You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We're relying on LLVM to print the list of supported CPUs and features, and as a result the output is confusing as it contains
For example, llc -mcpu=mycpu -mattr=+feature1,-feature2
For us I believe the syntax is -C cpuname,+feature1,-feature2;...;..., where ; delimits different configurations to support. Each configuration can also contain clone_all, in which case all code will be (re)generated for that architecture, or base(n) meaning we will try to share code with configuration #n (0-based I think?).
We're relying on LLVM to print the list of supported CPUs and features, and as a result the output is confusing as it contains
For us I believe the syntax is
-C cpuname,+feature1,-feature2;...;...
, where;
delimits different configurations to support. Each configuration can also containclone_all
, in which case all code will be (re)generated for that architecture, orbase(n)
meaning we will try to share code with configuration #n (0-based I think?).There is a comment about this in
jl_init_llvm
.See #21849
The text was updated successfully, but these errors were encountered: