Everywhere where cuda.syncthreads or __syncthreads is used, use the
properly namespaced nvvm.barrier0 instead.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274664 91177308-0d34-0410-b5e6-96231b3b80d8
Summary:
I've completed my audit of all the code that looks at noduplicate and
added handling of convergent where appropriate, so we no longer need
noduplicate on these intrinsics.
Reviewers: jholewinski
Subscribers: llvm-commits, jholewinski
Differential Revision: http://reviews.llvm.org/D18168
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264107 91177308-0d34-0410-b5e6-96231b3b80d8
Summary:
This is the attribute purpose-made for e.g. __syncthreads. It appears
that NoDuplicate may not be sufficient to prevent Sink from touching a
call to __syncthreads.
Reviewers: jingyue, hfinkel
Subscribers: llvm-commits, jholewinski, jhen, rnk, tra, majnemer
Differential Revision: http://reviews.llvm.org/D16941
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@260005 91177308-0d34-0410-b5e6-96231b3b80d8
The "noduplicate" function attribute exists to prevent certain optimizations
from duplicating calls to the function. This is important on platforms where
certain function call duplications are unsafe (for example execution barriers
for CUDA and OpenCL).
This patch makes it possible to specify intrinsics as "noduplicate" and
translates that to the appropriate function attribute.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204200 91177308-0d34-0410-b5e6-96231b3b80d8
