Re: [eigen] Eigen::Complex -- does it have a reason to exist?

[Mark Borgerding <mark@xxxxxxxxxxxxxx>]

Benoit Jacob wrote:
> There is still a problem: so far, we use the criterion "a packet
> consists of a single scalar", formulated
> ei_packet_traits<Scalar>::Size==1, to determine if a scalar type
> doesn't have vectorization. This test breaks on complex<double> as its
> sizeof is 16, so we need a more future-proof test.
>   

I think the benefits of explicitly optimizing std::complex<double> with 
SIMD intrinsics/asm would be fairly small for exactly the reason you 
specify: there is exactly one complex<double> in a 128 bit simd 
register.  This should be child's play for a compiler.  The only problem 
I can think of is that the compiler might not assume 16 byte alignment 
of complex<double>.  We may be able to work around that with 
__attribute__((__vector_size__(16)))

DANGER ZONE: I am speculating without the benefit of benchmarks.  Does 
anyone have any concrete knowledge to either support or refute my 
speculation?


> Thanks a lot for this observation. I am sorry for having suggested you
> to work on something we may end up not using after all, and I
> appreciate a lot that you brought this issue up!
>   
I don't think you really suggested it, at least not very strongly.  I 
was mainly curious to see if the idea would work to have an 'operator &' 
that returns a struct that is convertible to one of two pointer types.  
Although the basic idea is sound, I would not consider it a success.  I 
thought I should be the one to kill it, since I gave birth to the monster.

That is, unless someone can come up with a good reason for it to exist.  
Going once... Going twice ....