diff options
| author | Charlie Turner <charlie.turner@arm.com> | 2015-11-09 13:10:52 +0000 |
|---|---|---|
| committer | Charlie Turner <charlie.turner@arm.com> | 2015-11-09 13:10:52 +0000 |
| commit | 90dafb1b6db0b72059dee3360e82c46af1fda1d1 (patch) | |
| tree | 67799d2b812fc8e2c51115f93761f172f567068b /llvm/lib | |
| parent | 7b7b06f73786ff8c6139a6337c565886ff4ba60e (diff) | |
| download | bcm5719-llvm-90dafb1b6db0b72059dee3360e82c46af1fda1d1.tar.gz bcm5719-llvm-90dafb1b6db0b72059dee3360e82c46af1fda1d1.zip | |
[AArch64] Add UABDL patterns for log2 shuffle.
Summary:
This matches the sum-of-absdiff patterns emitted by the vectoriser using log2 shuffles.
Relies on D14207 to be able to match the `extract_subvector(..., 0)`
Reviewers: t.p.northover, jmolloy
Subscribers: aemerson, llvm-commits, rengolin
Differential Revision: http://reviews.llvm.org/D14208
llvm-svn: 252465
Diffstat (limited to 'llvm/lib')
| -rw-r--r-- | llvm/lib/Target/AArch64/AArch64InstrInfo.td | 36 |
1 files changed, 34 insertions, 2 deletions
diff --git a/llvm/lib/Target/AArch64/AArch64InstrInfo.td b/llvm/lib/Target/AArch64/AArch64InstrInfo.td index 5df147aa320..3b4fbc87b52 100644 --- a/llvm/lib/Target/AArch64/AArch64InstrInfo.td +++ b/llvm/lib/Target/AArch64/AArch64InstrInfo.td @@ -2622,6 +2622,40 @@ defm FMOV : FPMoveImmediate<"fmov">; // Advanced SIMD two vector instructions. //===----------------------------------------------------------------------===// +defm UABDL : SIMDLongThreeVectorBHSabdl<1, 0b0111, "uabdl", + uabsdiff>; +// Match UABDL in log2-shuffle patterns. +def : Pat<(xor (v8i16 (AArch64vashr v8i16:$src, (i32 15))), + (v8i16 (add (sub (zext (v8i8 V64:$opA)), + (zext (v8i8 V64:$opB))), + (AArch64vashr v8i16:$src, (i32 15))))), + (UABDLv8i8_v8i16 V64:$opA, V64:$opB)>; +def : Pat<(xor (v8i16 (AArch64vashr v8i16:$src, (i32 15))), + (v8i16 (add (sub (zext (extract_high_v16i8 V128:$opA)), + (zext (extract_high_v16i8 V128:$opB))), + (AArch64vashr v8i16:$src, (i32 15))))), + (UABDLv16i8_v8i16 V128:$opA, V128:$opB)>; +def : Pat<(xor (v4i32 (AArch64vashr v4i32:$src, (i32 31))), + (v4i32 (add (sub (zext (v4i16 V64:$opA)), + (zext (v4i16 V64:$opB))), + (AArch64vashr v4i32:$src, (i32 31))))), + (UABDLv4i16_v4i32 V64:$opA, V64:$opB)>; +def : Pat<(xor (v4i32 (AArch64vashr v4i32:$src, (i32 31))), + (v4i32 (add (sub (zext (extract_high_v8i16 V128:$opA)), + (zext (extract_high_v8i16 V128:$opB))), + (AArch64vashr v4i32:$src, (i32 31))))), + (UABDLv8i16_v4i32 V128:$opA, V128:$opB)>; +def : Pat<(xor (v2i64 (AArch64vashr v2i64:$src, (i32 63))), + (v2i64 (add (sub (zext (v2i32 V64:$opA)), + (zext (v2i32 V64:$opB))), + (AArch64vashr v2i64:$src, (i32 63))))), + (UABDLv2i32_v2i64 V64:$opA, V64:$opB)>; +def : Pat<(xor (v2i64 (AArch64vashr v2i64:$src, (i32 63))), + (v2i64 (add (sub (zext (extract_high_v4i32 V128:$opA)), + (zext (extract_high_v4i32 V128:$opB))), + (AArch64vashr v2i64:$src, (i32 63))))), + (UABDLv4i32_v2i64 V128:$opA, V128:$opB)>; + defm ABS : SIMDTwoVectorBHSD<0, 0b01011, "abs", int_aarch64_neon_abs>; def : Pat<(xor (v8i8 (AArch64vashr V64:$src, (i32 7))), (v8i8 (add V64:$src, (AArch64vashr V64:$src, (i32 7))))), @@ -3375,8 +3409,6 @@ defm SSUBW : SIMDWideThreeVectorBHS<0, 0b0011, "ssubw", BinOpFrag<(sub node:$LHS, (sext node:$RHS))>>; defm UABAL : SIMDLongThreeVectorTiedBHSabal<1, 0b0101, "uabal", uabsdiff>; -defm UABDL : SIMDLongThreeVectorBHSabdl<1, 0b0111, "uabdl", - uabsdiff>; defm UADDL : SIMDLongThreeVectorBHS<1, 0b0000, "uaddl", BinOpFrag<(add (zext node:$LHS), (zext node:$RHS))>>; defm UADDW : SIMDWideThreeVectorBHS<1, 0b0001, "uaddw", |
