diff options
| author | Alex MacLean <amaclean@nvidia.com> | 2025-04-14 16:50:19 -0700 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2025-04-14 16:50:19 -0700 |
| commit | 07fe9c6461dbdde6f619fc56bd021cd30b3f6327 (patch) | |
| tree | 131d285b8f0348b69da6e7f8af610c78b44b5897 /llvm/lib/Analysis/VectorUtils.cpp | |
| parent | 8e8d04870dbc3248301534796b4370f75e33f7c0 (diff) | |
| download | llvm-07fe9c6461dbdde6f619fc56bd021cd30b3f6327.zip llvm-07fe9c6461dbdde6f619fc56bd021cd30b3f6327.tar.gz llvm-07fe9c6461dbdde6f619fc56bd021cd30b3f6327.tar.bz2 | |
[NVPTX] Use 0 immediate for i1 trunc, cleanup dead code (#135646)
Update the instruction selection for truncation to i1 to use "setp.ne
%v, 0" as the zero immediate is a preferable canonical form. Also remove
some dead code relating to the "set" instruction which we do not
actually support currently.
Diffstat (limited to 'llvm/lib/Analysis/VectorUtils.cpp')
0 files changed, 0 insertions, 0 deletions