Re: [eigen] question about ploadu/pstoreu and alignment

[Konstantinos Margaritis <markos@xxxxxxxxxxx>]

On Thu, 18 Sep 2014 15:28:31 +0200
Gael Guennebaud <gael.guennebaud@xxxxxxxxx> wrote:

> Hi,
> On all other platforms we assume at least a 4 bytes alignment, so you
> should be fine here.

Cool, good to know that, in that case the packetmath test, shouldn't it
be modified not to fail in odd alignments? perhaps make it give a
warning that word-alignment works, but not arbitrary ones. Or test only
word-alignment (ie move in steps of sizeof(Scalar) and not +1).

> Right, ploadu/pstoreu are to handle cases for which aligned access
> cannot be guaranteed, but that does not mean that the data is not
> aligned at all. I don't really see why this would make the
> implementation of ploadu more difficult/expensive though.

Ok, that might actually be some bug on my code, seems that with zero
alignment it breaks. I'll take a closer look.

Konstantinos

Attachment: pgpXEdCdOTiLi.pgp
Description: PGP signature