[Openmp-dev] [RFC] Clarify the absence of API stability for the *device runtime* (aka. libomptarget-nvptx-sm_XX.bc)

[Ye Luo via Openmp-dev]

>
> The host runtimes on the other side are linked dynamically and IMO it's
> one of the strengths of libomp to mix-and-match versions. Unless
> there's a compelling reason, I'll make another plea to keep
> compatibility.
>

I also value allowing mix-and-match versions a critical feature long term.
If I wrote a library and compiled it based on one version while the
application is compiled and linked by another version, we hope everything
works without recompiling the library.
Unfortunately, this mix-and-match scheme is quite brittle at the current
stage. I still remember when I built libomptarget with CUDA9, CUDA 10.2
refused to link some device binary code.
This is even out of our control. So I believe we potentially need more
version checks to ensure no hassle to users.
At this point, we are quickly developing OpenMP features. Making fast moves
helps us reach the point when the implementation is relatively matured and
robust, then we can do mix-and-match.
I hope the consensus is that we are not giving up but just making a
trade-off at this point of time and we will adjust based on the situation.

Best,
Ye
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/openmp-dev/attachments/20200709/d69dfeba/attachment.html>