tuxd3v | raspbian, compiles not only the kernel but also the userspace with hardfloat support.. | 00:19 |
---|---|---|
tuxd3v | by default armv6 doesn't offer vfp on the base configuration, its an extra.. | 00:20 |
tuxd3v | but rpi arm1176jzf-s as support for hardfloat operations.. | 00:21 |
tuxd3v | debian armel, doesn't support hardfloat, because it is not in the base specs of armv6( its a extra.. ) | 00:22 |
tuxd3v | also armel suports >=armv5t til armv6 | 00:23 |
Tenkawa | tuxd3v: yes.. I know the hardware and the "foundation" software does | 00:23 |
Tenkawa | but its heavily modified to do that | 00:23 |
tuxd3v | I don't know but if armv6 doesn't offer be default hardfloat, for sure amdv5 also doesn't.. | 00:23 |
tuxd3v | Tenkawa, indeed | 00:23 |
tuxd3v | its a lot, lot of work for that.. | 00:24 |
Tenkawa | thats what I was pointing out earlier is that anyone not using raspios is going to have to deal with softfp | 00:24 |
Tenkawa | for armhf | 00:24 |
Tenkawa | pseudo armhf | 00:25 |
Tenkawa | (pi 1, zero, pi 2, anything non raspios) | 00:25 |
tuxd3v | you can build applications on armel with hardfloat support, but the userland provided by armel is softfloat | 00:25 |
Tenkawa | no you cant | 00:26 |
Tenkawa | not non-foundation | 00:26 |
Tenkawa | you "must" use their compiler | 00:26 |
tuxd3v | you can compile code on rpi 1/zeros with floating point support, but you need to compile it yourself, and have a compiler that support that options. | 00:26 |
Tenkawa | thats my point | 00:27 |
Tenkawa | theres not one | 00:27 |
Tenkawa | at least I havent been able to find one yet | 00:27 |
tuxd3v | I never tried, but maybe you are right, and maybe the compiler that exist in the repos doesn't support hardfloat, but I never tried.. | 00:28 |
Tenkawa | they had to build their own toolchain.. no it supports hard float but they built it and I'm not sure they ever released the sources | 00:29 |
Tenkawa | its possible they did.. | 00:29 |
Tenkawa | i think the linaro toolchain is.. the problem is its very old | 00:29 |
Tenkawa | honestly the zero/pi 1 isnt going to gain much between softfp vs hardfp anyway | 00:30 |
Tenkawa | to make it worth it | 00:30 |
Tenkawa | I'm going to focus efforts on 64 bit efficiency | 00:31 |
Tenkawa | (which pi 1 and zero won't be using heheh) | 00:31 |
tuxd3v | indeed | 00:32 |
Tenkawa | now if they would ever release a 64 bit zero :) | 00:32 |
Tenkawa | heheh | 00:32 |
Tenkawa | ok back to food search | 00:33 |
tuxd3v | :) | 00:34 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!