This revision was automatically updated to reflect the committed changes.
Closed by commit rG970569b6ccb0: [AMDGPU] __builtin_amdgcn_read_exec_* should
be implemented with llvm.amdgcn. (authored by ranapratap55).
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.l
ranapratap55 added inline comments.
Comment at: clang/lib/CodeGen/CGBuiltin.cpp:17496-17497
+CallInst *CI =
+cast(EmitAMDGCNBallotForExec(*this, E, Int64Ty, Int64Ty));
CI->setConvergent();
return CI;
arsenm wrote:
> Don't need the cast or se
