Add Saba_V, Sabal_V, Sabd_V, Sabdl_V, Uaba_V, Uabal_V; Update Uabd_V, Uabdl_V. Add 16 tests. (#204)

* Update AOpCodeTable.cs

* Update AInstEmitSimdArithmetic.cs

* Update AInstEmitSimdHelper.cs

* Update Instructions.cs

* Update CpuTest.cs

* Update CpuTestSimd.cs

* Update CpuTestSimdReg.cs

ChocolArm64/AOpCodeTable.cs

@@ -352,6 +352,10 @@ namespace ChocolArm64
             SetA64("0x1011100x100000000010xxxxxxxxxx", AInstEmit.Rev32_V,       typeof(AOpCodeSimd));
             SetA64("0x001110<<100000000010xxxxxxxxxx", AInstEmit.Rev64_V,       typeof(AOpCodeSimd));
             SetA64("0x101110<<1xxxxx011000xxxxxxxxxx", AInstEmit.Rsubhn_V,      typeof(AOpCodeSimdReg));
+            SetA64("0x001110<<1xxxxx011111xxxxxxxxxx", AInstEmit.Saba_V,        typeof(AOpCodeSimdReg));
+            SetA64("0x001110<<1xxxxx010100xxxxxxxxxx", AInstEmit.Sabal_V,       typeof(AOpCodeSimdReg));
+            SetA64("0x001110<<1xxxxx011101xxxxxxxxxx", AInstEmit.Sabd_V,        typeof(AOpCodeSimdReg));
+            SetA64("0x001110<<1xxxxx011100xxxxxxxxxx", AInstEmit.Sabdl_V,       typeof(AOpCodeSimdReg));
             SetA64("0x001110<<1xxxxx000100xxxxxxxxxx", AInstEmit.Saddw_V,       typeof(AOpCodeSimdReg));
             SetA64("x0011110xx100010000000xxxxxxxxxx", AInstEmit.Scvtf_Gp,      typeof(AOpCodeSimdCvt));
             SetA64("010111100x100001110110xxxxxxxxxx", AInstEmit.Scvtf_S,       typeof(AOpCodeSimd));
@@ -390,6 +394,8 @@ namespace ChocolArm64
             SetA64("0x001110000xxxxx0xx000xxxxxxxxxx", AInstEmit.Tbl_V,         typeof(AOpCodeSimdTbl));
             SetA64("0>001110<<0xxxxx001010xxxxxxxxxx", AInstEmit.Trn1_V,        typeof(AOpCodeSimdReg));
             SetA64("0>001110<<0xxxxx011010xxxxxxxxxx", AInstEmit.Trn2_V,        typeof(AOpCodeSimdReg));
+            SetA64("0x101110<<1xxxxx011111xxxxxxxxxx", AInstEmit.Uaba_V,        typeof(AOpCodeSimdReg));
+            SetA64("0x101110<<1xxxxx010100xxxxxxxxxx", AInstEmit.Uabal_V,       typeof(AOpCodeSimdReg));
             SetA64("0x101110<<1xxxxx011101xxxxxxxxxx", AInstEmit.Uabd_V,        typeof(AOpCodeSimdReg));
             SetA64("0x101110<<1xxxxx011100xxxxxxxxxx", AInstEmit.Uabdl_V,       typeof(AOpCodeSimdReg));
             SetA64("0x101110<<1xxxxx000000xxxxxxxxxx", AInstEmit.Uaddl_V,       typeof(AOpCodeSimdReg));

ChocolArm64/Instruction/AInstEmitSimdArithmetic.cs

@@ -22,19 +22,6 @@ namespace ChocolArm64.Instruction
             EmitVectorUnaryOpSx(Context, () => EmitAbs(Context));
         }
 
-        private static void EmitAbs(AILEmitterCtx Context)
-        {
-            AILLabel LblTrue = new AILLabel();
-
-            Context.Emit(OpCodes.Dup);
-            Context.Emit(OpCodes.Ldc_I4_0);
-            Context.Emit(OpCodes.Bge_S, LblTrue);
-
-            Context.Emit(OpCodes.Neg);
-
-            Context.MarkLabel(LblTrue);
-        }
-
         public static void Add_S(AILEmitterCtx Context)
         {
             EmitScalarBinaryOpZx(Context, () => Context.Emit(OpCodes.Add));
@@ -179,6 +166,19 @@ namespace ChocolArm64.Instruction
             }
         }
 
+        private static void EmitAbs(AILEmitterCtx Context)
+        {
+            AILLabel LblTrue = new AILLabel();
+
+            Context.Emit(OpCodes.Dup);
+            Context.Emit(OpCodes.Ldc_I4_0);
+            Context.Emit(OpCodes.Bge_S, LblTrue);
+
+            Context.Emit(OpCodes.Neg);
+
+            Context.MarkLabel(LblTrue);
+        }
+
         private static void EmitHighNarrow(AILEmitterCtx Context, Action Emit, bool Round)
         {
             AOpCodeSimdReg Op = (AOpCodeSimdReg)Context.CurrOp;
@@ -188,6 +188,8 @@ namespace ChocolArm64.Instruction
 
             int Part = Op.RegisterSize == ARegisterSize.SIMD128 ? Elems : 0;
 
+            long RoundConst = 1L << (ESize - 1);
+
             for (int Index = 0; Index < Elems; Index++)
             {
                 EmitVectorExtractZx(Context, Op.Rn, Index, Op.Size + 1);
@@ -197,7 +199,7 @@ namespace ChocolArm64.Instruction
 
                 if (Round)
                 {
-                    Context.EmitLdc_I8(1L << (ESize - 1));
+                    Context.EmitLdc_I8(RoundConst);
 
                     Context.Emit(OpCodes.Add);
                 }
@@ -220,11 +222,11 @@ namespace ChocolArm64.Instruction
             int Elems = (!Scalar ? 8 >> Op.Size : 1);
             int ESize = 8 << Op.Size;
 
+            int Part = (!Scalar & (Op.RegisterSize == ARegisterSize.SIMD128) ? Elems : 0);
+
             int TMaxValue = (SignedDst ? (1 << (ESize - 1)) - 1 : (int)((1L << ESize) - 1L));
             int TMinValue = (SignedDst ? -((1 << (ESize - 1))) : 0);
 
-            int Part = (!Scalar & (Op.RegisterSize == ARegisterSize.SIMD128) ? Elems : 0);
-
             Context.EmitLdc_I8(0L);
             Context.EmitSttmp();
 
@@ -1107,6 +1109,46 @@ namespace ChocolArm64.Instruction
             EmitHighNarrow(Context, () => Context.Emit(OpCodes.Sub), Round: true);
         }
 
+        public static void Saba_V(AILEmitterCtx Context)
+        {
+            EmitVectorTernaryOpSx(Context, () =>
+            {
+                Context.Emit(OpCodes.Sub);
+                EmitAbs(Context);
+
+                Context.Emit(OpCodes.Add);
+            });
+        }
+
+        public static void Sabal_V(AILEmitterCtx Context)
+        {
+            EmitVectorWidenRnRmTernaryOpSx(Context, () =>
+            {
+                Context.Emit(OpCodes.Sub);
+                EmitAbs(Context);
+
+                Context.Emit(OpCodes.Add);
+            });
+        }
+
+        public static void Sabd_V(AILEmitterCtx Context)
+        {
+            EmitVectorBinaryOpSx(Context, () =>
+            {
+                Context.Emit(OpCodes.Sub);
+                EmitAbs(Context);
+            });
+        }
+
+        public static void Sabdl_V(AILEmitterCtx Context)
+        {
+            EmitVectorWidenRnRmBinaryOpSx(Context, () =>
+            {
+                Context.Emit(OpCodes.Sub);
+                EmitAbs(Context);
+            });
+        }
+
         public static void Saddw_V(AILEmitterCtx Context)
         {
             EmitVectorWidenRmBinaryOpSx(Context, () => Context.Emit(OpCodes.Add));
@@ -1186,23 +1228,44 @@ namespace ChocolArm64.Instruction
             EmitHighNarrow(Context, () => Context.Emit(OpCodes.Sub), Round: false);
         }
 
+        public static void Uaba_V(AILEmitterCtx Context)
+        {
+            EmitVectorTernaryOpZx(Context, () =>
+            {
+                Context.Emit(OpCodes.Sub);
+                EmitAbs(Context);
+
+                Context.Emit(OpCodes.Add);
+            });
+        }
+
+        public static void Uabal_V(AILEmitterCtx Context)
+        {
+            EmitVectorWidenRnRmTernaryOpZx(Context, () =>
+            {
+                Context.Emit(OpCodes.Sub);
+                EmitAbs(Context);
+
+                Context.Emit(OpCodes.Add);
+            });
+        }
+
         public static void Uabd_V(AILEmitterCtx Context)
         {
-            EmitVectorBinaryOpZx(Context, () => EmitAbd(Context));
+            EmitVectorBinaryOpZx(Context, () =>
+            {
+                Context.Emit(OpCodes.Sub);
+                EmitAbs(Context);
+            });
         }
 
         public static void Uabdl_V(AILEmitterCtx Context)
         {
-            EmitVectorWidenRnRmBinaryOpZx(Context, () => EmitAbd(Context));
+            EmitVectorWidenRnRmBinaryOpZx(Context, () =>
-        }
-
-        private static void EmitAbd(AILEmitterCtx Context)
             {
                 Context.Emit(OpCodes.Sub);
-
+                EmitAbs(Context);
-            Type[] Types = new Type[] { typeof(long) };
+            });
-
-            Context.EmitCall(typeof(Math).GetMethod(nameof(Math.Abs), Types));
         }
 
         public static void Uaddl_V(AILEmitterCtx Context)

ChocolArm64/Instruction/AInstEmitSimdHelper.cs

@@ -483,6 +483,11 @@ namespace ChocolArm64.Instruction
             EmitVectorOp(Context, Emit, OperFlags.RnRm, true);
         }
 
+        public static void EmitVectorTernaryOpSx(AILEmitterCtx Context, Action Emit)
+        {
+            EmitVectorOp(Context, Emit, OperFlags.RdRnRm, true);
+        }
+
         public static void EmitVectorUnaryOpZx(AILEmitterCtx Context, Action Emit)
         {
             EmitVectorOp(Context, Emit, OperFlags.Rn, false);

Ryujinx.Tests/Cpu/CpuTest.cs

@@ -113,20 +113,20 @@ namespace Ryujinx.Tests.Cpu
             return GetThreadState();
         }
 
-        protected static Vector128<float> MakeVectorE0(double A0)
+        protected static Vector128<float> MakeVectorE0(double E0)
         {
-            return Sse.StaticCast<long, float>(Sse2.SetVector128(0, BitConverter.DoubleToInt64Bits(A0)));
+            return Sse.StaticCast<long, float>(Sse2.SetVector128(0, BitConverter.DoubleToInt64Bits(E0)));
         }
 
-        protected static Vector128<float> MakeVectorE0E1(double A0, double A1)
+        protected static Vector128<float> MakeVectorE0E1(double E0, double E1)
         {
-            return Sse.StaticCast<long, float>(Sse2.SetVector128(BitConverter.DoubleToInt64Bits(A1),
+            return Sse.StaticCast<long, float>(Sse2.SetVector128(BitConverter.DoubleToInt64Bits(E1),
-                                                                 BitConverter.DoubleToInt64Bits(A0)));
+                                                                 BitConverter.DoubleToInt64Bits(E0)));
         }
 
-        protected static Vector128<float> MakeVectorE1(double A1)
+        protected static Vector128<float> MakeVectorE1(double E1)
         {
-            return Sse.StaticCast<long, float>(Sse2.SetVector128(BitConverter.DoubleToInt64Bits(A1), 0));
+            return Sse.StaticCast<long, float>(Sse2.SetVector128(BitConverter.DoubleToInt64Bits(E1), 0));
         }
 
         protected static double VectorExtractDouble(Vector128<float> Vector, byte Index)
@@ -136,29 +136,29 @@ namespace Ryujinx.Tests.Cpu
             return BitConverter.Int64BitsToDouble(Value);
         }
 
-        protected static Vector128<float> MakeVectorE0(ulong A0)
+        protected static Vector128<float> MakeVectorE0(ulong E0)
         {
-            return Sse.StaticCast<ulong, float>(Sse2.SetVector128(0, A0));
+            return Sse.StaticCast<ulong, float>(Sse2.SetVector128(0, E0));
         }
 
-        protected static Vector128<float> MakeVectorE0E1(ulong A0, ulong A1)
+        protected static Vector128<float> MakeVectorE0E1(ulong E0, ulong E1)
         {
-            return Sse.StaticCast<ulong, float>(Sse2.SetVector128(A1, A0));
+            return Sse.StaticCast<ulong, float>(Sse2.SetVector128(E1, E0));
         }
 
-        protected static Vector128<float> MakeVectorE1(ulong A1)
+        protected static Vector128<float> MakeVectorE1(ulong E1)
         {
-            return Sse.StaticCast<ulong, float>(Sse2.SetVector128(A1, 0));
+            return Sse.StaticCast<ulong, float>(Sse2.SetVector128(E1, 0));
         }
 
         protected static ulong GetVectorE0(Vector128<float> Vector)
         {
-            return Sse41.Extract(Sse.StaticCast<float, ulong>(Vector), 0);
+            return Sse41.Extract(Sse.StaticCast<float, ulong>(Vector), (byte)0);
         }
 
         protected static ulong GetVectorE1(Vector128<float> Vector)
         {
-            return Sse41.Extract(Sse.StaticCast<float, ulong>(Vector), 1);
+            return Sse41.Extract(Sse.StaticCast<float, ulong>(Vector), (byte)1);
         }
     }
 }

Ryujinx.Tests/Cpu/CpuTestSimd.cs

@@ -834,8 +834,8 @@ namespace Ryujinx.Tests.Cpu
             Opcode |= ((size & 3) << 22);
             Bits Op = new Bits(Opcode);
 
-            ulong _X0 = TestContext.CurrentContext.Random.NextULong();
+            ulong _E0 = TestContext.CurrentContext.Random.NextULong();
-            Vector128<float> V0 = MakeVectorE0(_X0);
+            Vector128<float> V0 = MakeVectorE0(_E0);
             Vector128<float> V1 = MakeVectorE0E1(A0, A1);
             AThreadState ThreadState = SingleOpcode(Opcode, V0: V0, V1: V1);
 
@@ -845,7 +845,7 @@ namespace Ryujinx.Tests.Cpu
 
             Assert.Multiple(() =>
             {
-                Assert.That(GetVectorE0(ThreadState.V0), Is.EqualTo(_X0));
+                Assert.That(GetVectorE0(ThreadState.V0), Is.EqualTo(_E0));
                 Assert.That(GetVectorE1(ThreadState.V0), Is.EqualTo(AArch64.Vpart(64, 0, 1).ToUInt64()));
             });
             Assert.That(((ThreadState.Fpsr >> 27) & 1) != 0, Is.EqualTo(Shared.FPSR[27]));
@@ -910,8 +910,8 @@ namespace Ryujinx.Tests.Cpu
             Opcode |= ((size & 3) << 22);
             Bits Op = new Bits(Opcode);
 
-            ulong _X0 = TestContext.CurrentContext.Random.NextULong();
+            ulong _E0 = TestContext.CurrentContext.Random.NextULong();
-            Vector128<float> V0 = MakeVectorE0(_X0);
+            Vector128<float> V0 = MakeVectorE0(_E0);
             Vector128<float> V1 = MakeVectorE0E1(A0, A1);
             AThreadState ThreadState = SingleOpcode(Opcode, V0: V0, V1: V1);
 
@@ -921,7 +921,7 @@ namespace Ryujinx.Tests.Cpu
 
             Assert.Multiple(() =>
             {
-                Assert.That(GetVectorE0(ThreadState.V0), Is.EqualTo(_X0));
+                Assert.That(GetVectorE0(ThreadState.V0), Is.EqualTo(_E0));
                 Assert.That(GetVectorE1(ThreadState.V0), Is.EqualTo(AArch64.Vpart(64, 0, 1).ToUInt64()));
             });
             Assert.That(((ThreadState.Fpsr >> 27) & 1) != 0, Is.EqualTo(Shared.FPSR[27]));
@@ -986,8 +986,8 @@ namespace Ryujinx.Tests.Cpu
             Opcode |= ((size & 3) << 22);
             Bits Op = new Bits(Opcode);
 
-            ulong _X0 = TestContext.CurrentContext.Random.NextULong();
+            ulong _E0 = TestContext.CurrentContext.Random.NextULong();
-            Vector128<float> V0 = MakeVectorE0(_X0);
+            Vector128<float> V0 = MakeVectorE0(_E0);
             Vector128<float> V1 = MakeVectorE0E1(A0, A1);
             AThreadState ThreadState = SingleOpcode(Opcode, V0: V0, V1: V1);
 
@@ -997,7 +997,7 @@ namespace Ryujinx.Tests.Cpu
 
             Assert.Multiple(() =>
             {
-                Assert.That(GetVectorE0(ThreadState.V0), Is.EqualTo(_X0));
+                Assert.That(GetVectorE0(ThreadState.V0), Is.EqualTo(_E0));
                 Assert.That(GetVectorE1(ThreadState.V0), Is.EqualTo(AArch64.Vpart(64, 0, 1).ToUInt64()));
             });
             Assert.That(((ThreadState.Fpsr >> 27) & 1) != 0, Is.EqualTo(Shared.FPSR[27]));

Ryujinx.Tests/Cpu/CpuTestSimdReg.cs

@@ -176,8 +176,8 @@ namespace Ryujinx.Tests.Cpu
             Opcode |= ((size & 3) << 22);
             Bits Op = new Bits(Opcode);
 
-            ulong _X0 = TestContext.CurrentContext.Random.NextULong();
+            ulong _E0 = TestContext.CurrentContext.Random.NextULong();
-            Vector128<float> V0 = MakeVectorE0(_X0);
+            Vector128<float> V0 = MakeVectorE0(_E0);
             Vector128<float> V1 = MakeVectorE0E1(A0, A1);
             Vector128<float> V2 = MakeVectorE0E1(B0, B1);
             AThreadState ThreadState = SingleOpcode(Opcode, V0: V0, V1: V1, V2: V2);
@@ -190,7 +190,7 @@ namespace Ryujinx.Tests.Cpu
 
             Assert.Multiple(() =>
             {
-                Assert.That(GetVectorE0(ThreadState.V0), Is.EqualTo(_X0));
+                Assert.That(GetVectorE0(ThreadState.V0), Is.EqualTo(_E0));
                 Assert.That(GetVectorE1(ThreadState.V0), Is.EqualTo(AArch64.Vpart(64, 0, 1).ToUInt64()));
             });
         }
@@ -1157,8 +1157,8 @@ namespace Ryujinx.Tests.Cpu
             Opcode |= ((size & 3) << 22);
             Bits Op = new Bits(Opcode);
 
-            ulong _X0 = TestContext.CurrentContext.Random.NextULong();
+            ulong _E0 = TestContext.CurrentContext.Random.NextULong();
-            Vector128<float> V0 = MakeVectorE0(_X0);
+            Vector128<float> V0 = MakeVectorE0(_E0);
             Vector128<float> V1 = MakeVectorE0E1(A0, A1);
             Vector128<float> V2 = MakeVectorE0E1(B0, B1);
             AThreadState ThreadState = SingleOpcode(Opcode, V0: V0, V1: V1, V2: V2);
@@ -1171,7 +1171,7 @@ namespace Ryujinx.Tests.Cpu
 
             Assert.Multiple(() =>
             {
-                Assert.That(GetVectorE0(ThreadState.V0), Is.EqualTo(_X0));
+                Assert.That(GetVectorE0(ThreadState.V0), Is.EqualTo(_E0));
                 Assert.That(GetVectorE1(ThreadState.V0), Is.EqualTo(AArch64.Vpart(64, 0, 1).ToUInt64()));
             });
         }
@@ -1216,8 +1216,8 @@ namespace Ryujinx.Tests.Cpu
             Opcode |= ((size & 3) << 22);
             Bits Op = new Bits(Opcode);
 
-            ulong _X0 = TestContext.CurrentContext.Random.NextULong();
+            ulong _E0 = TestContext.CurrentContext.Random.NextULong();
-            Vector128<float> V0 = MakeVectorE0(_X0);
+            Vector128<float> V0 = MakeVectorE0(_E0);
             Vector128<float> V1 = MakeVectorE0E1(A0, A1);
             Vector128<float> V2 = MakeVectorE0E1(B0, B1);
             AThreadState ThreadState = SingleOpcode(Opcode, V0: V0, V1: V1, V2: V2);
@@ -1230,7 +1230,233 @@ namespace Ryujinx.Tests.Cpu
 
             Assert.Multiple(() =>
             {
-                Assert.That(GetVectorE0(ThreadState.V0), Is.EqualTo(_X0));
+                Assert.That(GetVectorE0(ThreadState.V0), Is.EqualTo(_E0));
+                Assert.That(GetVectorE1(ThreadState.V0), Is.EqualTo(AArch64.Vpart(64, 0, 1).ToUInt64()));
+            });
+        }
+
+        [Test, Description("SABA <Vd>.<T>, <Vn>.<T>, <Vm>.<T>")]
+        public void Saba_V_8B_4H_2S([ValueSource("_8B4H2S_")] [Random(1)] ulong _Z,
+                                    [ValueSource("_8B4H2S_")] [Random(1)] ulong A,
+                                    [ValueSource("_8B4H2S_")] [Random(1)] ulong B,
+                                    [Values(0b00u, 0b01u, 0b10u)] uint size) // <8B, 4H, 2S>
+        {
+            uint Opcode = 0x0E227C20; // SABA V0.8B, V1.8B, V2.8B
+            Opcode |= ((size & 3) << 22);
+            Bits Op = new Bits(Opcode);
+
+            Vector128<float> V0 = MakeVectorE0E1(_Z, TestContext.CurrentContext.Random.NextULong());
+            Vector128<float> V1 = MakeVectorE0(A);
+            Vector128<float> V2 = MakeVectorE0(B);
+            AThreadState ThreadState = SingleOpcode(Opcode, V0: V0, V1: V1, V2: V2);
+
+            AArch64.Vpart(0, 0, new Bits(_Z));
+            AArch64.V(1, new Bits(A));
+            AArch64.V(2, new Bits(B));
+            SimdFp.Saba_V(Op[30], Op[23, 22], Op[20, 16], Op[9, 5], Op[4, 0]);
+
+            Assert.Multiple(() =>
+            {
+                Assert.That(GetVectorE0(ThreadState.V0), Is.EqualTo(AArch64.V(64, 0).ToUInt64()));
+                Assert.That(GetVectorE1(ThreadState.V0), Is.Zero);
+            });
+        }
+
+        [Test, Pairwise, Description("SABA <Vd>.<T>, <Vn>.<T>, <Vm>.<T>")]
+        public void Saba_V_16B_8H_4S([ValueSource("_8B4H2S_")] [Random(1)] ulong _Z0,
+                                     [ValueSource("_8B4H2S_")] [Random(1)] ulong _Z1,
+                                     [ValueSource("_8B4H2S_")] [Random(1)] ulong A0,
+                                     [ValueSource("_8B4H2S_")] [Random(1)] ulong A1,
+                                     [ValueSource("_8B4H2S_")] [Random(1)] ulong B0,
+                                     [ValueSource("_8B4H2S_")] [Random(1)] ulong B1,
+                                     [Values(0b00u, 0b01u, 0b10u)] uint size) // <16B, 8H, 4S>
+        {
+            uint Opcode = 0x4E227C20; // SABA V0.16B, V1.16B, V2.16B
+            Opcode |= ((size & 3) << 22);
+            Bits Op = new Bits(Opcode);
+
+            Vector128<float> V0 = MakeVectorE0E1(_Z0, _Z1);
+            Vector128<float> V1 = MakeVectorE0E1(A0, A1);
+            Vector128<float> V2 = MakeVectorE0E1(B0, B1);
+            AThreadState ThreadState = SingleOpcode(Opcode, V0: V0, V1: V1, V2: V2);
+
+            AArch64.Vpart(0, 0, new Bits(_Z0));
+            AArch64.Vpart(0, 1, new Bits(_Z1));
+            AArch64.Vpart(1, 0, new Bits(A0));
+            AArch64.Vpart(1, 1, new Bits(A1));
+            AArch64.Vpart(2, 0, new Bits(B0));
+            AArch64.Vpart(2, 1, new Bits(B1));
+            SimdFp.Saba_V(Op[30], Op[23, 22], Op[20, 16], Op[9, 5], Op[4, 0]);
+
+            Assert.Multiple(() =>
+            {
+                Assert.That(GetVectorE0(ThreadState.V0), Is.EqualTo(AArch64.Vpart(64, 0, 0).ToUInt64()));
+                Assert.That(GetVectorE1(ThreadState.V0), Is.EqualTo(AArch64.Vpart(64, 0, 1).ToUInt64()));
+            });
+        }
+
+        [Test, Pairwise, Description("SABAL{2} <Vd>.<Ta>, <Vn>.<Tb>, <Vm>.<Tb>")]
+        public void Sabal_V_8B8H_4H4S_2S2D([ValueSource("_8B4H2S_")] [Random(1)] ulong _Z0,
+                                           [ValueSource("_8B4H2S_")] [Random(1)] ulong _Z1,
+                                           [ValueSource("_8B4H2S_")] [Random(1)] ulong A0,
+                                           [ValueSource("_8B4H2S_")] [Random(1)] ulong B0,
+                                           [Values(0b00u, 0b01u, 0b10u)] uint size) // <8B, 4H, 2S>
+        {
+            uint Opcode = 0x0E225020; // SABAL V0.8H, V1.8B, V2.8B
+            Opcode |= ((size & 3) << 22);
+            Bits Op = new Bits(Opcode);
+
+            Vector128<float> V0 = MakeVectorE0E1(_Z0, _Z1);
+            Vector128<float> V1 = MakeVectorE0(A0);
+            Vector128<float> V2 = MakeVectorE0(B0);
+            AThreadState ThreadState = SingleOpcode(Opcode, V0: V0, V1: V1, V2: V2);
+
+            AArch64.Vpart(0, 0, new Bits(_Z0));
+            AArch64.Vpart(0, 1, new Bits(_Z1));
+            AArch64.Vpart(1, 0, new Bits(A0));
+            AArch64.Vpart(2, 0, new Bits(B0));
+            SimdFp.Sabal_V(Op[30], Op[23, 22], Op[20, 16], Op[9, 5], Op[4, 0]);
+
+            Assert.Multiple(() =>
+            {
+                Assert.That(GetVectorE0(ThreadState.V0), Is.EqualTo(AArch64.Vpart(64, 0, 0).ToUInt64()));
+                Assert.That(GetVectorE1(ThreadState.V0), Is.EqualTo(AArch64.Vpart(64, 0, 1).ToUInt64()));
+            });
+        }
+
+        [Test, Pairwise, Description("SABAL{2} <Vd>.<Ta>, <Vn>.<Tb>, <Vm>.<Tb>")]
+        public void Sabal_V_16B8H_8H4S_4S2D([ValueSource("_8B4H2S_")] [Random(1)] ulong _Z0,
+                                            [ValueSource("_8B4H2S_")] [Random(1)] ulong _Z1,
+                                            [ValueSource("_8B4H2S_")] [Random(1)] ulong A1,
+                                            [ValueSource("_8B4H2S_")] [Random(1)] ulong B1,
+                                            [Values(0b00u, 0b01u, 0b10u)] uint size) // <16B, 8H, 4S>
+        {
+            uint Opcode = 0x4E225020; // SABAL2 V0.8H, V1.16B, V2.16B
+            Opcode |= ((size & 3) << 22);
+            Bits Op = new Bits(Opcode);
+
+            Vector128<float> V0 = MakeVectorE0E1(_Z0, _Z1);
+            Vector128<float> V1 = MakeVectorE1(A1);
+            Vector128<float> V2 = MakeVectorE1(B1);
+            AThreadState ThreadState = SingleOpcode(Opcode, V0: V0, V1: V1, V2: V2);
+
+            AArch64.Vpart(0, 0, new Bits(_Z0));
+            AArch64.Vpart(0, 1, new Bits(_Z1));
+            AArch64.Vpart(1, 1, new Bits(A1));
+            AArch64.Vpart(2, 1, new Bits(B1));
+            SimdFp.Sabal_V(Op[30], Op[23, 22], Op[20, 16], Op[9, 5], Op[4, 0]);
+
+            Assert.Multiple(() =>
+            {
+                Assert.That(GetVectorE0(ThreadState.V0), Is.EqualTo(AArch64.Vpart(64, 0, 0).ToUInt64()));
+                Assert.That(GetVectorE1(ThreadState.V0), Is.EqualTo(AArch64.Vpart(64, 0, 1).ToUInt64()));
+            });
+        }
+
+        [Test, Description("SABD <Vd>.<T>, <Vn>.<T>, <Vm>.<T>")]
+        public void Sabd_V_8B_4H_2S([ValueSource("_8B4H2S_")] [Random(1)] ulong A,
+                                    [ValueSource("_8B4H2S_")] [Random(1)] ulong B,
+                                    [Values(0b00u, 0b01u, 0b10u)] uint size) // <8B, 4H, 2S>
+        {
+            uint Opcode = 0x0E227420; // SABD V0.8B, V1.8B, V2.8B
+            Opcode |= ((size & 3) << 22);
+            Bits Op = new Bits(Opcode);
+
+            Vector128<float> V0 = MakeVectorE0E1(TestContext.CurrentContext.Random.NextULong(),
+                                                 TestContext.CurrentContext.Random.NextULong());
+            Vector128<float> V1 = MakeVectorE0(A);
+            Vector128<float> V2 = MakeVectorE0(B);
+            AThreadState ThreadState = SingleOpcode(Opcode, V0: V0, V1: V1, V2: V2);
+
+            AArch64.V(1, new Bits(A));
+            AArch64.V(2, new Bits(B));
+            SimdFp.Sabd_V(Op[30], Op[23, 22], Op[20, 16], Op[9, 5], Op[4, 0]);
+
+            Assert.Multiple(() =>
+            {
+                Assert.That(GetVectorE0(ThreadState.V0), Is.EqualTo(AArch64.V(64, 0).ToUInt64()));
+                Assert.That(GetVectorE1(ThreadState.V0), Is.Zero);
+            });
+        }
+
+        [Test, Pairwise, Description("SABD <Vd>.<T>, <Vn>.<T>, <Vm>.<T>")]
+        public void Sabd_V_16B_8H_4S([ValueSource("_8B4H2S_")] [Random(1)] ulong A0,
+                                     [ValueSource("_8B4H2S_")] [Random(1)] ulong A1,
+                                     [ValueSource("_8B4H2S_")] [Random(1)] ulong B0,
+                                     [ValueSource("_8B4H2S_")] [Random(1)] ulong B1,
+                                     [Values(0b00u, 0b01u, 0b10u)] uint size) // <16B, 8H, 4S>
+        {
+            uint Opcode = 0x4E227420; // SABD V0.16B, V1.16B, V2.16B
+            Opcode |= ((size & 3) << 22);
+            Bits Op = new Bits(Opcode);
+
+            Vector128<float> V0 = MakeVectorE0E1(TestContext.CurrentContext.Random.NextULong(),
+                                                 TestContext.CurrentContext.Random.NextULong());
+            Vector128<float> V1 = MakeVectorE0E1(A0, A1);
+            Vector128<float> V2 = MakeVectorE0E1(B0, B1);
+            AThreadState ThreadState = SingleOpcode(Opcode, V0: V0, V1: V1, V2: V2);
+
+            AArch64.Vpart(1, 0, new Bits(A0));
+            AArch64.Vpart(1, 1, new Bits(A1));
+            AArch64.Vpart(2, 0, new Bits(B0));
+            AArch64.Vpart(2, 1, new Bits(B1));
+            SimdFp.Sabd_V(Op[30], Op[23, 22], Op[20, 16], Op[9, 5], Op[4, 0]);
+
+            Assert.Multiple(() =>
+            {
+                Assert.That(GetVectorE0(ThreadState.V0), Is.EqualTo(AArch64.Vpart(64, 0, 0).ToUInt64()));
+                Assert.That(GetVectorE1(ThreadState.V0), Is.EqualTo(AArch64.Vpart(64, 0, 1).ToUInt64()));
+            });
+        }
+
+        [Test, Description("SABDL{2} <Vd>.<Ta>, <Vn>.<Tb>, <Vm>.<Tb>")]
+        public void Sabdl_V_8B8H_4H4S_2S2D([ValueSource("_8B4H2S_")] [Random(1)] ulong A0,
+                                           [ValueSource("_8B4H2S_")] [Random(1)] ulong B0,
+                                           [Values(0b00u, 0b01u, 0b10u)] uint size) // <8B, 4H, 2S>
+        {
+            uint Opcode = 0x0E227020; // SABDL V0.8H, V1.8B, V2.8B
+            Opcode |= ((size & 3) << 22);
+            Bits Op = new Bits(Opcode);
+
+            Vector128<float> V0 = MakeVectorE0E1(TestContext.CurrentContext.Random.NextULong(),
+                                                 TestContext.CurrentContext.Random.NextULong());
+            Vector128<float> V1 = MakeVectorE0(A0);
+            Vector128<float> V2 = MakeVectorE0(B0);
+            AThreadState ThreadState = SingleOpcode(Opcode, V0: V0, V1: V1, V2: V2);
+
+            AArch64.Vpart(1, 0, new Bits(A0));
+            AArch64.Vpart(2, 0, new Bits(B0));
+            SimdFp.Sabdl_V(Op[30], Op[23, 22], Op[20, 16], Op[9, 5], Op[4, 0]);
+
+            Assert.Multiple(() =>
+            {
+                Assert.That(GetVectorE0(ThreadState.V0), Is.EqualTo(AArch64.Vpart(64, 0, 0).ToUInt64()));
+                Assert.That(GetVectorE1(ThreadState.V0), Is.EqualTo(AArch64.Vpart(64, 0, 1).ToUInt64()));
+            });
+        }
+
+        [Test, Description("SABDL{2} <Vd>.<Ta>, <Vn>.<Tb>, <Vm>.<Tb>")]
+        public void Sabdl_V_16B8H_8H4S_4S2D([ValueSource("_8B4H2S_")] [Random(1)] ulong A1,
+                                            [ValueSource("_8B4H2S_")] [Random(1)] ulong B1,
+                                            [Values(0b00u, 0b01u, 0b10u)] uint size) // <16B, 8H, 4S>
+        {
+            uint Opcode = 0x4E227020; // SABDL2 V0.8H, V1.16B, V2.16B
+            Opcode |= ((size & 3) << 22);
+            Bits Op = new Bits(Opcode);
+
+            Vector128<float> V0 = MakeVectorE0E1(TestContext.CurrentContext.Random.NextULong(),
+                                                 TestContext.CurrentContext.Random.NextULong());
+            Vector128<float> V1 = MakeVectorE1(A1);
+            Vector128<float> V2 = MakeVectorE1(B1);
+            AThreadState ThreadState = SingleOpcode(Opcode, V0: V0, V1: V1, V2: V2);
+
+            AArch64.Vpart(1, 1, new Bits(A1));
+            AArch64.Vpart(2, 1, new Bits(B1));
+            SimdFp.Sabdl_V(Op[30], Op[23, 22], Op[20, 16], Op[9, 5], Op[4, 0]);
+
+            Assert.Multiple(() =>
+            {
+                Assert.That(GetVectorE0(ThreadState.V0), Is.EqualTo(AArch64.Vpart(64, 0, 0).ToUInt64()));
                 Assert.That(GetVectorE1(ThreadState.V0), Is.EqualTo(AArch64.Vpart(64, 0, 1).ToUInt64()));
             });
         }
@@ -1351,8 +1577,8 @@ namespace Ryujinx.Tests.Cpu
             Opcode |= ((size & 3) << 22);
             Bits Op = new Bits(Opcode);
 
-            ulong _X0 = TestContext.CurrentContext.Random.NextULong();
+            ulong _E0 = TestContext.CurrentContext.Random.NextULong();
-            Vector128<float> V0 = MakeVectorE0(_X0);
+            Vector128<float> V0 = MakeVectorE0(_E0);
             Vector128<float> V1 = MakeVectorE0E1(A0, A1);
             Vector128<float> V2 = MakeVectorE0E1(B0, B1);
             AThreadState ThreadState = SingleOpcode(Opcode, V0: V0, V1: V1, V2: V2);
@@ -1365,7 +1591,233 @@ namespace Ryujinx.Tests.Cpu
 
             Assert.Multiple(() =>
             {
-                Assert.That(GetVectorE0(ThreadState.V0), Is.EqualTo(_X0));
+                Assert.That(GetVectorE0(ThreadState.V0), Is.EqualTo(_E0));
+                Assert.That(GetVectorE1(ThreadState.V0), Is.EqualTo(AArch64.Vpart(64, 0, 1).ToUInt64()));
+            });
+        }
+
+        [Test, Description("UABA <Vd>.<T>, <Vn>.<T>, <Vm>.<T>")]
+        public void Uaba_V_8B_4H_2S([ValueSource("_8B4H2S_")] [Random(1)] ulong _Z,
+                                    [ValueSource("_8B4H2S_")] [Random(1)] ulong A,
+                                    [ValueSource("_8B4H2S_")] [Random(1)] ulong B,
+                                    [Values(0b00u, 0b01u, 0b10u)] uint size) // <8B, 4H, 2S>
+        {
+            uint Opcode = 0x2E227C20; // UABA V0.8B, V1.8B, V2.8B
+            Opcode |= ((size & 3) << 22);
+            Bits Op = new Bits(Opcode);
+
+            Vector128<float> V0 = MakeVectorE0E1(_Z, TestContext.CurrentContext.Random.NextULong());
+            Vector128<float> V1 = MakeVectorE0(A);
+            Vector128<float> V2 = MakeVectorE0(B);
+            AThreadState ThreadState = SingleOpcode(Opcode, V0: V0, V1: V1, V2: V2);
+
+            AArch64.Vpart(0, 0, new Bits(_Z));
+            AArch64.V(1, new Bits(A));
+            AArch64.V(2, new Bits(B));
+            SimdFp.Uaba_V(Op[30], Op[23, 22], Op[20, 16], Op[9, 5], Op[4, 0]);
+
+            Assert.Multiple(() =>
+            {
+                Assert.That(GetVectorE0(ThreadState.V0), Is.EqualTo(AArch64.V(64, 0).ToUInt64()));
+                Assert.That(GetVectorE1(ThreadState.V0), Is.Zero);
+            });
+        }
+
+        [Test, Pairwise, Description("UABA <Vd>.<T>, <Vn>.<T>, <Vm>.<T>")]
+        public void Uaba_V_16B_8H_4S([ValueSource("_8B4H2S_")] [Random(1)] ulong _Z0,
+                                     [ValueSource("_8B4H2S_")] [Random(1)] ulong _Z1,
+                                     [ValueSource("_8B4H2S_")] [Random(1)] ulong A0,
+                                     [ValueSource("_8B4H2S_")] [Random(1)] ulong A1,
+                                     [ValueSource("_8B4H2S_")] [Random(1)] ulong B0,
+                                     [ValueSource("_8B4H2S_")] [Random(1)] ulong B1,
+                                     [Values(0b00u, 0b01u, 0b10u)] uint size) // <16B, 8H, 4S>
+        {
+            uint Opcode = 0x6E227C20; // UABA V0.16B, V1.16B, V2.16B
+            Opcode |= ((size & 3) << 22);
+            Bits Op = new Bits(Opcode);
+
+            Vector128<float> V0 = MakeVectorE0E1(_Z0, _Z1);
+            Vector128<float> V1 = MakeVectorE0E1(A0, A1);
+            Vector128<float> V2 = MakeVectorE0E1(B0, B1);
+            AThreadState ThreadState = SingleOpcode(Opcode, V0: V0, V1: V1, V2: V2);
+
+            AArch64.Vpart(0, 0, new Bits(_Z0));
+            AArch64.Vpart(0, 1, new Bits(_Z1));
+            AArch64.Vpart(1, 0, new Bits(A0));
+            AArch64.Vpart(1, 1, new Bits(A1));
+            AArch64.Vpart(2, 0, new Bits(B0));
+            AArch64.Vpart(2, 1, new Bits(B1));
+            SimdFp.Uaba_V(Op[30], Op[23, 22], Op[20, 16], Op[9, 5], Op[4, 0]);
+
+            Assert.Multiple(() =>
+            {
+                Assert.That(GetVectorE0(ThreadState.V0), Is.EqualTo(AArch64.Vpart(64, 0, 0).ToUInt64()));
+                Assert.That(GetVectorE1(ThreadState.V0), Is.EqualTo(AArch64.Vpart(64, 0, 1).ToUInt64()));
+            });
+        }
+
+        [Test, Pairwise, Description("UABAL{2} <Vd>.<Ta>, <Vn>.<Tb>, <Vm>.<Tb>")]
+        public void Uabal_V_8B8H_4H4S_2S2D([ValueSource("_8B4H2S_")] [Random(1)] ulong _Z0,
+                                           [ValueSource("_8B4H2S_")] [Random(1)] ulong _Z1,
+                                           [ValueSource("_8B4H2S_")] [Random(1)] ulong A0,
+                                           [ValueSource("_8B4H2S_")] [Random(1)] ulong B0,
+                                           [Values(0b00u, 0b01u, 0b10u)] uint size) // <8B, 4H, 2S>
+        {
+            uint Opcode = 0x2E225020; // UABAL V0.8H, V1.8B, V2.8B
+            Opcode |= ((size & 3) << 22);
+            Bits Op = new Bits(Opcode);
+
+            Vector128<float> V0 = MakeVectorE0E1(_Z0, _Z1);
+            Vector128<float> V1 = MakeVectorE0(A0);
+            Vector128<float> V2 = MakeVectorE0(B0);
+            AThreadState ThreadState = SingleOpcode(Opcode, V0: V0, V1: V1, V2: V2);
+
+            AArch64.Vpart(0, 0, new Bits(_Z0));
+            AArch64.Vpart(0, 1, new Bits(_Z1));
+            AArch64.Vpart(1, 0, new Bits(A0));
+            AArch64.Vpart(2, 0, new Bits(B0));
+            SimdFp.Uabal_V(Op[30], Op[23, 22], Op[20, 16], Op[9, 5], Op[4, 0]);
+
+            Assert.Multiple(() =>
+            {
+                Assert.That(GetVectorE0(ThreadState.V0), Is.EqualTo(AArch64.Vpart(64, 0, 0).ToUInt64()));
+                Assert.That(GetVectorE1(ThreadState.V0), Is.EqualTo(AArch64.Vpart(64, 0, 1).ToUInt64()));
+            });
+        }
+
+        [Test, Pairwise, Description("UABAL{2} <Vd>.<Ta>, <Vn>.<Tb>, <Vm>.<Tb>")]
+        public void Uabal_V_16B8H_8H4S_4S2D([ValueSource("_8B4H2S_")] [Random(1)] ulong _Z0,
+                                            [ValueSource("_8B4H2S_")] [Random(1)] ulong _Z1,
+                                            [ValueSource("_8B4H2S_")] [Random(1)] ulong A1,
+                                            [ValueSource("_8B4H2S_")] [Random(1)] ulong B1,
+                                            [Values(0b00u, 0b01u, 0b10u)] uint size) // <16B, 8H, 4S>
+        {
+            uint Opcode = 0x6E225020; // UABAL2 V0.8H, V1.16B, V2.16B
+            Opcode |= ((size & 3) << 22);
+            Bits Op = new Bits(Opcode);
+
+            Vector128<float> V0 = MakeVectorE0E1(_Z0, _Z1);
+            Vector128<float> V1 = MakeVectorE1(A1);
+            Vector128<float> V2 = MakeVectorE1(B1);
+            AThreadState ThreadState = SingleOpcode(Opcode, V0: V0, V1: V1, V2: V2);
+
+            AArch64.Vpart(0, 0, new Bits(_Z0));
+            AArch64.Vpart(0, 1, new Bits(_Z1));
+            AArch64.Vpart(1, 1, new Bits(A1));
+            AArch64.Vpart(2, 1, new Bits(B1));
+            SimdFp.Uabal_V(Op[30], Op[23, 22], Op[20, 16], Op[9, 5], Op[4, 0]);
+
+            Assert.Multiple(() =>
+            {
+                Assert.That(GetVectorE0(ThreadState.V0), Is.EqualTo(AArch64.Vpart(64, 0, 0).ToUInt64()));
+                Assert.That(GetVectorE1(ThreadState.V0), Is.EqualTo(AArch64.Vpart(64, 0, 1).ToUInt64()));
+            });
+        }
+
+        [Test, Description("UABD <Vd>.<T>, <Vn>.<T>, <Vm>.<T>")]
+        public void Uabd_V_8B_4H_2S([ValueSource("_8B4H2S_")] [Random(1)] ulong A,
+                                    [ValueSource("_8B4H2S_")] [Random(1)] ulong B,
+                                    [Values(0b00u, 0b01u, 0b10u)] uint size) // <8B, 4H, 2S>
+        {
+            uint Opcode = 0x2E227420; // UABD V0.8B, V1.8B, V2.8B
+            Opcode |= ((size & 3) << 22);
+            Bits Op = new Bits(Opcode);
+
+            Vector128<float> V0 = MakeVectorE0E1(TestContext.CurrentContext.Random.NextULong(),
+                                                 TestContext.CurrentContext.Random.NextULong());
+            Vector128<float> V1 = MakeVectorE0(A);
+            Vector128<float> V2 = MakeVectorE0(B);
+            AThreadState ThreadState = SingleOpcode(Opcode, V0: V0, V1: V1, V2: V2);
+
+            AArch64.V(1, new Bits(A));
+            AArch64.V(2, new Bits(B));
+            SimdFp.Uabd_V(Op[30], Op[23, 22], Op[20, 16], Op[9, 5], Op[4, 0]);
+
+            Assert.Multiple(() =>
+            {
+                Assert.That(GetVectorE0(ThreadState.V0), Is.EqualTo(AArch64.V(64, 0).ToUInt64()));
+                Assert.That(GetVectorE1(ThreadState.V0), Is.Zero);
+            });
+        }
+
+        [Test, Pairwise, Description("UABD <Vd>.<T>, <Vn>.<T>, <Vm>.<T>")]
+        public void Uabd_V_16B_8H_4S([ValueSource("_8B4H2S_")] [Random(1)] ulong A0,
+                                     [ValueSource("_8B4H2S_")] [Random(1)] ulong A1,
+                                     [ValueSource("_8B4H2S_")] [Random(1)] ulong B0,
+                                     [ValueSource("_8B4H2S_")] [Random(1)] ulong B1,
+                                     [Values(0b00u, 0b01u, 0b10u)] uint size) // <16B, 8H, 4S>
+        {
+            uint Opcode = 0x6E227420; // UABD V0.16B, V1.16B, V2.16B
+            Opcode |= ((size & 3) << 22);
+            Bits Op = new Bits(Opcode);
+
+            Vector128<float> V0 = MakeVectorE0E1(TestContext.CurrentContext.Random.NextULong(),
+                                                 TestContext.CurrentContext.Random.NextULong());
+            Vector128<float> V1 = MakeVectorE0E1(A0, A1);
+            Vector128<float> V2 = MakeVectorE0E1(B0, B1);
+            AThreadState ThreadState = SingleOpcode(Opcode, V0: V0, V1: V1, V2: V2);
+
+            AArch64.Vpart(1, 0, new Bits(A0));
+            AArch64.Vpart(1, 1, new Bits(A1));
+            AArch64.Vpart(2, 0, new Bits(B0));
+            AArch64.Vpart(2, 1, new Bits(B1));
+            SimdFp.Uabd_V(Op[30], Op[23, 22], Op[20, 16], Op[9, 5], Op[4, 0]);
+
+            Assert.Multiple(() =>
+            {
+                Assert.That(GetVectorE0(ThreadState.V0), Is.EqualTo(AArch64.Vpart(64, 0, 0).ToUInt64()));
+                Assert.That(GetVectorE1(ThreadState.V0), Is.EqualTo(AArch64.Vpart(64, 0, 1).ToUInt64()));
+            });
+        }
+
+        [Test, Description("UABDL{2} <Vd>.<Ta>, <Vn>.<Tb>, <Vm>.<Tb>")]
+        public void Uabdl_V_8B8H_4H4S_2S2D([ValueSource("_8B4H2S_")] [Random(1)] ulong A0,
+                                           [ValueSource("_8B4H2S_")] [Random(1)] ulong B0,
+                                           [Values(0b00u, 0b01u, 0b10u)] uint size) // <8B, 4H, 2S>
+        {
+            uint Opcode = 0x2E227020; // UABDL V0.8H, V1.8B, V2.8B
+            Opcode |= ((size & 3) << 22);
+            Bits Op = new Bits(Opcode);
+
+            Vector128<float> V0 = MakeVectorE0E1(TestContext.CurrentContext.Random.NextULong(),
+                                                 TestContext.CurrentContext.Random.NextULong());
+            Vector128<float> V1 = MakeVectorE0(A0);
+            Vector128<float> V2 = MakeVectorE0(B0);
+            AThreadState ThreadState = SingleOpcode(Opcode, V0: V0, V1: V1, V2: V2);
+
+            AArch64.Vpart(1, 0, new Bits(A0));
+            AArch64.Vpart(2, 0, new Bits(B0));
+            SimdFp.Uabdl_V(Op[30], Op[23, 22], Op[20, 16], Op[9, 5], Op[4, 0]);
+
+            Assert.Multiple(() =>
+            {
+                Assert.That(GetVectorE0(ThreadState.V0), Is.EqualTo(AArch64.Vpart(64, 0, 0).ToUInt64()));
+                Assert.That(GetVectorE1(ThreadState.V0), Is.EqualTo(AArch64.Vpart(64, 0, 1).ToUInt64()));
+            });
+        }
+
+        [Test, Description("UABDL{2} <Vd>.<Ta>, <Vn>.<Tb>, <Vm>.<Tb>")]
+        public void Uabdl_V_16B8H_8H4S_4S2D([ValueSource("_8B4H2S_")] [Random(1)] ulong A1,
+                                            [ValueSource("_8B4H2S_")] [Random(1)] ulong B1,
+                                            [Values(0b00u, 0b01u, 0b10u)] uint size) // <16B, 8H, 4S>
+        {
+            uint Opcode = 0x6E227020; // UABDL2 V0.8H, V1.16B, V2.16B
+            Opcode |= ((size & 3) << 22);
+            Bits Op = new Bits(Opcode);
+
+            Vector128<float> V0 = MakeVectorE0E1(TestContext.CurrentContext.Random.NextULong(),
+                                                 TestContext.CurrentContext.Random.NextULong());
+            Vector128<float> V1 = MakeVectorE1(A1);
+            Vector128<float> V2 = MakeVectorE1(B1);
+            AThreadState ThreadState = SingleOpcode(Opcode, V0: V0, V1: V1, V2: V2);
+
+            AArch64.Vpart(1, 1, new Bits(A1));
+            AArch64.Vpart(2, 1, new Bits(B1));
+            SimdFp.Uabdl_V(Op[30], Op[23, 22], Op[20, 16], Op[9, 5], Op[4, 0]);
+
+            Assert.Multiple(() =>
+            {
+                Assert.That(GetVectorE0(ThreadState.V0), Is.EqualTo(AArch64.Vpart(64, 0, 0).ToUInt64()));
                 Assert.That(GetVectorE1(ThreadState.V0), Is.EqualTo(AArch64.Vpart(64, 0, 1).ToUInt64()));
             });
         }

Ryujinx.Tests/Cpu/Tester/Instructions.cs

@@ -4074,6 +4074,184 @@ namespace Ryujinx.Tests.Cpu.Tester
             Vpart(d, part, result);
         }
 
+        // saba_advsimd.html
+        public static void Saba_V(bool Q, Bits size, Bits Rm, Bits Rn, Bits Rd)
+        {
+            const bool U = false;
+            const bool ac = true;
+
+            /* Decode */
+            int d = (int)UInt(Rd);
+            int n = (int)UInt(Rn);
+            int m = (int)UInt(Rm);
+
+            /* if size == '11' then ReservedValue(); */
+
+            int esize = 8 << (int)UInt(size);
+            int datasize = (Q ? 128 : 64);
+            int elements = datasize / esize;
+
+            bool unsigned = (U == true);
+            bool accumulate = (ac == true);
+
+            /* Operation */
+            /* CheckFPAdvSIMDEnabled64(); */
+
+            Bits operand1 = V(datasize, n);
+            Bits operand2 = V(datasize, m);
+            BigInteger element1;
+            BigInteger element2;
+            Bits absdiff;
+
+            Bits result = (accumulate ? V(datasize, d) : Zeros(datasize));
+
+            for (int e = 0; e <= elements - 1; e++)
+            {
+                element1 = Int(Elem(operand1, e, esize), unsigned);
+                element2 = Int(Elem(operand2, e, esize), unsigned);
+
+                absdiff = Abs(element1 - element2).SubBigInteger(esize - 1, 0);
+
+                Elem(result, e, esize, Elem(result, e, esize) + absdiff);
+            }
+
+            V(d, result);
+        }
+
+        // sabal_advsimd.html
+        public static void Sabal_V(bool Q, Bits size, Bits Rm, Bits Rn, Bits Rd)
+        {
+            const bool U = false;
+            const bool op = false;
+
+            /* Decode */
+            int d = (int)UInt(Rd);
+            int n = (int)UInt(Rn);
+            int m = (int)UInt(Rm);
+
+            /* if size == '11' then ReservedValue(); */
+
+            int esize = 8 << (int)UInt(size);
+            int datasize = 64;
+            int part = (int)UInt(Q);
+            int elements = datasize / esize;
+
+            bool unsigned = (U == true);
+            bool accumulate = (op == false);
+
+            /* Operation */
+            /* CheckFPAdvSIMDEnabled64(); */
+
+            Bits operand1 = Vpart(datasize, n, part);
+            Bits operand2 = Vpart(datasize, m, part);
+            BigInteger element1;
+            BigInteger element2;
+            Bits absdiff;
+
+            Bits result = (accumulate ? V(2 * datasize, d) : Zeros(2 * datasize));
+
+            for (int e = 0; e <= elements - 1; e++)
+            {
+                element1 = Int(Elem(operand1, e, esize), unsigned);
+                element2 = Int(Elem(operand2, e, esize), unsigned);
+
+                absdiff = Abs(element1 - element2).SubBigInteger(2 * esize - 1, 0);
+
+                Elem(result, e, 2 * esize, Elem(result, e, 2 * esize) + absdiff);
+            }
+
+            V(d, result);
+        }
+
+        // sabd_advsimd.html
+        public static void Sabd_V(bool Q, Bits size, Bits Rm, Bits Rn, Bits Rd)
+        {
+            const bool U = false;
+            const bool ac = false;
+
+            /* Decode */
+            int d = (int)UInt(Rd);
+            int n = (int)UInt(Rn);
+            int m = (int)UInt(Rm);
+
+            /* if size == '11' then ReservedValue(); */
+
+            int esize = 8 << (int)UInt(size);
+            int datasize = (Q ? 128 : 64);
+            int elements = datasize / esize;
+
+            bool unsigned = (U == true);
+            bool accumulate = (ac == true);
+
+            /* Operation */
+            /* CheckFPAdvSIMDEnabled64(); */
+
+            Bits operand1 = V(datasize, n);
+            Bits operand2 = V(datasize, m);
+            BigInteger element1;
+            BigInteger element2;
+            Bits absdiff;
+
+            Bits result = (accumulate ? V(datasize, d) : Zeros(datasize));
+
+            for (int e = 0; e <= elements - 1; e++)
+            {
+                element1 = Int(Elem(operand1, e, esize), unsigned);
+                element2 = Int(Elem(operand2, e, esize), unsigned);
+
+                absdiff = Abs(element1 - element2).SubBigInteger(esize - 1, 0);
+
+                Elem(result, e, esize, Elem(result, e, esize) + absdiff);
+            }
+
+            V(d, result);
+        }
+
+        // sabdl_advsimd.html
+        public static void Sabdl_V(bool Q, Bits size, Bits Rm, Bits Rn, Bits Rd)
+        {
+            const bool U = false;
+            const bool op = true;
+
+            /* Decode */
+            int d = (int)UInt(Rd);
+            int n = (int)UInt(Rn);
+            int m = (int)UInt(Rm);
+
+            /* if size == '11' then ReservedValue(); */
+
+            int esize = 8 << (int)UInt(size);
+            int datasize = 64;
+            int part = (int)UInt(Q);
+            int elements = datasize / esize;
+
+            bool unsigned = (U == true);
+            bool accumulate = (op == false);
+
+            /* Operation */
+            /* CheckFPAdvSIMDEnabled64(); */
+
+            Bits operand1 = Vpart(datasize, n, part);
+            Bits operand2 = Vpart(datasize, m, part);
+            BigInteger element1;
+            BigInteger element2;
+            Bits absdiff;
+
+            Bits result = (accumulate ? V(2 * datasize, d) : Zeros(2 * datasize));
+
+            for (int e = 0; e <= elements - 1; e++)
+            {
+                element1 = Int(Elem(operand1, e, esize), unsigned);
+                element2 = Int(Elem(operand2, e, esize), unsigned);
+
+                absdiff = Abs(element1 - element2).SubBigInteger(2 * esize - 1, 0);
+
+                Elem(result, e, 2 * esize, Elem(result, e, 2 * esize) + absdiff);
+            }
+
+            V(d, result);
+        }
+
         // sub_advsimd.html#SUB_asisdsame_only
         public static void Sub_S(Bits size, Bits Rm, Bits Rn, Bits Rd)
         {
@@ -4217,6 +4395,184 @@ namespace Ryujinx.Tests.Cpu.Tester
 
             Vpart(d, part, result);
         }
+
+        // uaba_advsimd.html
+        public static void Uaba_V(bool Q, Bits size, Bits Rm, Bits Rn, Bits Rd)
+        {
+            const bool U = true;
+            const bool ac = true;
+
+            /* Decode */
+            int d = (int)UInt(Rd);
+            int n = (int)UInt(Rn);
+            int m = (int)UInt(Rm);
+
+            /* if size == '11' then ReservedValue(); */
+
+            int esize = 8 << (int)UInt(size);
+            int datasize = (Q ? 128 : 64);
+            int elements = datasize / esize;
+
+            bool unsigned = (U == true);
+            bool accumulate = (ac == true);
+
+            /* Operation */
+            /* CheckFPAdvSIMDEnabled64(); */
+
+            Bits operand1 = V(datasize, n);
+            Bits operand2 = V(datasize, m);
+            BigInteger element1;
+            BigInteger element2;
+            Bits absdiff;
+
+            Bits result = (accumulate ? V(datasize, d) : Zeros(datasize));
+
+            for (int e = 0; e <= elements - 1; e++)
+            {
+                element1 = Int(Elem(operand1, e, esize), unsigned);
+                element2 = Int(Elem(operand2, e, esize), unsigned);
+
+                absdiff = Abs(element1 - element2).SubBigInteger(esize - 1, 0);
+
+                Elem(result, e, esize, Elem(result, e, esize) + absdiff);
+            }
+
+            V(d, result);
+        }
+
+        // uabal_advsimd.html
+        public static void Uabal_V(bool Q, Bits size, Bits Rm, Bits Rn, Bits Rd)
+        {
+            const bool U = true;
+            const bool op = false;
+
+            /* Decode */
+            int d = (int)UInt(Rd);
+            int n = (int)UInt(Rn);
+            int m = (int)UInt(Rm);
+
+            /* if size == '11' then ReservedValue(); */
+
+            int esize = 8 << (int)UInt(size);
+            int datasize = 64;
+            int part = (int)UInt(Q);
+            int elements = datasize / esize;
+
+            bool unsigned = (U == true);
+            bool accumulate = (op == false);
+
+            /* Operation */
+            /* CheckFPAdvSIMDEnabled64(); */
+
+            Bits operand1 = Vpart(datasize, n, part);
+            Bits operand2 = Vpart(datasize, m, part);
+            BigInteger element1;
+            BigInteger element2;
+            Bits absdiff;
+
+            Bits result = (accumulate ? V(2 * datasize, d) : Zeros(2 * datasize));
+
+            for (int e = 0; e <= elements - 1; e++)
+            {
+                element1 = Int(Elem(operand1, e, esize), unsigned);
+                element2 = Int(Elem(operand2, e, esize), unsigned);
+
+                absdiff = Abs(element1 - element2).SubBigInteger(2 * esize - 1, 0);
+
+                Elem(result, e, 2 * esize, Elem(result, e, 2 * esize) + absdiff);
+            }
+
+            V(d, result);
+        }
+
+        // uabd_advsimd.html
+        public static void Uabd_V(bool Q, Bits size, Bits Rm, Bits Rn, Bits Rd)
+        {
+            const bool U = true;
+            const bool ac = false;
+
+            /* Decode */
+            int d = (int)UInt(Rd);
+            int n = (int)UInt(Rn);
+            int m = (int)UInt(Rm);
+
+            /* if size == '11' then ReservedValue(); */
+
+            int esize = 8 << (int)UInt(size);
+            int datasize = (Q ? 128 : 64);
+            int elements = datasize / esize;
+
+            bool unsigned = (U == true);
+            bool accumulate = (ac == true);
+
+            /* Operation */
+            /* CheckFPAdvSIMDEnabled64(); */
+
+            Bits operand1 = V(datasize, n);
+            Bits operand2 = V(datasize, m);
+            BigInteger element1;
+            BigInteger element2;
+            Bits absdiff;
+
+            Bits result = (accumulate ? V(datasize, d) : Zeros(datasize));
+
+            for (int e = 0; e <= elements - 1; e++)
+            {
+                element1 = Int(Elem(operand1, e, esize), unsigned);
+                element2 = Int(Elem(operand2, e, esize), unsigned);
+
+                absdiff = Abs(element1 - element2).SubBigInteger(esize - 1, 0);
+
+                Elem(result, e, esize, Elem(result, e, esize) + absdiff);
+            }
+
+            V(d, result);
+        }
+
+        // uabdl_advsimd.html
+        public static void Uabdl_V(bool Q, Bits size, Bits Rm, Bits Rn, Bits Rd)
+        {
+            const bool U = true;
+            const bool op = true;
+
+            /* Decode */
+            int d = (int)UInt(Rd);
+            int n = (int)UInt(Rn);
+            int m = (int)UInt(Rm);
+
+            /* if size == '11' then ReservedValue(); */
+
+            int esize = 8 << (int)UInt(size);
+            int datasize = 64;
+            int part = (int)UInt(Q);
+            int elements = datasize / esize;
+
+            bool unsigned = (U == true);
+            bool accumulate = (op == false);
+
+            /* Operation */
+            /* CheckFPAdvSIMDEnabled64(); */
+
+            Bits operand1 = Vpart(datasize, n, part);
+            Bits operand2 = Vpart(datasize, m, part);
+            BigInteger element1;
+            BigInteger element2;
+            Bits absdiff;
+
+            Bits result = (accumulate ? V(2 * datasize, d) : Zeros(2 * datasize));
+
+            for (int e = 0; e <= elements - 1; e++)
+            {
+                element1 = Int(Elem(operand1, e, esize), unsigned);
+                element2 = Int(Elem(operand2, e, esize), unsigned);
+
+                absdiff = Abs(element1 - element2).SubBigInteger(2 * esize - 1, 0);
+
+                Elem(result, e, 2 * esize, Elem(result, e, 2 * esize) + absdiff);
+            }
+
+            V(d, result);
+        }
 #endregion
     }
 }