------- Comment #25 from gdr at integrable-solutions dot net 2010-02-18 04:55
-------
Subject: Re: mixed complex<T> multiplication horribly
inefficient
On Wed, Feb 17, 2010 at 4:54 PM, jan at epgmod dot phys dot tue dot nl
<gcc-bugzi...@gcc.gnu.org> wrote:
> ------- Comment #13 from jan at epgmod dot phys dot tue dot nl 2010-02-17
> 22:54 -------
> (In reply to comment #10)
>> Regarding the specific semantics, Richard, there is little to say: we want
>> the
>> C99 semantics, by and large. Maybe there are some missing details, but we
>> decided already that, when we switched the complex multiplication to be the
>> slow, C99 annex, by default. Thus, in a first, and second too, approx, the
>> C++
>> standard doesn't count, really, this is just, sketched:
>>
>> __complex__ double val;
>> double d;
>> val *= d;
>>
>
> But the point in question is that gcc *does not implement Annex G-4.1 (item 2)
> correctly*. That clearly states that u*(x+iy) = (ux)+i(uy), whereas gcc does
> (u+0i)*(x+iy), leading to wrongly signed zeros. The patch fixes exactly that
> (for C++) by circumventing gcc's flawed implementation of mixed arithmetic
> involving C99's complex type.
then, the proper fix is to fix should be in the compiler. Not in the library.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43108
