This is the mail archive of the gcc-bugs@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[Bug target/33671] incorrect vector codegen on PPU at O3 level optimization



------- Comment #4 from j dot m dot taylor at dur dot ac dot uk  2007-10-05 22:06 -------
> Yep, including -mabi=altivec fixes the issue so this is not a real issue.
Thankyou for your prompt reply. I was not aware that I needed to do this. I am
glad to hear it's a compiler option issue rather than a true codegen bug, but I
still would argue that there is a bug in here, even if it's only a UI one.

I have passed gcc a series of options and it has compiled the source without
any error messages or warnings, but has produced INCORRECT CODE which it has
taken me a whole day to narrow down. It seems to me that if gcc is going to
produce incorrect code then it should surely be warning me about this.

All the man page says is that I "may also need to set -mabi=altivec".
Presumably there is some bizarre circumstance in which it might be preferable
not to set that option, but this case is clearly not one of them! I cannot see
how it can be possible to write any code whatsoever using the altivec
extensions without risking encountering this same issue.

The only other discussion I can find on this issue is here:
http://gcc.gnu.org/ml/gcc-patches/2007-08/msg00438.html
where someone else also seems to be arguing that this situation needs
addressing.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33671


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]