On Tue, Aug 09, 2016 at 03:13:30PM -0700, David Wohlferd wrote: > I'm looking at gcc/config/i386/avx512fintrin.h, and I see this: > > extern __inline __m256i > __attribute__ ((__gnu_inline__, __always_inline__, __artificial__)) > _mm512_cvtsepi64_epi32 (__m512i __A) > { > __v8si __O; > return (__m256i) __builtin_ia32_pmovsqd512_mask ((__v8di) __A, > (__v8si) > _mm256_undefined_si256 (), > (__mmask8) -1); > } > > Since r208793, __O appears to be unused? > > For someone with checkin permissions (ie not me), removing this seems like > an "as obvious."
Indeed, I've committed following to trunk: Index: ChangeLog =================================================================== --- ChangeLog (revision 239309) +++ ChangeLog (working copy) @@ -1,3 +1,8 @@ +2016-08-09 David Wohlferd <d...@limegreensocks.com> + + * config/i3836/avx512fintrin.h (_mm512_cvtsepi64_epi32): Remove + unused variable __O. + 2016-08-09 Martin Liska <mli...@suse.cz> * doc/gcov.texi: Change _gcov_dump to __gcov_dump and Index: config/i386/avx512fintrin.h =================================================================== --- config/i386/avx512fintrin.h (revision 239309) +++ config/i386/avx512fintrin.h (working copy) @@ -4903,7 +4903,6 @@ extern __inline __m256i __attribute__ ((__gnu_inline__, __always_inline__, __artificial__)) _mm512_cvtsepi64_epi32 (__m512i __A) { - __v8si __O; return (__m256i) __builtin_ia32_pmovsqd512_mask ((__v8di) __A, (__v8si) _mm256_undefined_si256 (), Jakub