diff options
| author | Richard Biener <rguenther@suse.de> | 2024-11-13 14:43:27 +0100 |
|---|---|---|
| committer | Richard Biener <rguenth@gcc.gnu.org> | 2024-11-14 09:06:16 +0100 |
| commit | 4b4aa47ed296968507b2fde70d2e651129ff3b36 (patch) | |
| tree | 658c698467e4ebc22ae925a40b0b20a25191b494 /gcc/cfgexpand.cc | |
| parent | 735f5260fb42919a651ed8e381c9e1a11e753345 (diff) | |
| download | gcc-4b4aa47ed296968507b2fde70d2e651129ff3b36.zip gcc-4b4aa47ed296968507b2fde70d2e651129ff3b36.tar.gz gcc-4b4aa47ed296968507b2fde70d2e651129ff3b36.tar.bz2 | |
tree-optimization/117556 - SLP of live stmts from load-lanes
The following fixes SLP live lane generation for load-lanes which
fails to analyze for gcc.dg/vect/vect-live-slp-3.c because the
VLA division doesn't work out but it would also wrongly use the
transposed vector defs I think. The following properly disables
the actual load-lanes SLP node from live lane processing and instead
relies on the SLP permute node representing the live lane where we
can use extract-last to extract the last lane. This also fixes
the reported Ada miscompile.
PR tree-optimization/117556
PR tree-optimization/117553
* tree-vect-stmts.cc (vect_analyze_stmt): Do not analyze
the SLP load-lanes node for live lanes, but only the
permute node.
(vect_transform_stmt): Likewise for the transform.
* gcc.dg/vect/vect-live-slp-3.c: Expect us to SLP even for
VLA vectors (in single-lane mode).
Diffstat (limited to 'gcc/cfgexpand.cc')
0 files changed, 0 insertions, 0 deletions