32x32 matrix transpose in metal using subgroups

bitmap_transpose.metal

inline uint shuffle_round(uint a, uint m, ushort s) {
    uint b = simd_shuffle_xor(a, s);
    uint c;
    if ((tix & s) == 0) {
        c = b << s;
    } else {
        m = ~m;
        c = b >> s;
    }
    return (a & m) | (c & ~m);
}

    // Transpose bitmask. Assumes subgroup size = 32
    uint bitmask; // value to transpose
    bitmask = shuffle_round(bitmask, 0xffff, 16);
    bitmask = shuffle_round(bitmask, 0xff00ff, 8);
    bitmask = shuffle_round(bitmask, 0xf0f0f0f, 4);
    bitmask = shuffle_round(bitmask, 0x33333333, 2);
    bitmask = shuffle_round(bitmask, 0x55555555, 1);