[wasm] PackedSimd, implement store methods

[radekdoulik]

No description provided.

[dotnet-issue-labeler]

Note regarding the new-api-needs-documentation label:

This serves as a reminder for when your PR is modifying a ref *.cs file and adding/modifying public APIs, please make sure the API implementation in the src *.cs file is documented with triple slash comments, so the PR reviewers can sign off that change.

[ghost]

Tagging subscribers to this area: @dotnet/area-system-runtime-intrinsics
See info in area-owners.md if you want to be subscribed.

Issue Details

---

null

Author:	radekdoulik
Assignees:	-
Labels:	area-System.Runtime.Intrinsics, new-api-needs-documentation
Milestone:	-

[radekdoulik]

Test methods:

    unsafe static void TestStoreLane()
    {
        var mem = new byte[32];
        var v = RandomByteVector;
        var v2 = RandomByteVector;
        System.Console.WriteLine($"store  v: {v}");
        System.Console.WriteLine($"store v2: {v2}");
        System.Console.Write($"mem:     ");
        for (int i = 0; i < 16; i++) {
            fixed(byte* ptr = &mem[i]) {
                AddAndStoreLaneByte(v, v2, ptr);
                System.Console.Write($"  {mem[i]}");
            }
        }
        System.Console.WriteLine();
        var memUInt = new uint[32];
        var vuint = RandomUIntVector;
        var vuint2 = RandomUIntVector;
        System.Console.WriteLine($"store  v: {vuint}");
        System.Console.WriteLine($"store v2: {vuint2}");
        System.Console.Write($"mem:     ");
        for (int i = 0; i < 4; i++) {
            fixed(uint* ptr = &memUInt[i]) {
                AddAndStoreLaneUInt(vuint, vuint2, ptr);
                System.Console.Write($"  {memUInt[i]}");
            }
        }
        System.Console.WriteLine();
        var memFloat = new float[32];
        var vFloat = RandomFloatVector;
        var vFloat2 = RandomFloatVector;
        System.Console.WriteLine($"store  v: {vFloat}");
        System.Console.WriteLine($"store v2: {vFloat2}");
        System.Console.Write($"mem:     ");
        for (int i = 0; i < 2; i++) {
            fixed(float* ptr = &memFloat[i]) {
                AddAndStoreLaneFloat(vFloat, vFloat2, ptr);
                System.Console.Write($"  {memFloat[i]}");
            }
        }
        System.Console.WriteLine();
        var memdouble = new double[32];
        var vdouble = RandomDoubleVector;
        var vdouble2 = RandomDoubleVector;
        System.Console.WriteLine($"store  v: {vdouble}");
        System.Console.WriteLine($"store v2: {vdouble2}");
        System.Console.Write($"mem:     ");
        for (int i = 0; i < 2; i++) {
            fixed(double* ptr = &memdouble[i]) {
                AddAndStoreLaneDouble(vdouble, vdouble2, ptr);
                System.Console.Write($"  {memdouble[i]}");
            }
        }
        System.Console.WriteLine();
    }

    unsafe static void TestStore()
    {
        var mem = new byte[32];
        for (int i = 0; i < 16; i++) {
            var v = RandomByteVector;
            fixed(byte* ptr = &mem[i]) {
                PackedSimd.Store(ptr, v);
                System.Console.WriteLine($"store v: {v}");
                System.Console.Write("mem:    ");
                for (int j=0; j < 16; j++)
                    System.Console.Write($"  {mem[i+j]}");
                System.Console.WriteLine();
            }
        }
        var memUInt = new uint[8];
        for (int i = 0; i < 4; i++) {
            var v = RandomUIntVector;
            fixed(uint* ptr = &memUInt[i]) {
                PackedSimd.Store(ptr, v);
                System.Console.WriteLine($"store v: {v}");
                System.Console.Write("mem:    ");
                for (int j=0; j < 4; j++)
                    System.Console.Write($"  {memUInt[i+j]}");
                System.Console.WriteLine();
            }
        }
        var memDouble = new double[4];
        for (int i = 0; i < 2; i++) {
            var v = RandomDoubleVector;
            fixed(double* ptr = &memDouble[i]) {
                PackedSimd.Store(ptr, v);
                System.Console.WriteLine($"store v: {v}");
                System.Console.Write("mem:    ");
                for (int j=0; j < 2; j++)
                    System.Console.Write($"  {memDouble[i+j]}");
                System.Console.WriteLine();
            }
        }
    }

    [MethodImpl(MethodImplOptions.NoInlining)]
    static unsafe void AddAndStoreLaneByte(Vector128<byte> v1, Vector128<byte> v2, byte* ptr)
    {
        var v = PackedSimd.Add (v1, v2);
        PackedSimd.StoreSelectedScalar (ptr, v, (byte)5);
        PackedSimd.StoreSelectedScalar (ptr+1, v, (byte)2);
        PackedSimd.StoreSelectedScalar (ptr+2, v, (byte)7);
        PackedSimd.StoreSelectedScalar (ptr+3, v, (byte)11);
    }

    [MethodImpl(MethodImplOptions.NoInlining)]
    static unsafe void AddAndStoreLaneUInt(Vector128<uint> v1, Vector128<uint> v2, uint* ptr)
    {
        var v = PackedSimd.Add (v1, v2);
        PackedSimd.StoreSelectedScalar (ptr, v, (byte)2);
        PackedSimd.StoreSelectedScalar (ptr+1, v, (byte)3);
        PackedSimd.StoreSelectedScalar (ptr+2, v, (byte)1);
    }

    [MethodImpl(MethodImplOptions.NoInlining)]
    static unsafe void AddAndStoreLaneFloat(Vector128<float> v1, Vector128<float> v2, float* ptr)
    {
        var v = PackedSimd.Add (v1, v2);
        PackedSimd.StoreSelectedScalar (ptr, v, (byte)2);
        PackedSimd.StoreSelectedScalar (ptr+1, v, (byte)3);
        PackedSimd.StoreSelectedScalar (ptr+2, v, (byte)1);
    }

    [MethodImpl(MethodImplOptions.NoInlining)]
    static unsafe void AddAndStoreLaneDouble(Vector128<double> v1, Vector128<double> v2, double* ptr)
    {
        var v = PackedSimd.Add (v1, v2);
        PackedSimd.StoreSelectedScalar (ptr, v, (byte)1);
        PackedSimd.StoreSelectedScalar (ptr+2, v, (byte)0);
    }

output:

store v: <117, 173, 66, 144, 52, 239, 14, 55, 0, 194, 116, 105, 250, 87, 206, 155>
  mem:      117  173  66  144  52  239  14  55  0  194  116  105  250  87  206  155
  store v: <63, 109, 19, 55, 2, 220, 160, 164, 204, 154, 134, 112, 161, 5, 140, 64>
  mem:      63  109  19  55  2  220  160  164  204  154  134  112  161  5  140  64
  store v: <26, 101, 1, 239, 137, 222, 59, 195, 146, 216, 83, 88, 200, 212, 46, 157>
  mem:      26  101  1  239  137  222  59  195  146  216  83  88  200  212  46  157
  store v: <142, 177, 59, 49, 117, 182, 252, 208, 12, 115, 156, 0, 220, 235, 134, 202>
  mem:      142  177  59  49  117  182  252  208  12  115  156  0  220  235  134  202
  store v: <217, 241, 160, 0, 92, 170, 4, 169, 185, 202, 209, 183, 132, 165, 244, 21>
  mem:      217  241  160  0  92  170  4  169  185  202  209  183  132  165  244  21
  store v: <244, 78, 253, 232, 143, 44, 49, 37, 252, 113, 213, 9, 114, 199, 42, 65>
  mem:      244  78  253  232  143  44  49  37  252  113  213  9  114  199  42  65
  store v: <182, 19, 98, 30, 223, 133, 126, 194, 151, 131, 68, 140, 200, 35, 4, 6>
  mem:      182  19  98  30  223  133  126  194  151  131  68  140  200  35  4  6
  store v: <54, 108, 27, 227, 148, 59, 50, 145, 136, 244, 43, 198, 182, 207, 221, 168>
  mem:      54  108  27  227  148  59  50  145  136  244  43  198  182  207  221  168
  store v: <87, 134, 188, 110, 97, 7, 174, 51, 0, 127, 220, 153, 168, 12, 228, 211>
  mem:      87  134  188  110  97  7  174  51  0  127  220  153  168  12  228  211
  store v: <92, 133, 247, 249, 204, 134, 154, 155, 10, 146, 211, 201, 127, 160, 88, 32>
  mem:      92  133  247  249  204  134  154  155  10  146  211  201  127  160  88  32
  store v: <96, 5, 203, 0, 240, 58, 103, 130, 109, 214, 196, 248, 145, 42, 252, 170>
  mem:      96  5  203  0  240  58  103  130  109  214  196  248  145  42  252  170
  store v: <144, 207, 254, 244, 157, 21, 0, 96, 125, 204, 211, 109, 118, 53, 158, 6>
  mem:      144  207  254  244  157  21  0  96  125  204  211  109  118  53  158  6
  store v: <98, 116, 184, 75, 248, 201, 12, 235, 86, 230, 188, 142, 65, 128, 167, 9>
  mem:      98  116  184  75  248  201  12  235  86  230  188  142  65  128  167  9
  store v: <11, 25, 218, 151, 72, 76, 219, 202, 249, 126, 59, 184, 244, 4, 135, 49>
  mem:      11  25  218  151  72  76  219  202  249  126  59  184  244  4  135  49
  store v: <55, 12, 66, 128, 33, 84, 16, 163, 117, 229, 54, 216, 206, 84, 171, 154>
  mem:      55  12  66  128  33  84  16  163  117  229  54  216  206  84  171  154
  store v: <21, 142, 71, 53, 204, 33, 134, 17, 149, 46, 72, 49, 97, 189, 76, 11>
  mem:      21  142  71  53  204  33  134  17  149  46  72  49  97  189  76  11
  store v: <724078740, 108220663, 656527309, 1502734651>
  mem:      724078740  108220663  656527309  1502734651
  store v: <1922879322, 483655944, 1103094010, 1431166438>
  mem:      1922879322  483655944  1103094010  1431166438
  store v: <2078901068, 814392698, 427374792, 975870130>
  mem:      2078901068  814392698  427374792  975870130
  store v: <1299043465, 1606985734, 1245800259, 800066743>
  mem:      1299043465  1606985734  1245800259  800066743
  store v: <4.4542889197194E+153, -8.945975260274256E+153>
  mem:      4.4542889197194E+153  -8.945975260274256E+153
  store v: <-1.0813302999187009E+154, 1.2876150067212106E+154>
  mem:      -1.0813302999187009E+154  1.2876150067212106E+154
  store  v: <18, 132, 103, 108, 112, 175, 117, 35, 107, 118, 156, 7, 131, 247, 38, 235>
  store v2: <222, 206, 245, 126, 171, 181, 220, 165, 49, 25, 21, 26, 144, 83, 187, 254>
  mem:       100  100  100  100  100  100  100  100  100  100  100  100  100  100  100  100
  store  v: <698743856, 1370067327, 1301073778, 812248499>
  store v2: <2116578198, 352804974, 1298483573, 1864132694>
  mem:       2599557351  2599557351  2599557351  2599557351
  store  v: <1.2762932E+19, -4.5795344E+18, 7.776362E+18, -6.0136815E+18>
  store v2: <-1.1721688E+19, -9.542314E+18, 1.8291103E+19, 9.44995E+18>
  mem:       2.6067464E+19  2.6067464E+19
  store  v: <-3.309235460147638E+152, 1.2595791886048362E+154>
  store v2: <-2.061471279565945E+153, -6.855126615164575E+153>
  mem:       5.740665270883787E+153  5.740665270883787E+153

emitted code check:

> wa-info -d -f TestStore src/mono/sample/wasm/console-node/bin/Debug/AppBundle/dotnet.native.wasm|grep 'v128.store *\[SIMD\]'
     v128.store    [SIMD]
> wa-info -d -f AddAndStoreLane src/mono/sample/wasm/console-node/bin/Debug/AppBundle/dotnet.native.wasm|grep lane
 v128.store8.lane 11    [SIMD]
 v128.store8.lane 7    [SIMD]
 v128.store8.lane 2    [SIMD]
 v128.store8.lane 5    [SIMD]
 v128.store32.lane 1    [SIMD]
 v128.store32.lane 3    [SIMD]
 v128.store32.lane 2    [SIMD]
 v128.store32.lane 1    [SIMD]
 v128.store32.lane 3    [SIMD]
 v128.store32.lane 2    [SIMD]
 v128.store64.lane 0    [SIMD]
 v128.store64.lane 1    [SIMD]