https://gcc.gnu.org/bugzilla/show_bug.cgi?id=56253
Matthias Kretz changed:
What|Removed |Added
Status|NEW |RESOLVED
Resolution|---
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=56253
--- Comment #15 from Marc Glisse ---
Oups, sorry, or_ps may be harder, because representing it with vector
extensions requires casts to integer vectors, which makes it much harder to
actually generate or_ps in the backend (there is at least one P
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=56253
--- Comment #14 from Marc Glisse ---
(In reply to Agner Fog from comment #13)
> Is it possible to do the same optimization with boolean vector intrinsics,
> such as _mm_and_epi32 and _mm_or_ps to enable optimizations such as
> algebraic reduction
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=56253
--- Comment #13 from Agner Fog ---
Thank you. I agree that integer overflow should be well-defined when using
intrinsics.
Is it possible to do the same optimization with boolean vector intrinsics, such
as _mm_and_epi32 and _mm_or_ps to enable op
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=56253
--- Comment #12 from Marc Glisse ---
(In reply to Agner Fog from comment #11)
> Thanks for the links Marc.
> You are right, the discussion in the gcc-patches mailing list ignores
> integer vectors. You need a solution that also allows optimizati
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=56253
--- Comment #11 from Agner Fog ---
Thanks for the links Marc.
You are right, the discussion in the gcc-patches mailing list ignores integer
vectors. You need a solution that also allows optimizations on integer
intrinsic functions (perhaps cast
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=56253
--- Comment #10 from Marc Glisse ---
Two random links into the latest conversation:
https://gcc.gnu.org/ml/gcc-patches/2014-07/msg01812.html
https://gcc.gnu.org/ml/gcc-patches/2014-06/msg02288.html
That version of the patch doesn't handle integ
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=56253
--- Comment #9 from Agner Fog ---
Many programmers are using a vector class library rather than writing intrinsic
functions directly. Such libraries have overloaded operators which are inlined
to produce intrinsic functions. Therefore, we cannot
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=56253
Agner Fog changed:
What|Removed |Added
CC||agner at agner dot org
--- Comment #8 from A
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56253
--- Comment #7 from Richard Biener 2013-02-11
11:36:29 UTC ---
So to re-cap, for vector intrinsics we can use the vector GCC extensions
while for the scalar intrinsics we would have to use target builtin folding.
Not sure if it's worth the
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56253
--- Comment #6 from Uros Bizjak 2013-02-08 14:58:42
UTC ---
(In reply to comment #5)
> Hmm, I wonder how the vectorizer then accesses add/sub patterns without
> SSE math. It just queries optabs ...
>
> We cannot handle the FMA case wit
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56253
--- Comment #5 from Richard Biener 2013-02-08
13:42:23 UTC ---
(In reply to comment #3)
> (In reply to comment #1)
>
> > not sure why we use builtins for these basic operations...
>
> Because they have to be emitted also for non-SSE m
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56253
--- Comment #4 from Richard Biener 2013-02-08
13:30:14 UTC ---
(In reply to comment #3)
> (In reply to comment #1)
>
> > not sure why we use builtins for these basic operations...
>
> Because they have to be emitted also for non-SSE m
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56253
--- Comment #3 from Uros Bizjak 2013-02-08 12:58:49
UTC ---
(In reply to comment #1)
> not sure why we use builtins for these basic operations...
Because they have to be emitted also for non-SSE math.
>From config/i386/sse.md:
;;
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56253
--- Comment #2 from Richard Biener 2013-02-08
12:52:32 UTC ---
For _mm256_fmadd_pd and friends the only possibility is to fold the target
builtins via targetm.fold_builtin to FMA_EXPR. Which is of course also
possible for the simple add/m
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56253
Richard Biener changed:
What|Removed |Added
Target||x86_64-*-*, i?86-*-*
16 matches
Mail list logo