projects / gcc.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 74cc0b2) | patch
re PR target/80846 (auto-vectorized AVX2 horizontal sum should narrow to 128b right...
authorJakub Jelinek <jakub@redhat.com>
Tue, 1 Aug 2017 16:12:31 +0000 (18:12 +0200)
committerJakub Jelinek <jakub@gcc.gnu.org>
Tue, 1 Aug 2017 16:12:31 +0000 (18:12 +0200)
commit8e1863eca8c8914cc5e05369dacdc5266cdca179
tree2768456226ce5912779782a5949c5a6a61a7bebatree
parent74cc0b2b66126e4598c4412ccbd879caa287ca81commit | diff
re PR target/80846 (auto-vectorized AVX2 horizontal sum should narrow to 128b right away, to be more efficient for Ryzen and Intel)

PR target/80846
* config/rs6000/vsx.md (vextract_fp_from_shorth,
vextract_fp_from_shortl): Add element mode after mode in gen_vec_init*
calls.

From-SVN: r250784
gcc/ChangeLog diff | blob | history
gcc/config/rs6000/vsx.md diff | blob | history