/external/llvm/test/CodeGen/X86/ |
D | avx512-round.ll | 6 %res = call <16 x float> @llvm.floor.v16f32(<16 x float> %a) 9 declare <16 x float> @llvm.floor.v16f32(<16 x float> %p) 22 %res = call <16 x float> @llvm.ceil.v16f32(<16 x float> %a) 25 declare <16 x float> @llvm.ceil.v16f32(<16 x float> %p) 38 %res = call <16 x float> @llvm.trunc.v16f32(<16 x float> %a) 41 declare <16 x float> @llvm.trunc.v16f32(<16 x float> %p) 54 %res = call <16 x float> @llvm.rint.v16f32(<16 x float> %a) 57 declare <16 x float> @llvm.rint.v16f32(<16 x float> %p) 70 %res = call <16 x float> @llvm.nearbyint.v16f32(<16 x float> %a) 73 declare <16 x float> @llvm.nearbyint.v16f32(<16 x float> %p)
|
D | wide-fma-contraction.ll | 25 …%ret = tail call <16 x float> @llvm.fmuladd.v16f32(<16 x float> %a, <16 x float> %b, <16 x float> … 29 declare <16 x float> @llvm.fmuladd.v16f32(<16 x float>, <16 x float>, <16 x float>) nounwind readno…
|
/external/swiftshader/third_party/llvm-7.0/llvm/test/Transforms/LoopVectorize/X86/ |
D | gather_scatter.ll | 34 ; AVX512-NEXT: [[WIDE_MASKED_GATHER:%.*]] = call <16 x float> @llvm.masked.gather.v16f32.v16p0f3… 38 ; AVX512-NEXT: call void @llvm.masked.store.v16f32.p0v16f32(<16 x float> [[TMP7]], <16 x float>*… 49 ; AVX512-NEXT: [[WIDE_MASKED_GATHER_1:%.*]] = call <16 x float> @llvm.masked.gather.v16f32.v16p0… 53 ; AVX512-NEXT: call void @llvm.masked.store.v16f32.p0v16f32(<16 x float> [[TMP17]], <16 x float>… 64 ; AVX512-NEXT: [[WIDE_MASKED_GATHER_2:%.*]] = call <16 x float> @llvm.masked.gather.v16f32.v16p0… 68 ; AVX512-NEXT: call void @llvm.masked.store.v16f32.p0v16f32(<16 x float> [[TMP27]], <16 x float>… 79 ; AVX512-NEXT: [[WIDE_MASKED_GATHER_3:%.*]] = call <16 x float> @llvm.masked.gather.v16f32.v16p0… 83 ; AVX512-NEXT: call void @llvm.masked.store.v16f32.p0v16f32(<16 x float> [[TMP37]], <16 x float>… 167 ; AVX512-NEXT: [[WIDE_MASKED_GATHER7:%.*]] = call <16 x float> @llvm.masked.gather.v16f32.v16p0f… 170 ; AVX512-NEXT: call void @llvm.masked.scatter.v16f32.v16p0f32(<16 x float> [[TMP3]], <16 x float… [all …]
|
/external/swiftshader/third_party/llvm-7.0/llvm/lib/Target/X86/ |
D | X86InstrVecCompiler.td | 91 def : Pat<(v8f64 (bitconvert (v16f32 VR512:$src))), (v8f64 VR512:$src)>; 92 def : Pat<(v16f32 (bitconvert (v8i64 VR512:$src))), (v16f32 VR512:$src)>; 93 def : Pat<(v16f32 (bitconvert (v16i32 VR512:$src))), (v16f32 VR512:$src)>; 94 def : Pat<(v16f32 (bitconvert (v32i16 VR512:$src))), (v16f32 VR512:$src)>; 95 def : Pat<(v16f32 (bitconvert (v64i8 VR512:$src))), (v16f32 VR512:$src)>; 96 def : Pat<(v16f32 (bitconvert (v8f64 VR512:$src))), (v16f32 VR512:$src)>; 101 def : Pat<(v8i64 (bitconvert (v16f32 VR512:$src))), (v8i64 VR512:$src)>; 103 def : Pat<(v16i32 (bitconvert (v16f32 VR512:$src))), (v16i32 VR512:$src)>; 111 def : Pat<(v32i16 (bitconvert (v16f32 VR512:$src))), (v32i16 VR512:$src)>; 116 def : Pat<(v64i8 (bitconvert (v16f32 VR512:$src))), (v64i8 VR512:$src)>; [all …]
|
D | X86TargetTransformInfo.cpp | 494 { ISD::FADD, MVT::v16f32, 1 }, // Skylake from http://www.agner.org/ in getArithmeticInstrCost() 495 { ISD::FSUB, MVT::v16f32, 1 }, // Skylake from http://www.agner.org/ in getArithmeticInstrCost() 496 { ISD::FMUL, MVT::v16f32, 1 }, // Skylake from http://www.agner.org/ in getArithmeticInstrCost() 901 { TTI::SK_Broadcast, MVT::v16f32, 1 }, // vbroadcastps in getShuffleCost() 906 { TTI::SK_Reverse, MVT::v16f32, 1 }, // vpermps in getShuffleCost() 913 { TTI::SK_PermuteSingleSrc, MVT::v16f32, 1 }, // vpermps in getShuffleCost() 925 { TTI::SK_PermuteTwoSrc, MVT::v16f32, 1 }, // vpermt2ps in getShuffleCost() 1183 { ISD::FP_EXTEND, MVT::v8f64, MVT::v16f32, 3 }, in getCastInstrCost() 1204 { ISD::SINT_TO_FP, MVT::v16f32, MVT::v16i1, 3 }, in getCastInstrCost() 1206 { ISD::SINT_TO_FP, MVT::v16f32, MVT::v16i8, 2 }, in getCastInstrCost() [all …]
|
/external/swiftshader/third_party/llvm-7.0/llvm/test/CodeGen/X86/ |
D | vmaskmov-offset.ll | 4 declare void @llvm.masked.store.v16f32.p0v16f32(<16 x float>, <16 x float>*, i32, <16 x i1>) 5 declare <16 x float> @llvm.masked.load.v16f32.p0v16f32(<16 x float>*, i32, <16 x i1>, <16 x float>) 19 …%masked_loaded_vec = call <16 x float> @llvm.masked.load.v16f32.p0v16f32(<16 x float>* nonnull %st… 20 …call void @llvm.masked.store.v16f32.p0v16f32(<16 x float> %masked_loaded_vec, <16 x float>* nonnul…
|
D | vec-copysign-avx512.ll | 41 define <16 x float> @v16f32(<16 x float> %a, <16 x float> %b) nounwind { 42 ; AVX512VL-LABEL: v16f32: 49 ; AVX512VLDQ-LABEL: v16f32: 55 %tmp = tail call <16 x float> @llvm.copysign.v16f32( <16 x float> %a, <16 x float> %b ) 108 declare <16 x float> @llvm.copysign.v16f32(<16 x float> %Mag, <16 x float> %Sgn)
|
D | wide-fma-contraction.ll | 41 …%ret = tail call <16 x float> @llvm.fmuladd.v16f32(<16 x float> %a, <16 x float> %b, <16 x float> … 45 declare <16 x float> @llvm.fmuladd.v16f32(<16 x float>, <16 x float>, <16 x float>) nounwind readno…
|
D | avx512-rndscale.ll | 9 declare <16 x float> @llvm.floor.v16f32(<16 x float> %p) 15 declare <16 x float> @llvm.ceil.v16f32(<16 x float> %p) 21 declare <16 x float> @llvm.trunc.v16f32(<16 x float> %p) 27 declare <16 x float> @llvm.rint.v16f32(<16 x float> %p) 33 declare <16 x float> @llvm.nearbyint.v16f32(<16 x float> %p) 85 %t = call <16 x float> @llvm.floor.v16f32(<16 x float> %p) 145 %t = call <16 x float> @llvm.floor.v16f32(<16 x float> %p) 222 %t = call <16 x float> @llvm.floor.v16f32(<16 x float> %p) 294 %t = call <16 x float> @llvm.floor.v16f32(<16 x float> %p) 372 %t = call <16 x float> @llvm.floor.v16f32(<16 x float> %p) [all …]
|
D | bitcast-setcc-512.ll | 143 define i16 @v16f32(<16 x float> %a, <16 x float> %b) { 144 ; SSE-LABEL: v16f32: 157 ; AVX1-LABEL: v16f32: 171 ; AVX2-LABEL: v16f32: 184 ; AVX512F-LABEL: v16f32: 192 ; AVX512BW-LABEL: v16f32:
|
/external/swiftshader/third_party/llvm-7.0/llvm/test/Analysis/CostModel/X86/ |
D | fround.ll | 21 …ated cost of 172 for instruction: %V16F32 = call <16 x float> @llvm.ceil.v16f32(<16 x float> undef) 32 …imated cost of 8 for instruction: %V16F32 = call <16 x float> @llvm.ceil.v16f32(<16 x float> undef) 43 …imated cost of 4 for instruction: %V16F32 = call <16 x float> @llvm.ceil.v16f32(<16 x float> undef) 54 …imated cost of 1 for instruction: %V16F32 = call <16 x float> @llvm.ceil.v16f32(<16 x float> undef) 65 …imated cost of 8 for instruction: %V16F32 = call <16 x float> @llvm.ceil.v16f32(<16 x float> undef) 76 …imated cost of 8 for instruction: %V16F32 = call <16 x float> @llvm.ceil.v16f32(<16 x float> undef) 87 …imated cost of 4 for instruction: %V16F32 = call <16 x float> @llvm.ceil.v16f32(<16 x float> undef) 97 %V16F32 = call <16 x float> @llvm.ceil.v16f32(<16 x float> undef) 112 …ted cost of 172 for instruction: %V16F32 = call <16 x float> @llvm.floor.v16f32(<16 x float> undef) 123 …mated cost of 8 for instruction: %V16F32 = call <16 x float> @llvm.floor.v16f32(<16 x float> undef) [all …]
|
D | arith-fma.ll | 13 …timated cost of 4 for instruction: %V16F32 = call <16 x float> @llvm.fma.v16f32(<16 x float> undef… 24 …timated cost of 4 for instruction: %V16F32 = call <16 x float> @llvm.fma.v16f32(<16 x float> undef… 35 …timated cost of 1 for instruction: %V16F32 = call <16 x float> @llvm.fma.v16f32(<16 x float> undef… 45 …%V16F32 = call <16 x float> @llvm.fma.v16f32(<16 x float> undef, <16 x float> undef, <16 x float> … 58 declare <16 x float> @llvm.fma.v16f32(<16 x float>, <16 x float>, <16 x float>)
|
D | masked-intrinsic-cost.ll | 297 …cost of 30 for instruction: %res = call <16 x float> @llvm.masked.gather.v16f32.v16p0f32(<16 x flo… 303 …cost of 24 for instruction: %res = call <16 x float> @llvm.masked.gather.v16f32.v16p0f32(<16 x flo… 309 …cost of 18 for instruction: %res = call <16 x float> @llvm.masked.gather.v16f32.v16p0f32(<16 x flo… 315 …cost of 18 for instruction: %res = call <16 x float> @llvm.masked.gather.v16f32.v16p0f32(<16 x flo… 321 …%res = call <16 x float> @llvm.masked.gather.v16f32.v16p0f32(<16 x float*> %gep.v, i32 4, <16 x i1… 329 …cost of 62 for instruction: %res = call <16 x float> @llvm.masked.gather.v16f32.v16p0f32(<16 x flo… 335 …cost of 24 for instruction: %res = call <16 x float> @llvm.masked.gather.v16f32.v16p0f32(<16 x flo… 341 …cost of 18 for instruction: %res = call <16 x float> @llvm.masked.gather.v16f32.v16p0f32(<16 x flo… 347 …cost of 18 for instruction: %res = call <16 x float> @llvm.masked.gather.v16f32.v16p0f32(<16 x flo… 353 …%res = call <16 x float> @llvm.masked.gather.v16f32.v16p0f32(<16 x float*> %gep.v, i32 4, <16 x i1… [all …]
|
D | arith-fp.ll | 509 …ated cost of 224 for instruction: %V16F32 = call <16 x float> @llvm.sqrt.v16f32(<16 x float> undef) 520 …mated cost of 72 for instruction: %V16F32 = call <16 x float> @llvm.sqrt.v16f32(<16 x float> undef) 531 …mated cost of 56 for instruction: %V16F32 = call <16 x float> @llvm.sqrt.v16f32(<16 x float> undef) 542 …mated cost of 28 for instruction: %V16F32 = call <16 x float> @llvm.sqrt.v16f32(<16 x float> undef) 553 …imated cost of 1 for instruction: %V16F32 = call <16 x float> @llvm.sqrt.v16f32(<16 x float> undef) 564 …ated cost of 160 for instruction: %V16F32 = call <16 x float> @llvm.sqrt.v16f32(<16 x float> undef) 575 …ated cost of 148 for instruction: %V16F32 = call <16 x float> @llvm.sqrt.v16f32(<16 x float> undef) 586 …mated cost of 56 for instruction: %V16F32 = call <16 x float> @llvm.sqrt.v16f32(<16 x float> undef) 596 %V16F32 = call <16 x float> @llvm.sqrt.v16f32(<16 x float> undef) 611 …imated cost of 8 for instruction: %V16F32 = call <16 x float> @llvm.fabs.v16f32(<16 x float> undef) [all …]
|
/external/llvm/test/Transforms/LoopVectorize/X86/ |
D | gather_scatter.ll | 21 ;AVX512: llvm.masked.gather.v16f32 22 ;AVX512: llvm.masked.store.v16f32 99 ;AVX512: llvm.masked.gather.v16f32 100 ;AVX512: llvm.masked.store.v16f32 174 ;AVX512: llvm.masked.gather.v16f32 177 ;AVX512: llvm.masked.scatter.v16f32 236 declare void @llvm.masked.scatter.v16f32(<16 x float>, <16 x float*>, i32, <16 x i1>)
|
/external/llvm/include/llvm/CodeGen/ |
D | MachineValueType.h | 114 v16f32 = 59, // 16 x f32 enumerator 264 return (SimpleTy == MVT::v16f32 || SimpleTy == MVT::v8f64 || in is512BitVector() 363 case v16f32: return f32; in getVectorElementType() 394 case v16f32: return 16; in getVectorNumElements() 501 case v16f32: in getSizeInBits() 653 if (NumElements == 16) return MVT::v16f32; in getVectorVT()
|
/external/swiftshader/third_party/llvm-7.0/llvm/include/llvm/Support/ |
D | MachineValueType.h | 156 v16f32 = 92, // 16 x f32 enumerator 368 return (SimpleTy == MVT::v16f32 || SimpleTy == MVT::v8f64 || in is512BitVector() 501 case v16f32: in getVectorElementType() 547 case v16f32: in getVectorNumElements() 732 case v16f32: in getSizeInBits() 894 if (NumElements == 16) return MVT::v16f32; in getVectorVT()
|
/external/swiftshader/third_party/llvm-7.0/llvm/test/Verifier/ |
D | scatter_gather.ll | 6 …%res = call <16 x float> @llvm.masked.gather.v16f32.v16p0f32(<16 x float*> %ptrs, i32 4, <16 x i1>… 9 declare <16 x float> @llvm.masked.gather.v16f32.v16p0f32(<16 x float*>, i32, <16 x i1>*, <16 x floa… 70 …call void @llvm.masked.scatter.v16f32.v16p0f32(<16 x float> %value, <16 x float*> %ptrs, i32 4, <1… 73 declare void @llvm.masked.scatter.v16f32.v16p0f32(<16 x float>, <16 x float*>, i32, <16 x i1>*)
|
/external/llvm/lib/Target/ARM/ |
D | ARMTargetTransformInfo.cpp | 148 { ISD::SINT_TO_FP, MVT::v16f32, MVT::v16i16, 8 }, in getCastInstrCost() 149 { ISD::UINT_TO_FP, MVT::v16f32, MVT::v16i16, 8 }, in getCastInstrCost() 150 { ISD::SINT_TO_FP, MVT::v16f32, MVT::v16i32, 4 }, in getCastInstrCost() 151 { ISD::UINT_TO_FP, MVT::v16f32, MVT::v16i32, 4 }, in getCastInstrCost() 175 { ISD::FP_TO_SINT, MVT::v16i16, MVT::v16f32, 8 }, in getCastInstrCost() 176 { ISD::FP_TO_UINT, MVT::v16i16, MVT::v16f32, 8 } in getCastInstrCost()
|
/external/llvm/test/Analysis/CostModel/X86/ |
D | masked-intrinsic-cost.ll | 131 …%res = call <16 x float> @llvm.masked.gather.v16f32(<16 x float*> %gep.v, i32 4, <16 x i1> <i1 tru… 149 …%res = call <16 x float> @llvm.masked.gather.v16f32(<16 x float*> %gep.v, i32 4, <16 x i1> %mask, … 167 …%res = call <16 x float> @llvm.masked.gather.v16f32(<16 x float*> %gep.v, i32 4, <16 x i1> %mask, … 188 …%res = call <16 x float> @llvm.masked.gather.v16f32(<16 x float*> %gep.random, i32 4, <16 x i1> <i… 280 declare <16 x float> @llvm.masked.gather.v16f32(<16 x float*> %gep.v, i32, <16 x i1> %mask, <16 x f… 290 declare void @llvm.masked.store.v16f32.p0v16f32(<16 x float>, <16 x float>*, i32, <16 x i1>) 291 declare <16 x float> @llvm.masked.load.v16f32.p0v16f32(<16 x float>*, i32, <16 x i1>, <16 x float>)
|
/external/swiftshader/third_party/llvm-7.0/llvm/lib/Target/ARM/ |
D | ARMTargetTransformInfo.cpp | 209 { ISD::SINT_TO_FP, MVT::v16f32, MVT::v16i16, 8 }, in getCastInstrCost() 210 { ISD::UINT_TO_FP, MVT::v16f32, MVT::v16i16, 8 }, in getCastInstrCost() 211 { ISD::SINT_TO_FP, MVT::v16f32, MVT::v16i32, 4 }, in getCastInstrCost() 212 { ISD::UINT_TO_FP, MVT::v16f32, MVT::v16i32, 4 }, in getCastInstrCost() 236 { ISD::FP_TO_SINT, MVT::v16i16, MVT::v16f32, 8 }, in getCastInstrCost() 237 { ISD::FP_TO_UINT, MVT::v16i16, MVT::v16f32, 8 } in getCastInstrCost()
|
/external/swiftshader/third_party/llvm-7.0/llvm/lib/Target/AMDGPU/ |
D | AMDGPUCallingConv.td | 114 …CCIfType<[i64, f64, v2i32, v2f32, v4i32, v4f32, v8i32, v8f32, v16i32, v16f32, v2i64, v2f64, v4i16,… 119 CCIfType<[v16i32, v16f32], CCAssignToStack<64, 4>> 131 …CCIfType<[i64, f64, v2i32, v2f32, v4i32, v4f32, v8i32, v8f32, v16i32, v16f32, v2i64, v2f64, v4i16,…
|
/external/llvm/lib/Target/X86/ |
D | X86CallingConv.td | 68 CCIfType<[v64i8, v32i16, v16i32, v8i64, v16f32, v8f64], 122 CCIfType<[v16f32, v8f64, v16i32, v8i64], 149 CCIfType<[v64i8, v32i16, v16i32, v8i64, v16f32, v8f64], 346 CCIfNotVarArg<CCIfType<[v64i8, v32i16, v16i32, v8i64, v16f32, v8f64], 366 CCIfType<[v16i32, v8i64, v16f32, v8f64], 406 CCIfType<[v16i32, v16f32, v8f64, v8i64], CCPassIndirect<i64>>, 449 CCIfType<[v64i8, v32i16, v16i32, v8i64, v16f32, v8f64], 524 CCIfType<[v64i8, v32i16, v16i32, v8i64, v16f32, v8f64], 541 CCIfNotVarArg<CCIfType<[v64i8, v32i16, v16i32, v8i64, v16f32, v8f64], 560 CCIfNotVarArg<CCIfType<[v64i8, v32i16, v16i32, v8i64, v16f32, v8f64], [all …]
|
/external/swiftshader/third_party/llvm-7.0/llvm/test/CodeGen/AMDGPU/ |
D | ftrunc.ll | 10 declare <16 x float> @llvm.trunc.v16f32(<16 x float>) nounwind readnone 117 %y = call <16 x float> @llvm.trunc.v16f32(<16 x float> %x) nounwind readnone
|
/external/llvm/test/CodeGen/AMDGPU/ |
D | ftrunc.ll | 10 declare <16 x float> @llvm.trunc.v16f32(<16 x float>) nounwind readnone 117 %y = call <16 x float> @llvm.trunc.v16f32(<16 x float> %x) nounwind readnone
|