float>, <4 x float>) 300 %val = call <4 x float> @llvm.aarch64.neon.vminnm.v4f32(<4 x float> %lhs, <4 x float> %rhs)
|
/external/llvm/lib/Target/PowerPC/ |
PPCCallingConv.td | 36 CCIfType<[v16i8, v8i16, v4i32, v4f32], CCAssignToReg<[V2]>> 69 CCIfType<[v16i8, v8i16, v4i32, v4f32], CCAssignToStack<16, 16>> 83 CCIfType<[v16i8, v8i16, v4i32, v4f32],
|
/external/llvm/test/CodeGen/PowerPC/ |
recipest.ll | 8 declare <4 x float> @llvm.sqrt.v4f32(<4 x float>) 105 %x = call <4 x float> @llvm.sqrt.v4f32(<4 x float> %b) 215 %r = call <4 x float> @llvm.sqrt.v4f32(<4 x float> %a)
|
vec_fmuladd.ll | 7 declare <4 x float> @llvm.fmuladd.v4f32(<4 x float> %val, <4 x float>, <4 x float>) 23 %fmuladd = call <4 x float> @llvm.fmuladd.v4f32 (<4 x float> %x, <4 x float> %x, <4 x float> %x)
|
vec_sqrt.ll | 10 declare <4 x float> @llvm.sqrt.v4f32(<4 x float> %val) 29 %sqrt = call <4 x float> @llvm.sqrt.v4f32 (<4 x float> %x)
|
/external/llvm/lib/Target/R600/ |
AMDGPUCallingConv.td | 41 CCIfType<[v4i32, v4f32], CCAssignToStack <16, 16>>,
|
R600Instructions.td | 379 multiclass TexPattern<bits<32> TextureOp, Instruction inst, ValueType vt = v4f32> { 474 (INSERT_SUBREG (v4f32 (IMPLICIT_DEF)), R600_Reg32:$reg, sub0), 480 (INSERT_SUBREG (v4f32 (IMPLICIT_DEF)), R600_Reg32:$reg, sub0), 486 (v4f32 (IMPLICIT_DEF)), imm:$type, 0, 7, 7, 7, 7, cf_inst, 0) 491 (v4f32 (IMPLICIT_DEF)), 1, 60, 7, 7, 7, 7, cf_inst, 0) 494 def : Pat<(EXPORT (v4f32 R600_Reg128:$src), (i32 imm:$base), (i32 imm:$type), 505 def : Pat<(int_R600_store_stream_output (v4f32 R600_Reg128:$src), 510 def : Pat<(int_R600_store_stream_output (v4f32 R600_Reg128:$src), 515 def : Pat<(int_R600_store_stream_output (v4f32 R600_Reg128:$src), 520 def : Pat<(int_R600_store_stream_output (v4f32 R600_Reg128:$src) [all...] |