Bug on supporting bit_cast<uint64_t>(double)

[kaizhangNV]

Following code will cause internal error:

RWStructuredBuffer<uint64_t> out;

uint64_t cast(double x)
{
    return bit_cast<uint64_t>(x);
}

[shader("compute")]
[numthreads(1,1,1)]
void computeMain()
{
    double x = 1.0;

    out[0] = cast(x);
}

note, there is no problem just doing

  out[0] = bit_cast<uint64_t>(1.0lf);

[kaizhangNV]

The error message is:

test1.slang(6): internal error 99999: unimplemented feature in Slang compiler: unexpected IR opcode during code emit
    return bit_cast<uint64_t>(x);

[kaizhangNV]

I tried other targets
It has no problem on cpp, cuda, spirv.

[kaizhangNV]

on all the shader language targets, we don't handle double base type during processing kIROp_BitCast.

[kaizhangNV]

There is a TODO in hlsl target
// TODO: There is an asdouble function
// for converting two 32-bit integer values into
// one double. We could use that for
// bit casts of 64-bit values with a bit of
// extra work, but doing so might be best
// handled in an IR pass that legalizes
// bit-casts.
//

[csyonghe]

We should work on top of #5020 for this issue.