This revision was landed with ongoing or failed builds.
This revision was automatically updated to reflect the committed changes.
Closed by commit rGdee7704829bd: [AMDGPU] Add __builtin_amdgcn_grid_size
(authored by JonChesterfield).
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTIO
yaxunl accepted this revision.
yaxunl added a comment.
This revision is now accepted and ready to land.
LGTM. Thanks.
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D90251/new/
https://reviews.llvm.org/D90251
yaxunl added a comment.
I think there is value in this change. It may help simplify device library and
hide some compiler details.
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D90251/new/
https://reviews.llvm.org/D90251
__
JonChesterfield added a comment.
Consistency really. It seemed strange to have a builtin for reading the
workgroup size and not one for the grid size.
There's probably a range limit that can be set on this one too, I'm just not
sure what it is.
Repository:
rG LLVM Github Monorepo
CHANGES S
arsenm added a comment.
What's the point of this? The reason for the other case was because there was
no other way to attach the range metadata. The invariant load here is
redundant since with AMDGPU AA the load from constant will be treated as
invariant anyway
Repository:
rG LLVM Github M
JonChesterfield created this revision.
JonChesterfield added reviewers: yaxunl, arsenm, b-sumner, cfang, rjmccall,
Anastasia.
Herald added subscribers: openmp-commits, cfe-commits, dexonsmith, kerbowa,
t-tye, tpr, dstuttard, nhaehnle, jvesely, kzhuravl.
Herald added projects: clang, OpenMP.
JonCh
