[Issue 19940] New: core.simd.{load,store}Unaligned functions do not support 32-byte vectors

[d-bugmail at puremagic.com]

https://issues.dlang.org/show_bug.cgi?id=19940

          Issue ID: 19940
           Summary: core.simd.{load,store}Unaligned functions do not
                    support 32-byte vectors
           Product: D
           Version: D2
          Hardware: All
                OS: All
            Status: NEW
          Severity: normal
          Priority: P1
         Component: druntime
          Assignee: nobody at puremagic.com
          Reporter: slavo5150 at yahoo.com

void copy32(T)(T* d, const T* s)
{
    import core.simd;
    static foreach(i; 0..T.sizeof/32)
    {
        // Error: {load,store}Unaligned is not supported for 32-byte vectors
        storeUnaligned(cast(void32*)(d) + i, loadUnaligned(cast(const
void32*)(s) + i));
    }
}

void copy16(T)(T* d, const T* s)
{
    import core.simd;
    static foreach(i; 0..T.sizeof/16)
    {
        // This works fine for 16-byte vectors
        storeUnaligned(cast(void16*)(d) + i, loadUnaligned(cast(const
void16*)(s) + i));
    }
}

struct S(size_t n)
{
    align(32)
        ubyte[n] x;   
}

void main()
{
    S!64 s1;
    S!64 s2;

    copy16(&s2, &s1);
    copy32(&s2, &s1);
}

Compile with `dmd -mcpu=avx`.  The {load,store}Unaligned calls with 32-byte
vector types will fail to compile, but will work fine for 16-byte vector types.

--