<div dir="ltr"><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
The host runtimes on the other side are linked dynamically and IMO it's<br>
one of the strengths of libomp to mix-and-match versions. Unless<br>
there's a compelling reason, I'll make another plea to keep<br>
compatibility.<br></blockquote><div> </div><div>I also value allowing mix-and-match versions a critical feature long term.</div><div>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.</div><div>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.</div><div>This is even out of our control. So I believe we potentially need more version checks to ensure no hassle to users.<br></div><div> 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.</div><div>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.<br></div><div><br></div><div>Best,</div><div>Ye<br></div></div></div>