Re: vector alignment

[Richard Biener <richard dot guenther at gmail dot com>]

On Tue, Apr 2, 2019 at 6:20 PM Martin Sebor <msebor@gmail.com> wrote:
>
> GCC tries to align a vector on its natural boundary, i.e., that
> given by its size, up to MAX_OBJECT_ALIGNMENT.  Vectors that are
> bigger than that are either silently [mis]aligned on that same
> maximum boundary (PR 89798), silently truncated (and misaligned),
> or cause an ICE (PR 89797).  Compiling the following:
>
>    __attribute__ ((vector_size (N))) char v;
>
>    _Static_assert (sizeof (v) == N, "size");
>    _Static_assert (__alignof__ (v) == N, "alignment");
>
> with N set to 1LLU << I shows these failures:
>
>    I < 29   succeeds
>    I < 31   fails alignment
>    I < 32   ICE
>    I >= 32  fails alignment and size
>
> Attribute aligned doesn't seem to have any effect on types or
> variables declared with attribute vector_size.  The alignment
> set by the latter prevails.
>
> This happens no matter what scope the vector is defined in (i.e.,
> file or local).
>
> I have some questions:
>
> 1) Is there some reason to align vectors on the same boundary
>     as their size no matter how big it is?  I can't find such
>     a requirement in the ABIs I looked at.  Or would it be more
>     appropriate to align the big ones on the preferred boundary
>     for the target?  For instance, does it make more sense to
>     align a 64KB vector on a 64KB boundary than on, say,
>     a 64-byte boundary (or some other boundary less than 64K?)

I don't think there's a good reason.  Instead I think that
BIGGEST_ALIGNMENT is what we should go for as upper limit,
anything bigger doesn't make sense (unless the user explicitely
requests it).

> 2) If not, is it then appropriate to underalign very large
>     vectors on a boundary less than their size?

Yes.

> 3) Should the aligned attribute not override the default vector
>     alignment?

Yes, but doesn't it already?

> I would like to think the answer to (1) is that vectors should
> be aligned on the preferred boundary for the target/ABI.  If
> that's feasible, it should also obviate question (2).
>
> I believe the answer to (3) is yes.  If not, GCC should issue
> a warning that it doesn't honor the aligned attribute.
>
> Thanks
> Martin