From 1e93466c3a5556db0bd87755e10e2938c2a43c1f Mon Sep 17 00:00:00 2001 From: Owen Anderson Date: Thu, 21 Oct 2010 18:55:04 +0000 Subject: [PATCH] Add correct NEON encodings for vhadd and vrhadd. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117047 91177308-0d34-0410-b5e6-96231b3b80d8 --- lib/Target/ARM/ARMInstrNEON.td | 39 ++++- test/MC/ARM/neon-fp-encoding.ll | 248 ++++++++++++++++++++++++++++++++ 2 files changed, 281 insertions(+), 6 deletions(-) diff --git a/lib/Target/ARM/ARMInstrNEON.td b/lib/Target/ARM/ARMInstrNEON.td index 45d10e3a505..e9fb12575aa 100644 --- a/lib/Target/ARM/ARMInstrNEON.td +++ b/lib/Target/ARM/ARMInstrNEON.td @@ -1289,10 +1289,22 @@ class N3VDInt op21_20, bits<4> op11_8, bit op4, Format f, InstrItinClass itin, string OpcodeStr, string Dt, ValueType ResTy, ValueType OpTy, Intrinsic IntOp, bit Commutable> : N3V { + (outs DPR:$Dd), (ins DPR:$Dn, DPR:$Dm), f, itin, + OpcodeStr, Dt, "$Dd, $Dn, $Dm", "", + [(set DPR:$Dd, (ResTy (IntOp (OpTy DPR:$Dn), (OpTy DPR:$Dm))))]> { let isCommutable = Commutable; + + // Instruction operands. + bits<5> Dd; + bits<5> Dn; + bits<5> Dm; + + let Inst{15-12} = Dd{3-0}; + let Inst{22} = Dd{4}; + let Inst{19-16} = Dn{3-0}; + let Inst{7} = Dn{4}; + let Inst{3-0} = Dm{3-0}; + let Inst{5} = Dm{4}; } class N3VDIntSL op21_20, bits<4> op11_8, InstrItinClass itin, string OpcodeStr, string Dt, ValueType Ty, Intrinsic IntOp> @@ -1320,10 +1332,25 @@ class N3VQInt op21_20, bits<4> op11_8, bit op4, Format f, InstrItinClass itin, string OpcodeStr, string Dt, ValueType ResTy, ValueType OpTy, Intrinsic IntOp, bit Commutable> : N3V { + (outs QPR:$Qd), (ins QPR:$Qn, QPR:$Qm), f, itin, + OpcodeStr, Dt, "$Qd, $Qn, $Qm", "", + [(set QPR:$Qd, (ResTy (IntOp (OpTy QPR:$Qn), (OpTy QPR:$Qm))))]> { let isCommutable = Commutable; + + // Instruction operands. + bits<4> Qd; + bits<4> Qn; + bits<4> Qm; + + let Inst{15-13} = Qd{2-0}; + let Inst{22} = Qd{3}; + let Inst{12} = 0; + let Inst{19-17} = Qn{2-0}; + let Inst{7} = Qn{3}; + let Inst{16} = 0; + let Inst{3-1} = Qm{2-0}; + let Inst{5} = Qm{3}; + let Inst{0} = 0; } class N3VQIntSL op21_20, bits<4> op11_8, InstrItinClass itin, string OpcodeStr, string Dt, diff --git a/test/MC/ARM/neon-fp-encoding.ll b/test/MC/ARM/neon-fp-encoding.ll index 9cd5bd77008..dbcb859b4cb 100644 --- a/test/MC/ARM/neon-fp-encoding.ll +++ b/test/MC/ARM/neon-fp-encoding.ll @@ -179,3 +179,251 @@ define <2 x i64> @vaddwu_2xi32(<2 x i64>* %A, <2 x i32>* %B) nounwind { %tmp4 = add <2 x i64> %tmp1, %tmp3 ret <2 x i64> %tmp4 } + +declare <8 x i8> @llvm.arm.neon.vhadds.v8i8(<8 x i8>, <8 x i8>) nounwind readnone +declare <4 x i16> @llvm.arm.neon.vhadds.v4i16(<4 x i16>, <4 x i16>) nounwind readnone +declare <2 x i32> @llvm.arm.neon.vhadds.v2i32(<2 x i32>, <2 x i32>) nounwind readnone + +; CHECK: vhadds_8xi8 +define <8 x i8> @vhadds_8xi8(<8 x i8>* %A, <8 x i8>* %B) nounwind { + %tmp1 = load <8 x i8>* %A + %tmp2 = load <8 x i8>* %B +; CHECK: vhadd.s8 d16, d16, d17 @ encoding: [0xa1,0x00,0x40,0xf2] + %tmp3 = call <8 x i8> @llvm.arm.neon.vhadds.v8i8(<8 x i8> %tmp1, <8 x i8> %tmp2) + ret <8 x i8> %tmp3 +} + +; CHECK: vhadds_4xi16 +define <4 x i16> @vhadds_4xi16(<4 x i16>* %A, <4 x i16>* %B) nounwind { + %tmp1 = load <4 x i16>* %A + %tmp2 = load <4 x i16>* %B +; CHECK: vhadd.s16 d16, d16, d17 @ encoding: [0xa1,0x00,0x50,0xf2] + %tmp3 = call <4 x i16> @llvm.arm.neon.vhadds.v4i16(<4 x i16> %tmp1, <4 x i16> %tmp2) + ret <4 x i16> %tmp3 +} + +; CHECK: vhadds_2xi32 +define <2 x i32> @vhadds_2xi32(<2 x i32>* %A, <2 x i32>* %B) nounwind { + %tmp1 = load <2 x i32>* %A + %tmp2 = load <2 x i32>* %B +; CHECK: vhadd.s32 d16, d16, d17 @ encoding: [0xa1,0x00,0x60,0xf2] + %tmp3 = call <2 x i32> @llvm.arm.neon.vhadds.v2i32(<2 x i32> %tmp1, <2 x i32> %tmp2) + ret <2 x i32> %tmp3 +} + +declare <8 x i8> @llvm.arm.neon.vhaddu.v8i8(<8 x i8>, <8 x i8>) nounwind readnone +declare <4 x i16> @llvm.arm.neon.vhaddu.v4i16(<4 x i16>, <4 x i16>) nounwind readnone +declare <2 x i32> @llvm.arm.neon.vhaddu.v2i32(<2 x i32>, <2 x i32>) nounwind readnone + +; CHECK: vhaddu_8xi8 +define <8 x i8> @vhaddu_8xi8(<8 x i8>* %A, <8 x i8>* %B) nounwind { + %tmp1 = load <8 x i8>* %A + %tmp2 = load <8 x i8>* %B +; CHECK: vhadd.u8 d16, d16, d17 @ encoding: [0xa1,0x00,0x40,0xf3] + %tmp3 = call <8 x i8> @llvm.arm.neon.vhaddu.v8i8(<8 x i8> %tmp1, <8 x i8> %tmp2) + ret <8 x i8> %tmp3 +} + +; CHECK: vhaddu_4xi16 +define <4 x i16> @vhaddu_4xi16(<4 x i16>* %A, <4 x i16>* %B) nounwind { + %tmp1 = load <4 x i16>* %A + %tmp2 = load <4 x i16>* %B +; CHECK: vhadd.u16 d16, d16, d17 @ encoding: [0xa1,0x00,0x50,0xf3] + %tmp3 = call <4 x i16> @llvm.arm.neon.vhaddu.v4i16(<4 x i16> %tmp1, <4 x i16> %tmp2) + ret <4 x i16> %tmp3 +} + +; CHECK: vhaddu_2xi32 +define <2 x i32> @vhaddu_2xi32(<2 x i32>* %A, <2 x i32>* %B) nounwind { + %tmp1 = load <2 x i32>* %A + %tmp2 = load <2 x i32>* %B +; CHECK: vhadd.u32 d16, d16, d17 @ encoding: [0xa1,0x00,0x60,0xf3] + %tmp3 = call <2 x i32> @llvm.arm.neon.vhaddu.v2i32(<2 x i32> %tmp1, <2 x i32> %tmp2) + ret <2 x i32> %tmp3 +} + +declare <16 x i8> @llvm.arm.neon.vhadds.v16i8(<16 x i8>, <16 x i8>) nounwind readnone +declare <8 x i16> @llvm.arm.neon.vhadds.v8i16(<8 x i16>, <8 x i16>) nounwind readnone +declare <4 x i32> @llvm.arm.neon.vhadds.v4i32(<4 x i32>, <4 x i32>) nounwind readnone + +; CHECK: vhadds_16xi8 +define <16 x i8> @vhadds_16xi8(<16 x i8>* %A, <16 x i8>* %B) nounwind { + %tmp1 = load <16 x i8>* %A + %tmp2 = load <16 x i8>* %B +; CHECK: vhadd.s8 q8, q8, q9 @ encoding: [0xe2,0x00,0x40,0xf2] + %tmp3 = call <16 x i8> @llvm.arm.neon.vhadds.v16i8(<16 x i8> %tmp1, <16 x i8> %tmp2) + ret <16 x i8> %tmp3 +} + +; CHECK: vhadds_8xi16 +define <8 x i16> @vhadds_8xi16(<8 x i16>* %A, <8 x i16>* %B) nounwind { + %tmp1 = load <8 x i16>* %A + %tmp2 = load <8 x i16>* %B +; CHECK: vhadd.s16 q8, q8, q9 @ encoding: [0xe2,0x00,0x50,0xf2] + %tmp3 = call <8 x i16> @llvm.arm.neon.vhadds.v8i16(<8 x i16> %tmp1, <8 x i16> %tmp2) + ret <8 x i16> %tmp3 +} + +; CHECK: vhadds_4xi32 +define <4 x i32> @vhadds_4xi32(<4 x i32>* %A, <4 x i32>* %B) nounwind { + %tmp1 = load <4 x i32>* %A + %tmp2 = load <4 x i32>* %B +; CHECK: vhadd.s32 q8, q8, q9 @ encoding: [0xe2,0x00,0x60,0xf2] + %tmp3 = call <4 x i32> @llvm.arm.neon.vhadds.v4i32(<4 x i32> %tmp1, <4 x i32> %tmp2) + ret <4 x i32> %tmp3 +} + +declare <16 x i8> @llvm.arm.neon.vhaddu.v16i8(<16 x i8>, <16 x i8>) nounwind readnone +declare <8 x i16> @llvm.arm.neon.vhaddu.v8i16(<8 x i16>, <8 x i16>) nounwind readnone +declare <4 x i32> @llvm.arm.neon.vhaddu.v4i32(<4 x i32>, <4 x i32>) nounwind readnone + +; CHECK: vhaddu_16xi8 +define <16 x i8> @vhaddu_16xi8(<16 x i8>* %A, <16 x i8>* %B) nounwind { + %tmp1 = load <16 x i8>* %A + %tmp2 = load <16 x i8>* %B +; CHECK: vhadd.u8 q8, q8, q9 @ encoding: [0xe2,0x00,0x40,0xf3] + %tmp3 = call <16 x i8> @llvm.arm.neon.vhaddu.v16i8(<16 x i8> %tmp1, <16 x i8> %tmp2) + ret <16 x i8> %tmp3 +} + +; CHECK: vhaddu_8xi16 +define <8 x i16> @vhaddu_8xi16(<8 x i16>* %A, <8 x i16>* %B) nounwind { + %tmp1 = load <8 x i16>* %A + %tmp2 = load <8 x i16>* %B +; CHECK: vhadd.u16 q8, q8, q9 @ encoding: [0xe2,0x00,0x50,0xf3] + %tmp3 = call <8 x i16> @llvm.arm.neon.vhaddu.v8i16(<8 x i16> %tmp1, <8 x i16> %tmp2) + ret <8 x i16> %tmp3 +} + +; CHECK: vhaddu_4xi32 +define <4 x i32> @vhaddu_4xi32(<4 x i32>* %A, <4 x i32>* %B) nounwind { + %tmp1 = load <4 x i32>* %A + %tmp2 = load <4 x i32>* %B +; CHECK: vhadd.u32 q8, q8, q9 @ encoding: [0xe2,0x00,0x60,0xf3] + %tmp3 = call <4 x i32> @llvm.arm.neon.vhaddu.v4i32(<4 x i32> %tmp1, <4 x i32> %tmp2) + ret <4 x i32> %tmp3 +} + +declare <8 x i8> @llvm.arm.neon.vrhadds.v8i8(<8 x i8>, <8 x i8>) nounwind readnone +declare <4 x i16> @llvm.arm.neon.vrhadds.v4i16(<4 x i16>, <4 x i16>) nounwind readnone +declare <2 x i32> @llvm.arm.neon.vrhadds.v2i32(<2 x i32>, <2 x i32>) nounwind readnone + +; CHECK: vrhadds_8xi8 +define <8 x i8> @vrhadds_8xi8(<8 x i8>* %A, <8 x i8>* %B) nounwind { + %tmp1 = load <8 x i8>* %A + %tmp2 = load <8 x i8>* %B +; CHECK: vrhadd.s8 d16, d16, d17 @ encoding: [0xa1,0x01,0x40,0xf2] + %tmp3 = call <8 x i8> @llvm.arm.neon.vrhadds.v8i8(<8 x i8> %tmp1, <8 x i8> %tmp2) + ret <8 x i8> %tmp3 +} + +; CHECK: vrhadds_4xi16 +define <4 x i16> @vrhadds_4xi16(<4 x i16>* %A, <4 x i16>* %B) nounwind { + %tmp1 = load <4 x i16>* %A + %tmp2 = load <4 x i16>* %B +; CHECK: vrhadd.s16 d16, d16, d17 @ encoding: [0xa1,0x01,0x50,0xf2] + %tmp3 = call <4 x i16> @llvm.arm.neon.vrhadds.v4i16(<4 x i16> %tmp1, <4 x i16> %tmp2) + ret <4 x i16> %tmp3 +} + +; CHECK: vrhadds_2xi32 +define <2 x i32> @vrhadds_2xi32(<2 x i32>* %A, <2 x i32>* %B) nounwind { + %tmp1 = load <2 x i32>* %A + %tmp2 = load <2 x i32>* %B +; CHECK: vrhadd.s32 d16, d16, d17 @ encoding: [0xa1,0x01,0x60,0xf2] + %tmp3 = call <2 x i32> @llvm.arm.neon.vrhadds.v2i32(<2 x i32> %tmp1, <2 x i32> %tmp2) + ret <2 x i32> %tmp3 +} + +declare <8 x i8> @llvm.arm.neon.vrhaddu.v8i8(<8 x i8>, <8 x i8>) nounwind readnone +declare <4 x i16> @llvm.arm.neon.vrhaddu.v4i16(<4 x i16>, <4 x i16>) nounwind readnone +declare <2 x i32> @llvm.arm.neon.vrhaddu.v2i32(<2 x i32>, <2 x i32>) nounwind readnone + +; CHECK: vrhaddu_8xi8 +define <8 x i8> @vrhaddu_8xi8(<8 x i8>* %A, <8 x i8>* %B) nounwind { + %tmp1 = load <8 x i8>* %A + %tmp2 = load <8 x i8>* %B +; CHECK: vrhadd.u8 d16, d16, d17 @ encoding: [0xa1,0x01,0x40,0xf3] + %tmp3 = call <8 x i8> @llvm.arm.neon.vrhaddu.v8i8(<8 x i8> %tmp1, <8 x i8> %tmp2) + ret <8 x i8> %tmp3 +} + +; CHECK: vrhaddu_4xi16 +define <4 x i16> @vrhaddu_4xi16(<4 x i16>* %A, <4 x i16>* %B) nounwind { + %tmp1 = load <4 x i16>* %A + %tmp2 = load <4 x i16>* %B +; CHECK: vrhadd.u16 d16, d16, d17 @ encoding: [0xa1,0x01,0x50,0xf3] + %tmp3 = call <4 x i16> @llvm.arm.neon.vrhaddu.v4i16(<4 x i16> %tmp1, <4 x i16> %tmp2) + ret <4 x i16> %tmp3 +} + +; CHECK: vrhaddu_2xi32 +define <2 x i32> @vrhaddu_2xi32(<2 x i32>* %A, <2 x i32>* %B) nounwind { + %tmp1 = load <2 x i32>* %A + %tmp2 = load <2 x i32>* %B +; CHECK: vrhadd.u32 d16, d16, d17 @ encoding: [0xa1,0x01,0x60,0xf3] + %tmp3 = call <2 x i32> @llvm.arm.neon.vrhaddu.v2i32(<2 x i32> %tmp1, <2 x i32> %tmp2) + ret <2 x i32> %tmp3 +} + +declare <16 x i8> @llvm.arm.neon.vrhadds.v16i8(<16 x i8>, <16 x i8>) nounwind readnone +declare <8 x i16> @llvm.arm.neon.vrhadds.v8i16(<8 x i16>, <8 x i16>) nounwind readnone +declare <4 x i32> @llvm.arm.neon.vrhadds.v4i32(<4 x i32>, <4 x i32>) nounwind readnone + +; CHECK: vrhadds_16xi8 +define <16 x i8> @vrhadds_16xi8(<16 x i8>* %A, <16 x i8>* %B) nounwind { + %tmp1 = load <16 x i8>* %A + %tmp2 = load <16 x i8>* %B +; CHECK: vrhadd.s8 q8, q8, q9 @ encoding: [0xe2,0x01,0x40,0xf2] + %tmp3 = call <16 x i8> @llvm.arm.neon.vrhadds.v16i8(<16 x i8> %tmp1, <16 x i8> %tmp2) + ret <16 x i8> %tmp3 +} + +; CHECK: vrhadds_8xi16 +define <8 x i16> @vrhadds_8xi16(<8 x i16>* %A, <8 x i16>* %B) nounwind { + %tmp1 = load <8 x i16>* %A + %tmp2 = load <8 x i16>* %B +; CHECK: vrhadd.s16 q8, q8, q9 @ encoding: [0xe2,0x01,0x50,0xf2] + %tmp3 = call <8 x i16> @llvm.arm.neon.vrhadds.v8i16(<8 x i16> %tmp1, <8 x i16> %tmp2) + ret <8 x i16> %tmp3 +} + +; CHECK: vrhadds_4xi32 +define <4 x i32> @vrhadds_4xi32(<4 x i32>* %A, <4 x i32>* %B) nounwind { + %tmp1 = load <4 x i32>* %A + %tmp2 = load <4 x i32>* %B +; CHECK: vrhadd.s32 q8, q8, q9 @ encoding: [0xe2,0x01,0x60,0xf2] + %tmp3 = call <4 x i32> @llvm.arm.neon.vrhadds.v4i32(<4 x i32> %tmp1, <4 x i32> %tmp2) + ret <4 x i32> %tmp3 +} + +declare <16 x i8> @llvm.arm.neon.vrhaddu.v16i8(<16 x i8>, <16 x i8>) nounwind readnone +declare <8 x i16> @llvm.arm.neon.vrhaddu.v8i16(<8 x i16>, <8 x i16>) nounwind readnone +declare <4 x i32> @llvm.arm.neon.vrhaddu.v4i32(<4 x i32>, <4 x i32>) nounwind readnone + +; CHECK: vrhaddu_16xi8 +define <16 x i8> @vrhaddu_16xi8(<16 x i8>* %A, <16 x i8>* %B) nounwind { + %tmp1 = load <16 x i8>* %A + %tmp2 = load <16 x i8>* %B +; CHECK: vrhadd.u8 q8, q8, q9 @ encoding: [0xe2,0x01,0x40,0xf3] + %tmp3 = call <16 x i8> @llvm.arm.neon.vrhaddu.v16i8(<16 x i8> %tmp1, <16 x i8> %tmp2) + ret <16 x i8> %tmp3 +} + +; CHECK: vrhaddu_8xi16 +define <8 x i16> @vrhaddu_8xi16(<8 x i16>* %A, <8 x i16>* %B) nounwind { + %tmp1 = load <8 x i16>* %A + %tmp2 = load <8 x i16>* %B +; CHECK: vrhadd.u16 q8, q8, q9 @ encoding: [0xe2,0x01,0x50,0xf3] + %tmp3 = call <8 x i16> @llvm.arm.neon.vrhaddu.v8i16(<8 x i16> %tmp1, <8 x i16> %tmp2) + ret <8 x i16> %tmp3 +} + +; CHECK: vrhaddu_4xi32 +define <4 x i32> @vrhaddu_4xi32(<4 x i32>* %A, <4 x i32>* %B) nounwind { + %tmp1 = load <4 x i32>* %A + %tmp2 = load <4 x i32>* %B +; CHECK: vrhadd.u32 q8, q8, q9 @ encoding: [0xe2,0x01,0x60,0xf3] + %tmp3 = call <4 x i32> @llvm.arm.neon.vrhaddu.v4i32(<4 x i32> %tmp1, <4 x i32> %tmp2) + ret <4 x i32> %tmp3 +} -- 2.34.1