diff options
Diffstat (limited to 'lib/Target/X86/X86InstrAVX512.td')
| -rw-r--r-- | lib/Target/X86/X86InstrAVX512.td | 39 |
1 files changed, 39 insertions, 0 deletions
diff --git a/lib/Target/X86/X86InstrAVX512.td b/lib/Target/X86/X86InstrAVX512.td index c38c13bb9757..71d395244b4a 100644 --- a/lib/Target/X86/X86InstrAVX512.td +++ b/lib/Target/X86/X86InstrAVX512.td @@ -8631,6 +8631,20 @@ multiclass avx512_unary_rm_vl_all<bits<8> opc_b, bits<8> opc_w, defm VPABS : avx512_unary_rm_vl_all<0x1C, 0x1D, 0x1E, 0x1F, "vpabs", abs>; +// VPABS: Use 512bit version to implement 128/256 bit in case NoVLX. +let Predicates = [HasAVX512, NoVLX] in { + def : Pat<(v4i64 (abs VR256X:$src)), + (EXTRACT_SUBREG + (VPABSQZrr + (INSERT_SUBREG (v8i64 (IMPLICIT_DEF)), VR256X:$src, sub_ymm)), + sub_ymm)>; + def : Pat<(v2i64 (abs VR128X:$src)), + (EXTRACT_SUBREG + (VPABSQZrr + (INSERT_SUBREG (v8i64 (IMPLICIT_DEF)), VR128X:$src, sub_xmm)), + sub_xmm)>; +} + multiclass avx512_ctlz<bits<8> opc, string OpcodeStr, Predicate prd>{ defm NAME : avx512_unary_rm_vl_dq<opc, opc, OpcodeStr, ctlz, prd>; @@ -8639,6 +8653,31 @@ multiclass avx512_ctlz<bits<8> opc, string OpcodeStr, Predicate prd>{ defm VPLZCNT : avx512_ctlz<0x44, "vplzcnt", HasCDI>; defm VPCONFLICT : avx512_unary_rm_vl_dq<0xC4, 0xC4, "vpconflict", X86Conflict, HasCDI>; +// VPLZCNT: Use 512bit version to implement 128/256 bit in case NoVLX. +let Predicates = [HasCDI, NoVLX] in { + def : Pat<(v4i64 (ctlz VR256X:$src)), + (EXTRACT_SUBREG + (VPLZCNTQZrr + (INSERT_SUBREG (v8i64 (IMPLICIT_DEF)), VR256X:$src, sub_ymm)), + sub_ymm)>; + def : Pat<(v2i64 (ctlz VR128X:$src)), + (EXTRACT_SUBREG + (VPLZCNTQZrr + (INSERT_SUBREG (v8i64 (IMPLICIT_DEF)), VR128X:$src, sub_xmm)), + sub_xmm)>; + + def : Pat<(v8i32 (ctlz VR256X:$src)), + (EXTRACT_SUBREG + (VPLZCNTDZrr + (INSERT_SUBREG (v16i32 (IMPLICIT_DEF)), VR256X:$src, sub_ymm)), + sub_ymm)>; + def : Pat<(v4i32 (ctlz VR128X:$src)), + (EXTRACT_SUBREG + (VPLZCNTDZrr + (INSERT_SUBREG (v16i32 (IMPLICIT_DEF)), VR128X:$src, sub_xmm)), + sub_xmm)>; +} + //===---------------------------------------------------------------------===// // Replicate Single FP - MOVSHDUP and MOVSLDUP //===---------------------------------------------------------------------===// |