https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96918
--- Comment #10 from Hongtao.liu ---
(In reply to Hongtao.liu from comment #9)
> Or the backend add combine helper insn to match
>
> Failed to match this instruction:
> (set (reg:V8HI 90)
> (rotate:V8HI (reg:V8HI 91)
> (const_int 8
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96918
--- Comment #9 from Hongtao.liu ---
Or the backend add combine helper insn to match
Failed to match this instruction:
(set (reg:V8HI 90)
(rotate:V8HI (reg:V8HI 91)
(const_int 8 [0x8])))
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96918
Andrew Pinski changed:
What|Removed |Added
Ever confirmed|0 |1
Last reconfirmed|
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96918
--- Comment #7 from Richard Biener ---
(In reply to Jakub Jelinek from comment #6)
> Or the generic code could try to expand vector rotates by multiplies of
> BITS_PER_UNIT as vector permutations, perhaps only if there is no optab for
> it. Or
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96918
--- Comment #6 from Jakub Jelinek ---
Or the generic code could try to expand vector rotates by multiplies of
BITS_PER_UNIT as vector permutations, perhaps only if there is no optab for it.
Or trying to expand both permutation and rotate and
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96918
--- Comment #5 from Marc Glisse ---
typedef unsigned short v8i16 __attribute__((vector_size(16)));
v8i16 bswap_epi16(v8i16 x)
{
return (x << 8) | (x >> 8);
}
We do recognize a rotate already in GENERIC
return x r<< 8;
But this is
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96918
--- Comment #4 from Gabriel Ravier ---
Oh, now I realise that those headers are actually in fact part of directly
GCC-provided headers and not some external package. I must say though, if those
functions are internal implementation detail and
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96918
Jakub Jelinek changed:
What|Removed |Added
CC||jakub at gcc dot gnu.org
--- Comment #3
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96918
--- Comment #2 from Gabriel Ravier ---
Yes, you can reproduce this with _mm_shuffle_epi8, _mm_slli_epi16 and
_mm_srli_epi16. I'm assuming GCC developers are more familiar with the internal
intrinsics than with the Intel-provided intrinsics, so I
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96918
--- Comment #1 from Richard Biener ---
Note it's not so interesting to handle testcases with __builtin_ia32_* which
are internal but I guess the same can be replicated using intrinsics?
10 matches
Mail list logo
