Lines Matching full:xop
2 ; RUN: llc < %s -fast-isel -mtriple=i686-unknown-unknown -mattr=+avx,+fma4,+xop | FileCheck %s --check-prefix=ALL --check-prefix=X32
3 ; RUN: llc < %s -fast-isel -mtriple=x86_64-unknown-unknown -mattr=+avx,+fma4,+xop | FileCheck %s --check-prefix=ALL --check-prefix=X64
5 ; NOTE: This should use IR equivalent to what is generated by clang/test/CodeGen/xop-builtins.c
20 %res = call <8 x i16> @llvm.x86.xop.vpmacssww(<8 x i16> %arg0, <8 x i16> %arg1, <8 x i16> %arg2)
24 declare <8 x i16> @llvm.x86.xop.vpmacssww(<8 x i16>, <8 x i16>, <8 x i16>) nounwind readnone
39 %res = call <8 x i16> @llvm.x86.xop.vpmacsww(<8 x i16> %arg0, <8 x i16> %arg1, <8 x i16> %arg2)
43 declare <8 x i16> @llvm.x86.xop.vpmacsww(<8 x i16>, <8 x i16>, <8 x i16>) nounwind readnone
58 %res = call <4 x i32> @llvm.x86.xop.vpmacsswd(<8 x i16> %arg0, <8 x i16> %arg1, <4 x i32> %arg2)
62 declare <4 x i32> @llvm.x86.xop.vpmacsswd(<8 x i16>, <8 x i16>, <4 x i32>) nounwind readnone
77 %res = call <4 x i32> @llvm.x86.xop.vpmacswd(<8 x i16> %arg0, <8 x i16> %arg1, <4 x i32> %arg2)
81 declare <4 x i32> @llvm.x86.xop.vpmacswd(<8 x i16>, <8 x i16>, <4 x i32>) nounwind readnone
96 %res = call <4 x i32> @llvm.x86.xop.vpmacssdd(<4 x i32> %arg0, <4 x i32> %arg1, <4 x i32> %arg2)
100 declare <4 x i32> @llvm.x86.xop.vpmacssdd(<4 x i32>, <4 x i32>, <4 x i32>) nounwind readnone
115 %res = call <4 x i32> @llvm.x86.xop.vpmacsdd(<4 x i32> %arg0, <4 x i32> %arg1, <4 x i32> %arg2)
119 declare <4 x i32> @llvm.x86.xop.vpmacsdd(<4 x i32>, <4 x i32>, <4 x i32>) nounwind readnone
133 %res = call <2 x i64> @llvm.x86.xop.vpmacssdql(<4 x i32> %arg0, <4 x i32> %arg1, <2 x i64> %a2)
136 declare <2 x i64> @llvm.x86.xop.vpmacssdql(<4 x i32>, <4 x i32>, <2 x i64>) nounwind readnone
150 %res = call <2 x i64> @llvm.x86.xop.vpmacsdql(<4 x i32> %arg0, <4 x i32> %arg1, <2 x i64> %a2)
153 declare <2 x i64> @llvm.x86.xop.vpmacsdql(<4 x i32>, <4 x i32>, <2 x i64>) nounwind readnone
167 %res = call <2 x i64> @llvm.x86.xop.vpmacssdqh(<4 x i32> %arg0, <4 x i32> %arg1, <2 x i64> %a2)
170 declare <2 x i64> @llvm.x86.xop.vpmacssdqh(<4 x i32>, <4 x i32>, <2 x i64>) nounwind readnone
184 %res = call <2 x i64> @llvm.x86.xop.vpmacsdqh(<4 x i32> %arg0, <4 x i32> %arg1, <2 x i64> %a2)
187 declare <2 x i64> @llvm.x86.xop.vpmacsdqh(<4 x i32>, <4 x i32>, <2 x i64>) nounwind readnone
202 %res = call <4 x i32> @llvm.x86.xop.vpmadcsswd(<8 x i16> %arg0, <8 x i16> %arg1, <4 x i32> %arg2)
206 declare <4 x i32> @llvm.x86.xop.vpmadcsswd(<8 x i16>, <8 x i16>, <4 x i32>) nounwind readnone
221 %res = call <4 x i32> @llvm.x86.xop.vpmadcswd(<8 x i16> %arg0, <8 x i16> %arg1, <4 x i32> %arg2)
225 declare <4 x i32> @llvm.x86.xop.vpmadcswd(<8 x i16>, <8 x i16>, <4 x i32>) nounwind readnone
238 %res = call <8 x i16> @llvm.x86.xop.vphaddbw(<16 x i8> %arg0)
242 declare <8 x i16> @llvm.x86.xop.vphaddbw(<16 x i8>) nounwind readnone
255 %res = call <4 x i32> @llvm.x86.xop.vphaddbd(<16 x i8> %arg0)
259 declare <4 x i32> @llvm.x86.xop.vphaddbd(<16 x i8>) nounwind readnone
272 %res = call <2 x i64> @llvm.x86.xop.vphaddbq(<16 x i8> %arg0)
275 declare <2 x i64> @llvm.x86.xop.vphaddbq(<16 x i8>) nounwind readnone
288 %res = call <4 x i32> @llvm.x86.xop.vphaddwd(<8 x i16> %arg0)
292 declare <4 x i32> @llvm.x86.xop.vphaddwd(<8 x i16>) nounwind readnone
305 %res = call <2 x i64> @llvm.x86.xop.vphaddwq(<8 x i16> %arg0)
308 declare <2 x i64> @llvm.x86.xop.vphaddwq(<8 x i16>) nounwind readnone
321 %res = call <2 x i64> @llvm.x86.xop.vphadddq(<4 x i32> %arg0)
324 declare <2 x i64> @llvm.x86.xop.vphadddq(<4 x i32>) nounwind readnone
337 %res = call <8 x i16> @llvm.x86.xop.vphaddubw(<16 x i8> %arg0)
341 declare <8 x i16> @llvm.x86.xop.vphaddubw(<16 x i8>) nounwind readnone
354 %res = call <4 x i32> @llvm.x86.xop.vphaddubd(<16 x i8> %arg0)
358 declare <4 x i32> @llvm.x86.xop.vphaddubd(<16 x i8>) nounwind readnone
371 %res = call <2 x i64> @llvm.x86.xop.vphaddubq(<16 x i8> %arg0)
374 declare <2 x i64> @llvm.x86.xop.vphaddubq(<16 x i8>) nounwind readnone
387 %res = call <4 x i32> @llvm.x86.xop.vphadduwd(<8 x i16> %arg0)
391 declare <4 x i32> @llvm.x86.xop.vphadduwd(<8 x i16>) nounwind readnone
405 %res = call <2 x i64> @llvm.x86.xop.vphadduwq(<8 x i16> %arg0)
408 declare <2 x i64> @llvm.x86.xop.vphadduwq(<8 x i16>) nounwind readnone
421 %res = call <2 x i64> @llvm.x86.xop.vphaddudq(<4 x i32> %arg0)
424 declare <2 x i64> @llvm.x86.xop.vphaddudq(<4 x i32>) nounwind readnone
437 %res = call <8 x i16> @llvm.x86.xop.vphsubbw(<16 x i8> %arg0)
441 declare <8 x i16> @llvm.x86.xop.vphsubbw(<16 x i8>) nounwind readnone
454 %res = call <4 x i32> @llvm.x86.xop.vphsubwd(<8 x i16> %arg0)
458 declare <4 x i32> @llvm.x86.xop.vphsubwd(<8 x i16>) nounwind readnone
471 %res = call <2 x i64> @llvm.x86.xop.vphsubdq(<4 x i32> %arg0)
474 declare <2 x i64> @llvm.x86.xop.vphsubdq(<4 x i32>) nounwind readnone
494 %res = call <2 x i64> @llvm.x86.xop.vpcmov(<2 x i64> %a0, <2 x i64> %a1, <2 x i64> %a2)
497 declare <2 x i64> @llvm.x86.xop.vpcmov(<2 x i64>, <2 x i64>, <2 x i64>) nounwind readnone
509 %res = call <4 x i64> @llvm.x86.xop.vpcmov.256(<4 x i64> %a0, <4 x i64> %a1, <4 x i64> %a2)
512 declare <4 x i64> @llvm.x86.xop.vpcmov.256(<4 x i64>, <4 x i64>, <4 x i64>) nounwind readnone
527 %res = call <16 x i8> @llvm.x86.xop.vpperm(<16 x i8> %arg0, <16 x i8> %arg1, <16 x i8> %arg2)
531 declare <16 x i8> @llvm.x86.xop.vpperm(<16 x i8>, <16 x i8>, <16 x i8>) nounwind readnone
545 %res = call <16 x i8> @llvm.x86.xop.vprotb(<16 x i8> %arg0, <16 x i8> %arg1)
549 declare <16 x i8> @llvm.x86.xop.vprotb(<16 x i8>, <16 x i8>) nounwind readnone
563 %res = call <8 x i16> @llvm.x86.xop.vprotw(<8 x i16> %arg0, <8 x i16> %arg1)
567 declare <8 x i16> @llvm.x86.xop.vprotw(<8 x i16>, <8 x i16>) nounwind readnone
581 %res = call <4 x i32> @llvm.x86.xop.vprotd(<4 x i32> %arg0, <4 x i32> %arg1)
585 declare <4 x i32> @llvm.x86.xop.vprotd(<4 x i32>, <4 x i32>) nounwind readnone
597 %res = call <2 x i64> @llvm.x86.xop.vprotq(<2 x i64> %a0, <2 x i64> %a1)
600 declare <2 x i64> @llvm.x86.xop.vprotq(<2 x i64>, <2 x i64>) nounwind readnone
613 %res = call <16 x i8> @llvm.x86.xop.vprotbi(<16 x i8> %arg0, i8 1)
617 declare <16 x i8> @llvm.x86.xop.vprotbi(<16 x i8>, i8) nounwind readnone
630 %res = call <8 x i16> @llvm.x86.xop.vprotwi(<8 x i16> %arg0, i8 50)
634 declare <8 x i16> @llvm.x86.xop.vprotwi(<8 x i16>, i8) nounwind readnone
647 %res = call <4 x i32> @llvm.x86.xop.vprotdi(<4 x i32> %arg0, i8 -30)
651 declare <4 x i32> @llvm.x86.xop.vprotdi(<4 x i32>, i8) nounwind readnone
663 %res = call <2 x i64> @llvm.x86.xop.vprotqi(<2 x i64> %a0, i8 100)
666 declare <2 x i64> @llvm.x86.xop.vprotqi(<2 x i64>, i8) nounwind readnone
680 %res = call <16 x i8> @llvm.x86.xop.vpshlb(<16 x i8> %arg0, <16 x i8> %arg1)
684 declare <16 x i8> @llvm.x86.xop.vpshlb(<16 x i8>, <16 x i8>) nounwind readnone
698 %res = call <8 x i16> @llvm.x86.xop.vpshlw(<8 x i16> %arg0, <8 x i16> %arg1)
702 declare <8 x i16> @llvm.x86.xop.vpshlw(<8 x i16>, <8 x i16>) nounwind readnone
716 %res = call <4 x i32> @llvm.x86.xop.vpshld(<4 x i32> %arg0, <4 x i32> %arg1)
720 declare <4 x i32> @llvm.x86.xop.vpshld(<4 x i32>, <4 x i32>) nounwind readnone
732 %res = call <2 x i64> @llvm.x86.xop.vpshlq(<2 x i64> %a0, <2 x i64> %a1)
735 declare <2 x i64> @llvm.x86.xop.vpshlq(<2 x i64>, <2 x i64>) nounwind readnone
749 %res = call <16 x i8> @llvm.x86.xop.vpshab(<16 x i8> %arg0, <16 x i8> %arg1)
753 declare <16 x i8> @llvm.x86.xop.vpshab(<16 x i8>, <16 x i8>) nounwind readnone
767 %res = call <8 x i16> @llvm.x86.xop.vpshaw(<8 x i16> %arg0, <8 x i16> %arg1)
771 declare <8 x i16> @llvm.x86.xop.vpshaw(<8 x i16>, <8 x i16>) nounwind readnone
785 %res = call <4 x i32> @llvm.x86.xop.vpshad(<4 x i32> %arg0, <4 x i32> %arg1)
789 declare <4 x i32> @llvm.x86.xop.vpshad(<4 x i32>, <4 x i32>) nounwind readnone
801 %res = call <2 x i64> @llvm.x86.xop.vpshaq(<2 x i64> %a0, <2 x i64> %a1)
804 declare <2 x i64> @llvm.x86.xop.vpshaq(<2 x i64>, <2 x i64>) nounwind readnone
818 %res = call <16 x i8> @llvm.x86.xop.vpcomub(<16 x i8> %arg0, <16 x i8> %arg1, i8 0)
822 declare <16 x i8> @llvm.x86.xop.vpcomub(<16 x i8>, <16 x i8>, i8) nounwind readnone
836 %res = call <8 x i16> @llvm.x86.xop.vpcomuw(<8 x i16> %arg0, <8 x i16> %arg1, i8 0)
840 declare <8 x i16> @llvm.x86.xop.vpcomuw(<8 x i16>, <8 x i16>, i8) nounwind readnone
854 %res = call <4 x i32> @llvm.x86.xop.vpcomud(<4 x i32> %arg0, <4 x i32> %arg1, i8 0)
858 declare <4 x i32> @llvm.x86.xop.vpcomud(<4 x i32>, <4 x i32>, i8) nounwind readnone
870 %res = call <2 x i64> @llvm.x86.xop.vpcomuq(<2 x i64> %a0, <2 x i64> %a1, i8 0)
873 declare <2 x i64> @llvm.x86.xop.vpcomuq(<2 x i64>, <2 x i64>, i8) nounwind readnone
887 %res = call <16 x i8> @llvm.x86.xop.vpcomb(<16 x i8> %arg0, <16 x i8> %arg1, i8 0)
891 declare <16 x i8> @llvm.x86.xop.vpcomb(<16 x i8>, <16 x i8>, i8) nounwind readnone
905 %res = call <8 x i16> @llvm.x86.xop.vpcomw(<8 x i16> %arg0, <8 x i16> %arg1, i8 0)
909 declare <8 x i16> @llvm.x86.xop.vpcomw(<8 x i16>, <8 x i16>, i8) nounwind readnone
923 %res = call <4 x i32> @llvm.x86.xop.vpcomd(<4 x i32> %arg0, <4 x i32> %arg1, i8 0)
927 declare <4 x i32> @llvm.x86.xop.vpcomd(<4 x i32>, <4 x i32>, i8) nounwind readnone
939 %res = call <2 x i64> @llvm.x86.xop.vpcomq(<2 x i64> %a0, <2 x i64> %a1, i8 0)
942 declare <2 x i64> @llvm.x86.xop.vpcomq(<2 x i64>, <2 x i64>, i8) nounwind readnone
954 %res = call <2 x double> @llvm.x86.xop.vpermil2pd(<2 x double> %a0, <2 x double> %a1, <2 x i64> %a2, i8 0)
957 declare <2 x double> @llvm.x86.xop.vpermil2pd(<2 x double>, <2 x double>, <2 x i64>, i8) nounwind readnone
969 %res = call <4 x double> @llvm.x86.xop.vpermil2pd.256(<4 x double> %a0, <4 x double> %a1, <4 x i64> %a2, i8 0)
972 declare <4 x double> @llvm.x86.xop.vpermil2pd.256(<4 x double>, <4 x double>, <4 x i64>, i8) nounwind readnone
985 %res = call <4 x float> @llvm.x86.xop.vpermil2ps(<4 x float> %a0, <4 x float> %a1, <4 x i32> %arg2, i8 0)
988 declare <4 x float> @llvm.x86.xop.vpermil2ps(<4 x float>, <4 x float>, <4 x i32>, i8) nounwind readnone
1001 %res = call <8 x float> @llvm.x86.xop.vpermil2ps.256(<8 x float> %a0, <8 x float> %a1, <8 x i32> %arg2, i8 0)
1004 declare <8 x float> @llvm.x86.xop.vpermil2ps.256(<8 x float>, <8 x float>, <8 x i32>, i8) nounwind readnone
1016 %res = call <4 x float> @llvm.x86.xop.vfrcz.ss(<4 x float> %a0)
1019 declare <4 x float> @llvm.x86.xop.vfrcz.ss(<4 x float>) nounwind readnone
1031 %res = call <2 x double> @llvm.x86.xop.vfrcz.sd(<2 x double> %a0)
1034 declare <2 x double> @llvm.x86.xop.vfrcz.sd(<2 x double>) nounwind readnone
1046 %res = call <4 x float> @llvm.x86.xop.vfrcz.ps(<4 x float> %a0)
1049 declare <4 x float> @llvm.x86.xop.vfrcz.ps(<4 x float>) nounwind readnone
1061 %res = call <2 x double> @llvm.x86.xop.vfrcz.pd(<2 x double> %a0)
1064 declare <2 x double> @llvm.x86.xop.vfrcz.pd(<2 x double>) nounwind readnone
1076 %res = call <8 x float> @llvm.x86.xop.vfrcz.ps.256(<8 x float> %a0)
1079 declare <8 x float> @llvm.x86.xop.vfrcz.ps.256(<8 x float>) nounwind readnone
1091 %res = call <4 x double> @llvm.x86.xop.vfrcz.pd.256(<4 x double> %a0)
1094 declare <4 x double> @llvm.x86.xop.vfrcz.pd.256(<4 x double>) nounwind readnone
