Add Rbit_V instruction. Add 8 tests (Rbit_V; Rev16_V, Rev32_V, Rev64_V). Improve CountSetBits8() algorithm. (#212)

* Update AOpCodeTable.cs

* Update AInstEmitSimdArithmetic.cs

* Update AInstEmitSimdLogical.cs

* Update AVectorHelper.cs

* Update ASoftFallback.cs

* Update Instructions.cs

* Update CpuTestSimd.cs

* Update CpuTestSimdReg.cs

* Improve CountSetBits8() algorithm.

* Improve CountSetBits8() algorithm.

AOpCodeTable.cs

@@ -348,6 +348,7 @@ namespace ChocolArm64
             SetA64("0x001110101xxxxx000111xxxxxxxxxx", AInstEmit.Orr_V,         typeof(AOpCodeSimdReg));
             SetA64("0x00111100000xxx<<x101xxxxxxxxxx", AInstEmit.Orr_Vi,        typeof(AOpCodeSimdImm));
             SetA64("0x101110<<1xxxxx010000xxxxxxxxxx", AInstEmit.Raddhn_V,      typeof(AOpCodeSimdReg));
+            SetA64("0x10111001100000010110xxxxxxxxxx", AInstEmit.Rbit_V,        typeof(AOpCodeSimd));
             SetA64("0x00111000100000000110xxxxxxxxxx", AInstEmit.Rev16_V,       typeof(AOpCodeSimd));
             SetA64("0x1011100x100000000010xxxxxxxxxx", AInstEmit.Rev32_V,       typeof(AOpCodeSimd));
             SetA64("0x001110<<100000000010xxxxxxxxxx", AInstEmit.Rev64_V,       typeof(AOpCodeSimd));

Instruction/AInstEmitSimdArithmetic.cs

@@ -151,9 +151,9 @@ namespace ChocolArm64.Instruction
             {
                 EmitVectorExtractZx(Context, Op.Rn, Index, 0);
 
-                Context.Emit(OpCodes.Conv_U1);
+                Context.Emit(OpCodes.Conv_U4);
 
-                AVectorHelper.EmitCall(Context, nameof(AVectorHelper.CountSetBits8));
+                ASoftFallback.EmitCall(Context, nameof(ASoftFallback.CountSetBits8));
 
                 Context.Emit(OpCodes.Conv_U8);
 

Instruction/AInstEmitSimdLogical.cs

@@ -56,8 +56,9 @@ namespace ChocolArm64.Instruction
             AOpCodeSimdReg Op = (AOpCodeSimdReg)Context.CurrOp;
 
             int Bytes = Context.CurrOp.GetBitsCount() >> 3;
+            int Elems = Bytes >> Op.Size;
 
-            for (int Index = 0; Index < (Bytes >> Op.Size); Index++)
+            for (int Index = 0; Index < Elems; Index++)
             {
                 EmitVectorExtractZx(Context, Op.Rd, Index, Op.Size);
                 EmitVectorExtractZx(Context, Op.Rn, Index, Op.Size);
@@ -145,6 +146,31 @@ namespace ChocolArm64.Instruction
             EmitVectorImmBinaryOp(Context, () => Context.Emit(OpCodes.Or));
         }
 
+        public static void Rbit_V(AILEmitterCtx Context)
+        {
+            AOpCodeSimd Op = (AOpCodeSimd)Context.CurrOp;
+
+            int Elems = Op.RegisterSize == ARegisterSize.SIMD128 ? 16 : 8;
+
+            for (int Index = 0; Index < Elems; Index++)
+            {
+                EmitVectorExtractZx(Context, Op.Rn, Index, 0);
+
+                Context.Emit(OpCodes.Conv_U4);
+
+                ASoftFallback.EmitCall(Context, nameof(ASoftFallback.ReverseBits8));
+
+                Context.Emit(OpCodes.Conv_U8);
+
+                EmitVectorInsert(Context, Op.Rd, Index, 0);
+            }
+
+            if (Op.RegisterSize == ARegisterSize.SIMD64)
+            {
+                EmitVectorZeroUpper(Context, Op.Rd);
+            }
+        }
+
         public static void Rev16_V(AILEmitterCtx Context)
         {
             EmitRev_V(Context, ContainerSize: 1);
@@ -164,18 +190,17 @@ namespace ChocolArm64.Instruction
         {
             AOpCodeSimd Op = (AOpCodeSimd)Context.CurrOp;
 
-            int Bytes = Context.CurrOp.GetBitsCount() >> 3;
-
-            int Elems = Bytes >> Op.Size;
-
             if (Op.Size >= ContainerSize)
             {
                 throw new InvalidOperationException();
             }
 
+            int Bytes = Context.CurrOp.GetBitsCount() >> 3;
+            int Elems = Bytes >> Op.Size;
+
             int ContainerMask = (1 << (ContainerSize - Op.Size)) - 1;
 
-            for (int Index = 0; Index < (Bytes >> Op.Size); Index++)
+            for (int Index = 0; Index < Elems; Index++)
             {
                 int RevIndex = Index ^ ContainerMask;
 

Instruction/ASoftFallback.cs

@@ -30,6 +30,14 @@ namespace ChocolArm64.Instruction
             return (ulong)Size;
         }
 
+        public static uint CountSetBits8(uint Value)
+        {
+            Value = ((Value >> 1) & 0x55) + (Value & 0x55);
+            Value = ((Value >> 2) & 0x33) + (Value & 0x33);
+
+            return (Value >> 4) + (Value & 0x0f);
+        }
+
         private const uint Crc32RevPoly  = 0xedb88320;
         private const uint Crc32cRevPoly = 0x82f63b78;
 
@@ -89,6 +97,14 @@ namespace ChocolArm64.Instruction
             return Crc;
         }
 
+        public static uint ReverseBits8(uint Value)
+        {
+            Value = ((Value & 0xaa) >> 1) | ((Value & 0x55) << 1);
+            Value = ((Value & 0xcc) >> 2) | ((Value & 0x33) << 2);
+
+            return (Value >> 4) | ((Value & 0x0f) << 4);
+        }
+
         public static uint ReverseBits32(uint Value)
         {
             Value = ((Value & 0xaaaaaaaa) >> 1) | ((Value & 0x55555555) << 1);
@@ -101,10 +117,10 @@ namespace ChocolArm64.Instruction
 
         public static ulong ReverseBits64(ulong Value)
         {
-            Value = ((Value & 0xaaaaaaaaaaaaaaaa) >>  1) | ((Value & 0x5555555555555555) <<  1);
+            Value = ((Value & 0xaaaaaaaaaaaaaaaa) >> 1 ) | ((Value & 0x5555555555555555) << 1 );
-            Value = ((Value & 0xcccccccccccccccc) >>  2) | ((Value & 0x3333333333333333) <<  2);
+            Value = ((Value & 0xcccccccccccccccc) >> 2 ) | ((Value & 0x3333333333333333) << 2 );
-            Value = ((Value & 0xf0f0f0f0f0f0f0f0) >>  4) | ((Value & 0x0f0f0f0f0f0f0f0f) <<  4);
+            Value = ((Value & 0xf0f0f0f0f0f0f0f0) >> 4 ) | ((Value & 0x0f0f0f0f0f0f0f0f) << 4 );
-            Value = ((Value & 0xff00ff00ff00ff00) >>  8) | ((Value & 0x00ff00ff00ff00ff) <<  8);
+            Value = ((Value & 0xff00ff00ff00ff00) >> 8 ) | ((Value & 0x00ff00ff00ff00ff) << 8 );
             Value = ((Value & 0xffff0000ffff0000) >> 16) | ((Value & 0x0000ffff0000ffff) << 16);
 
             return (Value >> 32) | (Value << 32);

Instruction/AVectorHelper.cs

@@ -93,14 +93,6 @@ namespace ChocolArm64.Instruction
                    Value < ulong.MinValue ? ulong.MinValue : (ulong)Value;
         }
 
-        public static int CountSetBits8(byte Value)
-        {
-            return ((Value >> 0) & 1) + ((Value >> 1) & 1) +
-                   ((Value >> 2) & 1) + ((Value >> 3) & 1) +
-                   ((Value >> 4) & 1) + ((Value >> 5) & 1) +
-                   ((Value >> 6) & 1) +  (Value >> 7);
-        }
-
         public static double Max(double LHS, double RHS)
         {
             if (LHS == 0.0 && RHS == 0.0)
@@ -646,4 +638,4 @@ namespace ChocolArm64.Instruction
             throw new PlatformNotSupportedException();
         }
     }
-}
+}