Opinions on packaging ATLAS (for the x86 architecture)

Jonathan Underwood jonathan.underwood at gmail.com
Fri Sep 25 17:25:19 UTC 2009


2009/9/25 Jussi Lehtola <jussilehtola at fedoraproject.org>:
> On Fri, 2009-09-25 at 16:30 +0100, Jonathan Underwood wrote:
>> Well, my point was that packages that do respect CFLAGS get built
>> currently with certain options, and however Atlas builds, the default
>> package should have comparable options. On x86, I see that is
>> presently -march=i686 -mtune=atom, and IIRC -march=i686 implies sse
>> and sse2 (not sure about sse3). Anyway, whatever it implies, I think
>> that should be the default for the main Atlas package too.
>
> No, -march=i686 does not use SSE; excerpt from gcc man page:
>
> i686 Same as "generic", but when used as "march" option, PentiumPro
>     instruction set will be used, so the code will run on all i686
>     family chips.

Ah, right, thanks. Then I am advocating that Atlas be built without
SSE for the main package (option 1 of Deji's proposals, I think).




More information about the devel mailing list