[PATCH] D12246: [NVPTX] change threading intrinsics from noduplicate to convergent
Bjarke Roune via llvm-commits
llvm-commits at lists.llvm.org
Mon Aug 31 10:01:14 PDT 2015
On Fri, Aug 28, 2015 at 11:01 PM, Owen Anderson <resistor at mac.com> wrote:
> I deliberately didn’t add any constraints on duplication, both because IME
> it’s difficult to deal with in practice, and because I have use cases in
> mind that don’t care about duplication the way that a barrier does.
> We're seeing that for convergent to best fit __syncthreads(), unrolling
and loop unswitching should not be allowed, but on the other hand e.g. full
unrolling and inlining should be allowed. I'm not sure if you're saying
that there should be no limitations as the use cases you had mind for
convergent are always OK with such duplication, or that we should be
figuring this stuff out on a case-by-case basis, or something else?
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the llvm-commits