]> granicus.if.org Git - clang/commit
[CUDA] Use activemask.b32 instruction to implement __activemask w/ CUDA-9.2+
authorArtem Belevich <tra@google.com>
Tue, 3 Sep 2019 17:31:58 +0000 (17:31 +0000)
committerArtem Belevich <tra@google.com>
Tue, 3 Sep 2019 17:31:58 +0000 (17:31 +0000)
commit4fbb9360b74b83cec392f205eb7df60838482304
tree961e77ad4e659df20dfec1c4a1c48a5cd385253f
parent192c1e3f91f4364f452ff23e57340aaaa2d83b80
[CUDA] Use activemask.b32 instruction to implement __activemask w/ CUDA-9.2+

vote.ballot instruction is gone in recent CUDA versions and
vote.sync.ballot can not be used because it needs a thread mask parameter.
Fortunately PTX 6.2 (introduced with CUDA-9.2) provides activemask.b32
instruction for this.

Differential Revision: https://reviews.llvm.org/D66665

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@370792 91177308-0d34-0410-b5e6-96231b3b80d8
lib/Headers/__clang_cuda_intrinsics.h