Re: armelfp: new architecture name for an armel variant
On Tue, Jul 06, 2010, Joey Hess wrote:
> Could the targeted CPU be used in the name? Ie, armelv7.
I would be a bit scared that this has a chance of getting out of date,
or be confusing because other ports might be v7 as well, or also
because this only reflects a subset of the ports' requirement (VFP
level for instance isn't reflected, such as vfpv2, vfpv3).
> FWIW, I've benchmarked a VFP libm to run benchmarks like whetstone 4x
> faster than regular armel libm. A VFP Mesa can speed up glxgears by
> nearly 2x on some hardware. It would be handy to have VFP versions of
> such targeted libraries. In general though VFP is not worth it.
Mesa is indeed an interesting package to provide a VFP version of, good
The approach of alternative packages is a bit painful:
- one can ship vfp and non-vfp lib in the same package, but then
everybody pays the disk space for both
- one can split them up, but how do they get properly pulled on user
- and the changes to the rules to have multiple passes might be
intrusive in some packages