Re: give back for ffmpeg on ppc64
On Thu, Jun 23, 2016 at 09:02:51AM +0200, Mathieu Malaterre wrote:
> Indeed. What are the machine amaterasu and tsukuyomi, do they have altivec ?
>
> I cannot reproduce FTBFS on the sole ppc64 porterbox we have: pizzetti.d.o:
>
> malat@pizzetti ~ % cat /proc/cpuinfo
> processor : 0
> cpu : POWER8E (raw), altivec supported
> clock : 3425.000000MHz
> revision : 2.1 (pvr 004b 0201)
>
> processor : 1
> cpu : POWER8E (raw), altivec supported
> clock : 3425.000000MHz
> revision : 2.1 (pvr 004b 0201)
>
> processor : 2
> cpu : POWER8E (raw), altivec supported
> clock : 3425.000000MHz
> revision : 2.1 (pvr 004b 0201)
>
> processor : 3
> cpu : POWER8E (raw), altivec supported
> clock : 3425.000000MHz
> revision : 2.1 (pvr 004b 0201)
>
> timebase : 512000000
> platform : pSeries
> model : IBM pSeries (emulated by qemu)
I don't have access to a 64 bit powerpc without altivec either, but I
tried to see if qemu would be an option since at least POWER5+ should
be a 64 bit CPU without altivec, as should the e5500.
Unfortunately trying qemu-ppc64 with e5500 as CPU blows up on everything
in ppc64 I tried. POWER5+ can run some stuff (like ls, but not ls -l),
so not much luck there. Trying to run qemu-system-ppc64 with CPU POWER5+
seems to fail to boot the debian powerpc netinst. It seems to get stuck
right after "returning from prom_init". So no luck trying anything
so far. Maybe qemu is just not quite ready for the task.
qemu-ppc64 with default CPU seems able to run everything just fine though,
but of course then it has altivec. I suppose digging up a very old
qemu from before altivec was added would be an option but it probably
has different bugs to deal with.
--
Len Sorensen
Reply to: