https://gcc.gnu.org/bugzilla/show_bug.cgi?id=120639
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Last reconfirmed| |2025-06-19
Blocks| |53947
Status|UNCONFIRMED |NEW
Component|middle-end |tree-optimization
Keywords| |missed-optimization
Ever confirmed|0 |1
--- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> ---
We could use scatter stores, building the index vector somehow cleverly with
i_width contiguous indexes interspaced by i_dst_stride. In fact this vector
could be built as inductions when building the i_height number of vectors
to store and concatenated the same way?
On no x86 uarch are scatters implemented efficiently enough to be worth this
though.
Stores with variable gaps in general ask for scatters, with constant known
gaps contiguous stores with mask might work. But the code we emit now is
quite efficient on the store side (it does force a higher VF sometimes,
due to the single-vector-size limitation).
Referenced Bugs:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=53947
[Bug 53947] [meta-bug] vectorizer missed-optimizations
