This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Another AltiVec problem
- From: Daniel Egger <degger at fhm dot edu>
- To: Aldy Hernandez <aldyh at redhat dot com>
- Cc: GCC Developer Mailinglist <gcc at gcc dot gnu dot org>
- Date: 16 Mar 2002 00:47:27 +0100
- Subject: Re: Another AltiVec problem
- References: <E73AC93B-3706-11D6-B8DB-000393750C1E@redhat.com>
Am Don, 2002-03-14 um 05.49 schrieb Aldy Hernandez:
> the answer should be 3*2 + 1 = 7, etc etc
I agree.
> [7 7 7] which is the answer i get from a fresh build from
> the gcc 3.1 tree. i also tried it with the apple compiler,
> and it agrees as well :).
Well, I get
5.000000 5.000000 5.000000 5.000000
and I have not the slightest idea what's going on. Also I figured out
that the PEM is inconsistent in the detailed description of the
mnemonic versus the list of mnemonics. It differs exactly in the
way my test differs from the expected result, i.e. two operands
are swapped.
Can we get an agreement, what exactly is supposed to be correct?
--
Servus,
Daniel